CN107851233A - Local analytics at assets - Google Patents
Local analytics at assets Download PDFInfo
- Publication number
- CN107851233A CN107851233A CN201680043854.5A CN201680043854A CN107851233A CN 107851233 A CN107851233 A CN 107851233A CN 201680043854 A CN201680043854 A CN 201680043854A CN 107851233 A CN107851233 A CN 107851233A
- Authority
- CN
- China
- Prior art keywords
- assets
- forecast model
- data
- workflow
- personalized
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000006116 polymerization reaction Methods 0.000 claims abstract description 139
- 238000000034 method Methods 0.000 claims abstract description 70
- 230000008859 change Effects 0.000 claims description 80
- 238000004891 communication Methods 0.000 claims description 53
- 230000003862 health status Effects 0.000 claims description 24
- 238000013396 workstream Methods 0.000 claims description 24
- 238000004519 manufacturing process Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000010304 firing Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 description 316
- 230000004048 modification Effects 0.000 description 54
- 238000012986 modification Methods 0.000 description 54
- 230000036541 health Effects 0.000 description 39
- 230000004044 response Effects 0.000 description 29
- 238000013500 data storage Methods 0.000 description 25
- 238000005259 measurement Methods 0.000 description 24
- 238000012545 processing Methods 0.000 description 24
- 108090000623 proteins and genes Proteins 0.000 description 18
- 230000006399 behavior Effects 0.000 description 14
- 230000001960 triggered effect Effects 0.000 description 13
- 238000012544 monitoring process Methods 0.000 description 12
- 238000012360 testing method Methods 0.000 description 12
- 238000001514 detection method Methods 0.000 description 11
- 238000003860 storage Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 8
- 230000002159 abnormal effect Effects 0.000 description 7
- 230000009471 action Effects 0.000 description 7
- 230000001419 dependent effect Effects 0.000 description 7
- 230000003137 locomotive effect Effects 0.000 description 7
- 230000005856 abnormality Effects 0.000 description 6
- 239000000446 fuel Substances 0.000 description 6
- 238000012423 maintenance Methods 0.000 description 6
- 230000004888 barrier function Effects 0.000 description 5
- 230000000704 physical effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 230000007257 malfunction Effects 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 4
- 230000000712 assembly Effects 0.000 description 3
- 238000000429 assembly Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 235000013399 edible fruits Nutrition 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 230000003449 preventive effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000008439 repair process Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000003442 weekly effect Effects 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000006837 decompression Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000005059 dormancy Effects 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000001976 improved effect Effects 0.000 description 1
- 238000010348 incorporation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 239000003973 paint Substances 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000008093 supporting effect Effects 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0208—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the configuration of the monitoring system
- G05B23/0213—Modular or universal configuration of the monitoring system, e.g. monitoring system having modules that may be combined to build monitoring program; monitoring system that can be applied to legacy systems; adaptable monitoring system; using different communication protocols
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0286—Modifications to the monitored process, e.g. stopping operation or adapting control
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0633—Workflow analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Automation & Control Theory (AREA)
- Manufacturing & Machinery (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Transfer Between Computers (AREA)
Abstract
There is disclosed herein the system related to assets, device and method and the forecast model related with Asset operation and corresponding workflow.In particular, example is related to;Definition and deployment polymerization forecast model and corresponding workflow;Definition and the personalized forecast model of deployment and/or corresponding workflow;And dynamically adjust the execution of model workflow pair.In addition, example, which is related to, is configured to receive and locally executes forecast model, local personalized forecast model and/or the assets for locally executing workflow or part thereof.
Description
The cross reference of related application
Present application advocates the priority of following item:(i) application on June 19th, 2015 and entitled it is used to locally execute
It polymerize forecast model and workflow (Aggregate Predictive Model&Workflow for Local Execution)
No. 14/744,352 U.S. Non-provisional Patent case;(ii) application on June 19th, 2015 and the entitled individual character for assets
Change the of forecast model and workflow (Individualized Predictive Model&Workflow for an Asset)
No. 14/744,369 U.S. Non-provisional Patent application case;And (iii) on December 8th, 2015 applies and the entitled " sheet at assets
14/963rd, No. 207 U.S. Non-provisional Patent application case of ground analysis (Local Analytics at an Asset) ", it is described
During the full content of each of non-provisional case is fully incorporated herein with incorporation way.Present application is also with reference
Mode is incorporated with the of application on June 5th, 2015 and entitled " assets health status fraction (Asset Health Score) "
The full content of No. 14/732,258 U.S. Non-provisional Patent application case.
Background technology
Today, machine are (also referred herein as " assets ") ubiquitous in many industries.From goods handling to
To helping nurse and doctor to save the Medical Devices of life, assets play an important role the locomotive of various countries in daily life.
Depending on assets role, its complexity and cost may be different.For example, some assets may include multiple subsystems
Unite (for example, the engine of locomotive, transmission device etc.), its must coordinated manipulation so that assets suitably operate.
, can be within limited downtime it is desirable to assets because assets play important role in daily life
Repaired.Therefore, some have developed each mechanism to monitor and detect the unusual condition in assets, so that promote may be
Maintenance assets in minimum downtime.
The content of the invention
Current method for monitoring assets is usually directed to computer in assets, and it is various in whole assets from being distributed in
Sensor and/or actuator reception signal, the operating conditions of sensor and/or actuator the monitoring assets.As a generation
Table example, if assets are locomotives, then sensor and/or actuator can monitor such as parameter of temperature, voltage and speed
And other parameters.If sensor and/or actuator signal from one or more of these devices reach some values, that
Computer can produce unusual condition designator, such as " failure code " in the assets, and it is that exception is had occurred and that in assets
The instruction of situation.
Generally, the defects of abnormal conditions can be at assets or its component, it may cause assets and/or the failure of component.
Thus, abnormal conditions can be associated with given failure or possible multiple failures, because abnormal conditions are given failure or multiple events
The sign of barrier.In fact, user's sensor that generally definition is associated with each unusual condition designator and corresponding sensor
Value.That is, user defines " normal " operating conditions (for example, not triggering the operating conditions of failure code) and the "abnormal" operation of assets
Situation (for example, operating conditions of triggering failure code).
After computer produces unusual condition designator in assets, designator and/or sensor signal can be passed to
Remote location, in the remote location, user, which can receive some instructions of unusual condition and/or sensor signal and determine, is
It is no to take action.The action that user may take is to assign machinist etc. to assess and possibly repair assets.Once place
At assets, computing device can be connected to assets by machinist and operation calculation device is to cause one or more locals of assets utilization
The reason for diagnostic tool is to promote to diagnose caused designator.
Although current assets monitoring system is typically effective, such system in terms of abnormal conditions designator is triggered
System is typically conservative.That is, when triggering designator to assets monitoring system, the failure in assets may have occurred and that (or
Will occur), this may cause downtime of a high price and other shortcomings.Further, since such assets monitoring system
In assets on abnormality detection mechanism simple essence, current assets monitoring method tends to be related to remote computing system to money
Production performs monitoring calculation, and if detecting problem, then to the assets firing order.When assets are moved to communication network
When outside coverage, due to network delay and/or infeasible, this is probably unfavorable.Further, since it is stored in assets
The essence of local diagnostic tool, current diagnostic program tends to not only poorly efficient but also trouble, because needing machinist to cause
The such instrument of assets utilization.
Instance system disclosed herein, device and method attempt help and solve one or more of these problems.In reality
In example embodiment, network configuration can include communication network, and it promotes the communication between assets and remote computing system.At some
In the case of, communication network can promote the secure communication between assets and remote computing system (for example, via encryption or other safety
Measure).
As described above, each assets can include the multiple sensors and/or actuator being distributed in whole assets, and it promotees
Enter to monitor the operating conditions of assets.Multiple assets can provide the corresponding of the operating conditions of each assets of instruction to remote computing system
Data, the remote computing system can be configured to be performed one or more operations based on the data provided.Generally, sensor
And/or actuator data can be used for the monitoring operation of in general assets.However, as described herein, remote computing system and/or money
Production can promote to perform more complicated operation using this data.
In example implementation, remote computing system can be configured to define and dispose related pre- of operation to assets
Survey model and corresponding workflow (referred to herein as " model-workflow to ").Assets can be configured to receive model-work
Stream pair, and using local analytics device according to model-workflow to operating.
Generally, model-workflow pair can cause assets to monitor some operating conditions, and when some situations be present, modification
The behavior that particular event occurs may be helped prevent.Specifically, forecast model can be received from asset sensor and/or actuating
The data of the specific collection of device as input, and export one or more particular events may be in following special time period in assets
Locate the possibility occurred.Workflow can relate to one that possibility based on one or more particular events exported by model performs or
Multiple operations.
In fact, the polymerization of remote computing system definable, forecast model and corresponding workflow, personalized forecast model and right
Answer workflow, or its certain combination." polymerization " model/workflow may refer to the model/workflow general to assets group, and
" personalization " model/workflow may refer to model/work for the single assets or assets subgroup customization from assets group
Flow.
In example implementation, remote computing system can pass through the historical data definition polymerization prediction based on multiple assets
Model starts.Using the data of multiple assets definition Billy can be promoted more accurately to predict mould with the operation data of single assets
Type.
At least operand of operating conditions of the given assets of instruction can be included by forming the basic historical data of polymerization model
According to.Specifically, the unusual condition data and/or refer to that operation data can include the situation for identifying when to break down at assets
Show the data of one or more physical properties measured when the situation occurs at assets.Data can also include instruction assets
Date that when environmental data of the environment operated wherein and instruction assets are utilized and the scheduling data of time, for fixed
Assets related data of adopted polymerization model-workflow pair etc..
Based on historical data, the polymerization model of remote computing system definable prediction particular event generation.Specific real
In example embodiment, the exportable probability that will be broken down in following special time period at assets of polymerization model.This model
Can referred to herein as " fault model ".In addition to other example forecast models, assets can be predicted in other polymerization models will be
The possibility of completion task in following special time period.
After polymerization model is defined, remote computing system can then define the polymerization corresponding to defined polymerization model
Workflow.Generally, workflow can include assets and can be operated based on one or more of corresponding model execution.That is, the defeated of model is corresponded to
Go out to cause assets to perform workflow operation.For example, polymerization model-workflow works as polymerization model output spy to can be defined so that
When determining the probability in scope, assets will perform particular workflow operation (such as local diagnostic tool).
Polymerization model-workflow is being defined to afterwards, it is described right that remote computing system can launch to one or more assets.One
Or multiple assets then can be according to polymerization model-workflow to operating.
In example implementation, remote computing system can be configured further to define the personalization of one or more assets
Forecast model and/or corresponding workflow.Remote computing system can be based on each given assets some characteristics and other considerations
To carry out above-mentioned definition.In example implementation, remote computing system can be using polymerization model-workflow to being opened as benchmark
Begin, and one or both in the polymerization model of the personalized given assets of characteristic based on assets and workflow.
In fact, remote computing system can be configured to determine to polymerization model-workflow to related asset character (example
Such as, characteristic of interest).The example of this class feature can include the assets time limit, assets behaviour in service, class of assets (for example, brand
And/or model), other characteristics such as the operating environment of assets health status and assets.
Then, remote computing system can determine that the characteristic of the given assets corresponding to characteristic of interest.At least based on given
Some in asset character, remote computing system can be configured with personalized polymerization model and/or corresponding workflow.
Define personalized model and/or workflow can relate to remote computing system and polymerization model and/or workflow are carried out
Some modifications.For example, in addition to other examples, personalized polymerization model may relate to change mode input, change model calculating,
And/or change the weight of the variable calculated or output etc..In addition to other examples, personalization polymerization workflow may relate to change
One or more operations for becoming workflow and/or the model output valve or value scope that change triggering workflow.
After the personalized model of the given assets of definition and/or workflow, remote computing system can be then to given money
Production transmitting personalized model and/or workflow.In the case that only one in wherein model or workflow is personalized, give
Assets can utilize model or the polymerization version not being personalized of workflow.Given assets can then according to its personalized model-
Workflow is to operating.
In example implementation, given assets can include local analytics device, and it can be configured to cause given assets
According to by model-workflow that remote computing system provides to operating.Local analytics device can be configured with using from money
The operation data (for example, being generally used for the data of other assets correlation purpose) of production sensor and/or actuator is predicted to run
Model.When local analytics device receives some operation datas, it can perform the model, and depend on the defeated of the model
Go out, can perform corresponding workflow.
Performing corresponding workflow can help to promote to prevent that unexpected event occurs at given assets.In this way, give
Assets can locally determine that particular event may occur, and then can perform particular workflow to help prevent generation event.If
Given communication between assets and remote computing system is obstructed, then this may be particularly useful.For example, in some cases, therefore
Barrier, which is likely to occur in, takes the order of preventive actions to be reached from remote computing system before giving assets.In such cases,
Local analytics device can be favourable, because thus it can avoid any network delay or due to given in locally generated order
Any problem caused by assets " offline ".Thus, local analytics device, which performs model-workflow pair, can promote to cause assets to fit
Answer its situation.
In some example implementations, prior to or just when model-workflow pair is performed first, local analytics device sheet
Its model-workflow pair for being received from remote computing system of body personalizable.Generally, local analytics device can be by assessing
Some or all predictions, hypothesis and/or the vague generalization related to given assets carried out during Definition Model-workflow pair comes individual
Property model-workflow pair.Based on assessment, local analytics device can change model-workflow pair so that model-workflow pair
Basic forecast, hypothesis and/or vague generalization more accurately reflect the virtual conditions of given assets.Local analytics device can then be held
Row personalized model-workflow pair, rather than its model-workflow pair initially received from remote computing system is performed, this can
Cause the more accurately monitoring to assets.
When given assets according to model-workflow to operation when, given assets can also continue to provide to remote computing system
Operation data.Based at least this data, remote computing system can change polymerization model-workflow pair and/or one or more individual characteies
Change model-workflow pair.Remote computing system may be modified due to many reasons.
In an example, if the new events that the model had not accounted for previously occur at assets, then long-range
Computing system can change model and/or workflow.For example, in fault model, new events are probably to be used to define in its data
The still nonevent new failure in any one of assets of polymerization model place.
In another example, if the thing occurs under the operating conditions that will not generally cause event to occur at assets
Part, then remote computing system can change model and/or workflow.For example, again return to fault model, if the past still
Do not result in and the failure occurs under the operating conditions of failure generation, then the fault model or corresponding workflow can be changed.
In a further example, if performed workflow fails the generation of prevention event, then remote computing system can
Change model and/or workflow.Specifically, if the output of model causes assets to perform the work for being intended to prevent event
Stream is but despite of that the event still occurs at assets, then remote computing system can change model and/or workflow.
The other examples for the reason for changing model and/or workflow are also possible.
Remote computing system can then distribute to any modification assets and/or and remote computation that its data causes modification
The other assets of system communication.In this way, remote computing system dynamically changes model and/or workflow, and based on individual
Such modification is distributed to cluster assets by the operating conditions of other assets.
In some example implementations, assets and/or remote computing system can be configured dynamically to adjust perform prediction
Model and/or workflow.In particular, assets and/or remote computing system can be configured to detect triggering on assets and/or
Remote computing system whether some events of the change of the responsibility of perform prediction model and/or workflow.
For example, in some cases, model-workflow is received to afterwards, assets can from remote computing system in assets
By model-workflow to being stored in data storage device, and then can be dependent on remote computing system intensively perform model-
Workflow centering it is part or all of.On the other hand, in other cases, remote computing system can be dependent on assets locally to hold
Row model-workflow centering it is part or all of.In the case of again other, remote computing system and assets can share execution mould
The responsibility of type-workflow pair.
Anyway, at some time point, it may occur however that triggering assets and/or remote computing system adjustment forecast model
And/or some events of the execution of workflow.For example, assets and/or remote computing system can detect and be coupled to assets remotely
Some characteristics of the communication network of computing system.Characteristic based on communication network, assets can adjust its whether be performed locally it is pre-
Survey model and/or workflow, and remote computing system can therefore change its whether concentrative implementation model and/or workflow.With this
Mode, assets and/or remote computing system are suitable for the situation of assets.
In particular instances, the detectable following instruction of assets:The letter of communication link between assets and remote computing system
Number intensity relatively weak (for example, assets can determine that will " offline "), network delay are of a relatively high, and/or network bandwidth is relative
It is relatively low.Therefore, assets can be programmed to undertake the responsibility of previously execution model-workflow pair by remote computing system disposal.
And then remote computing system can stop concentrative implementation model-workflow to some or all of.In this way, assets can be
Forecast model is locally executed, and is next based on perform prediction model to perform corresponding workflow potentially to help prevent at assets
Break down.
In addition, in some embodiments, assets and/or remote computing system can be similarly based on various other considerations and come
Adjustment performs (or may change) forecast model and/or workflow.For example, the disposal ability based on assets, assets can be in local
Model-workflow pair is performed, and therefore remote computing system can be also adjusted.In another example, based on assets are coupled
To the bandwidth of the communication network of remote computing system, executable the changed workflow of assets is (for example, according to data emission case
Data are launched to remote computing system with the emission rate of reduction).Other examples are also possible.
As discussed above, examples provided herein is related to the deployment and execution of forecast model.In an aspect, carry
A kind of computing system is supplied.The computing system includes at least one processor, non-transitory computer-readable media and storage
Programmed instruction in the non-transitory computer-readable media, described program instruction can be held by least one processor
Go to cause the computing system:(a) the corresponding operating data of multiple assets are received;(b) based on the received operand
According to the definition forecast model related to the operation of the multiple assets and corresponding workflow;And (c) into the multiple assets
At least one assets launch the forecast model and corresponding workflow so that at least one assets locally execute.
In another aspect, there is provided a kind of non-transitory computer-readable media for being stored with instruction above, the finger
Order can perform to cause computing system:(a) the corresponding operating data of multiple assets are received;(b) based on the received operation
Data, define the forecast model related to the operation of the multiple assets and corresponding workflow;And (c) into the multiple assets
At least one assets launch the forecast model and corresponding workflow so that at least one assets locally execute.
In another aspect, there is provided a kind of computer implemented method.Methods described includes:(a) multiple assets are received
Corresponding operating data;(b) based on the received operation data, the prediction mould related to the operation of the multiple assets is defined
Type and corresponding workflow;And at least one assets of (c) into the multiple assets launch the forecast model and corresponding work
Stream locally executes at least one assets.
As discussed above, examples provided herein is related to the deployment and execution of forecast model.In an aspect, carry
A kind of computing system is supplied.The computing system includes at least one processor, non-transitory computer-readable media and storage
Programmed instruction in the non-transitory computer-readable media, described program instruction can be held by least one processor
Go to cause the computing system:(a) operation data of multiple assets is received, wherein the multiple assets include the first assets;
(b) based on the received operation data, the polymerization forecast model related to the operation of the multiple assets and polymerization are defined
Corresponding workflow;(c) one or more characteristics of first assets are determined;(d) based on the described one or more of first assets
Individual characteristic and the polymerization forecast model and the corresponding workflow of the polymerization, define the operation phase with first assets
At least one of the personalized forecast model of pass or personalized corresponding workflow;And described in (e) to first assets transmitting
Defined at least one personalized forecast model or the personalized workflow that corresponds to are so that first assets locally execute.
In another aspect, there is provided a kind of non-transitory computer-readable media for being stored with instruction above, the finger
Order can perform to cause computing system:(a) operation data of multiple assets is received, wherein the multiple assets include the first money
Production;(b) based on the received operation data, define the polymerization forecast model related to the operation of the multiple assets and gather
Close corresponding workflow;(c) one or more characteristics of first assets are determined;(d) based on first assets described one or
Multiple characteristics and the polymerization forecast model and the corresponding workflow of the set, definition and the operation of first assets
At least one of related personalized forecast model or personalized corresponding workflow;And (e) launches institute to first assets
At least one personalized forecast model defined in stating or personalized corresponding workflow are so that first assets locally execute.
In another aspect, there is provided a kind of computer implemented method.Methods described includes:(a) multiple assets are received
Operation data, wherein the multiple assets include the first assets;(b) based on the received operation data, definition with it is described
The polymerization forecast model and the corresponding workflow of polymerization of the operation correlation of multiple assets;(c) the one or more of first assets are determined
Individual characteristic;(d) based on one or more characteristics described in first assets and the polymerization forecast model and the polymerization pair
Workflow is answered, is defined in the personalized forecast model related to the operation of first assets or personalized corresponding workflow
At least one;And (e) launches defined at least one personalized forecast model or personalization to first assets
Corresponding workflow locally executes for first assets.
As discussed above, examples provided herein, which is related at assets, receives simultaneously perform prediction model and/or work
Stream.In an aspect, there is provided a kind of computing device.The computing device includes:(i) asset interface, its be configured to by
The computing device is coupled to assets;(ii) network interface, it is configured to promote the computing device and the remote calculating
Communication between the computing system of device positioning;(iii) at least one processor;(iv) non-transitory computer-readable media;
And (v) is stored in the programmed instruction in the non-transitory computer-readable media, described program instruction can be by described at least one
Individual computing device is to cause the computing device:(a) received via the network interface related to the operation of the assets
Forecast model, wherein the forecast model is defined by the computing system based on the operation data of multiple assets;(b) via institute
State the operation data that asset interface receives the assets;(c) at least part of the operation data received based on the assets
Perform the forecast model;And (d) is based on performing the forecast model, performs the workflow corresponding to the forecast model, its
It is middle to perform the workflow including causing the assets to perform operation via the asset interface.
In another aspect, there is provided a kind of non-transitory computer-readable media for being stored with instruction above, the finger
Order can perform to cause the asset interface via computing device to be coupled to the computing device of assets:(a) via the calculating
The network interface of device receives the forecast model related to the operation of the assets, and the network of the computing device connects
Mouth is configured to promote the communication between the computing device and the computing system positioned away from the computing device, wherein described
Forecast model is defined by the computing system based on the operation data of multiple assets;(b) institute is received via the asset interface
State the operation data of assets;(c) at least part of the operation data received based on the assets performs the forecast model;
And (c) is based on performing the forecast model, the workflow corresponding to the forecast model is performed, wherein performing the workflow packages
Include to cause to state assets via the asset interface and perform operation.
In another aspect, there is provided a kind of computer implemented method.Methods described includes:(a) via computing device
Network interface receives the forecast model related to the operation of assets, and the network interface of the computing device is via described
The asset interface of computing device is coupled to the assets, wherein the forecast model is as described in away from computing device positioning
Computing system is defined based on the operation data of multiple assets;(b) institute is received via the asset interface by the computing device
State the operation data of assets;(b) held by least part of the operation data that is received of the computing device based on the assets
The row forecast model;And (c) is based on performing the forecast model, is performed by the computing device and corresponds to the forecast model
Workflow, wherein performing the workflow includes via the asset interface causing the assets to perform operation.
Art those skilled in the art are readily apparent that these and many other aspects when reading following disclosure.
Brief description of the drawings
Fig. 1 depict wherein can embodiment embodiment example network configuration.
Fig. 2 depicts the simplified block diagram of example assets.
Fig. 3 depicts the conceptual illustration of example unusual condition designator and triggering criterion.
Fig. 4 depicts the simplified block diagram of instance analysis system.
Fig. 5 depicts the example flow chart of the definition phase available for Definition Model-workflow pair.
Fig. 6 A depict the conceptual illustration of polymerization model-workflow pair.
Fig. 6 B depict the conceptual illustration of personalized model-workflow pair.
Fig. 6 C depict the conceptual illustration of another personalized model-workflow pair.
Fig. 6 D depict the conceptual illustration of changed model-workflow pair.
Fig. 7 depicts the example flow chart of the modelling phase of the forecast model available for definition output health indicator.
Fig. 8 depicts the conceptual illustration of the data for Definition Model.
Fig. 9 is depicted available for the example flow chart for locally executing the stage for locally executing forecast model.
Figure 10 depicts the example flow chart of the modification stage available for modification model-workflow pair.
Figure 11 depicts the example flow chart of the adjusting stage of the execution available for adjustment model-workflow pair.
Figure 12 depicts the flow chart of the case method for defining and disposing polymerization forecast model and corresponding workflow.
Figure 13 depicts the flow of the case method for defining and disposing personalized forecast model and/or corresponding workflow
Figure.
Figure 14 depicts the flow chart of the case method of the execution for dynamically changing model-workflow pair.
Figure 15 depicts the flow chart of the case method for receiving and locally executing model-workflow pair.
Embodiment
Content refer to the attached drawing disclosed below and several exemplary cases.One of ordinary skill in the art will be understood that,
Such reference is only used for the purpose explained, and is therefore not intended to limit.Disclosed system, the part in device and method or
All it can in a variety of ways rearrange, combine, add and/or remove, each of which mode is paid attention to herein.
I. example network configures
Turning now to schema, Fig. 1 depict wherein can embodiment embodiment example network configuration 100.As indicated, net
Network configuration 100 comprising assets 102, assets 104, communication network 106, can take analysis system form remote computing system
108th, output system 110 and data source 112.
Communication network 106 is communicatively coupled each of component in network configuration 100.For example, assets 102 and
104 can communicate via communication network 106 with analysis system 108.In some cases, assets 102 and 104 can with one or more
Between system (such as assets gateway (not describing)) communicate, the assets gateway and then communicated with analysis system 108.Similarly, divide
Analysis system 108 can communicate via communication network 106 with output system 110.In some cases, analysis system 108 can with one or
The communication of multiple intermediate systems (such as host server (not describing)), the host server and then communicates with output system 110.
Many other configurations are also possible.In the case of an instance, communication network 106 can promote the secure communication between networking component
(for example, via encryption or other safety measures).
Generally, assets 102 and 104, which can be taken, is configured to perform one or more operations (it can be defined based on field)
The form of any device, and can also include setting for the data for one or more operating conditions for being configured to the given assets of transmitting instruction
It is standby.In some instances, assets can include one or more subsystems for being configured to perform one or more corresponding operatings.It is actual
On, multiple subsystems parallel or can be operated sequentially so that Asset operation.
Example assets can include transportation machines (for example, locomotive, aircraft, passenger vehicle, semi-mounted truck, ship etc.), industry
Machine (for example, winning equipment, Architectural Equipment, automation equipment in factory etc.), medical machine are (for example, medical imaging devices, surgery
Surgical apparatus, medical monitoring system, medical laboratory's equipment etc.) and practical machine (for example, turbine, solar energy farm etc.) etc.
Deng.One of ordinary skill in the art are readily apparent that these are only several examples of assets, and several other assets are herein
In it is possible and pay attention to.
In example implementation, assets 102 and 104 can be respectively identical type (for example, one group of locomotive or aircraft,
Wind park or MRI machine set etc.), thereby increases and it is possible to it is identical category (for example, same brand and/or model).At it
In its example, assets 102 and 104 may be different in type, brand, model etc..It is discussed in further detail below with reference to Fig. 2
Assets.
As indicated, assets 102 and 104 and possible data source 112 can be logical via communication network 106 and analysis system 108
Letter.Generally, communication network 106, which can include, is configured to promote one or more computing systems for transmitting data among components of the networks
And network infrastructure.Communication network 106 can be or can include can be wired and/or wireless and supporting secure communication one or
Multiple wide area networks (WAN) and/or LAN (LAN).In some instances, communication network 106 can include one or more Cellular Networks
The network such as network and/or internet.Communication network 106 can according to such as LTE, CDMA, GSM, LPWAN, WiFi, bluetooth, Ethernet,
HTTP/S, TCP, CoAP/DTLS etc. one or more communication protocols are operated.Although communication network 106 is illustrated as single net
Network, it should be understood that communication network 106 can include the multiple different networks itself communicatedly linked.Communication network 106
Also other forms can be taken.
As described above, analysis system 108 can be configured to receive data from assets 102 and 104 and data source 112.One
As for, analysis system 108, which can include, to be configured to receive, handles, analyzing and one or more computing systems of output data, example
Such as server and database.Analysis system 108 can be according to data-oriented Flow Technique (such as TPL Dataflow or NiFi etc.)
To configure.Analysis system 108 is discussed in further detail below with reference to Fig. 3.
As indicated, analysis system 108 can be configured to launch data to assets 102 and 104 and/or output system 110.Institute
The specific data of transmitting can take various forms, and will be described in details further below.
Generally, output system 110, which can be taken, is configured to receive data and provides the computing system of some form of output
Or the form of device.Output system 110 can take various forms.In an example, output system 110 can be or comprising output
Device, it is configured to receive data and the sense of hearing, vision and/or tactile output is provided in response to the data.Generally, export
Device can include one or more input interfaces for being configured to receive user's input, and output device can be configured with based on this use
Family inputs and launches data by communication network 106.The example of output device includes tablet PC, smart phone, on knee
Computer, other mobile computing devices, desktop computer, intelligent television etc..
Another example of output system 110 can take work to make the form of system, and it is defeated that the work makes system be configured to
The machinist etc. that sends as an envoy to repairs the request of assets.Another example again of output system 110 can take the part being configured to assets
Place an order and export the form of the part order system of its receipt.Many other output systems are also possible.
Data source 112 can be configured to be communicated with analysis system 108.Generally, data source 112 can be or comprising one or more
Computing system, it is configured to collection, data storage and/or provides data to other systems (such as analysis system 108), described
Data can be related to the function of being performed by analysis system 108.Data source 112 can be configured to be produced independently of assets 102 and 104
And/or obtain data.Thus, the data provided by data source 112 are referred to alternatively as " external data " herein.Data source 112
It can be configured to provide current and/or historical data.In fact, what analysis system 108 can be provided by " subscription " by data source
Service to receive data from data source 112.However, analysis system 108 can also other manner from data source 112 receive data.
The example of data source 112 includes environmental data source, asset management data source and other data sources.Generally, environment number
The data of certain characteristic of the operating environment of instruction assets are provided according to source.The example in environmental data source, which includes, to be provided on given area
The nature in domain and the meteorological data server of the information of artificial feature, GLONASS (GNSS) server, map number
According to server and topological data server etc..
Generally, asset management data source provides the entity of operation or maintenance that instruction can influence assets (for example, other moneys
Production) event or state data (for example, assets can when and where operation or receive safeguard).Asset management data source
Example includes:Traffic data server, it is provided on aerial, waterborne and/or traffic above-ground information;Asset deployment service
Device, it provides the information in the desired path and/or position of specific date and/or special time on assets;Defect detector
System (also referred to as " hot tank " detector), it provides one or more operations on the assets by defect detector arrangement adjacent
The information of situation;Parts supplier server, it is provided on the part and its information of price in the stock of specific supplier;
And maintenance shop's server, it provides the information on maintenance shop's production capacity etc.;Etc..
The example of other data sources includes the grid service device for providing the information on power consumption and stores going through for assets
External data base of history operation data etc..One of ordinary skill in the art are readily apparent that these are only the several of data source
Individual example, and several other examples are possible.
It should be understood that network configuration 100 is an example of the network that can wherein implement embodiment described herein.
Several other arrangements are possible and paid attention to herein.For example, other network configurations can include the additional set do not described
Part and/or more or less described components.
II. example assets
Fig. 2 is turned to, depicts the simplified block diagram of example assets 200.Any one of assets 102 and 104 from Fig. 1 or
Both can be configured as assets 200.As indicated, assets 200 can include one or more subsystems 202, one or more
Sensor 204, one or more actuators 205, CPU 206, data storage device 208, network interface 210, user
Interface 212 and local analytics device 220, its all can be by system bus, network or other bindiny mechanisms (directly or indirectly
Ground) communicatedly link.One of ordinary skill in the art are readily apparent that, assets 200 can include the additional components that do not show and/or
More or less described components.
In general, assets 200 can include be configured to perform one or more operations one or more electrically, machinery and/
Or electromechanical assemblies.In some cases, one or more components can be grouped into in stator system 202.
Generally, subsystem 202 can include the associated component group of the part as assets 200.Single subsystem 202 can be only
On the spot perform one or more operations, or single subsystem 202 can be operated together with one or more other subsystems with perform one or
Multiple operations.Generally, the assets of different types of assets and even different classes of same type can include different sub-systems.
For example, under the background of transportation asset, the example of subsystem 202 can include engine, speed change gear, power train,
Fuel system, battery system, gas extraction system, brakes, electrical system, signal processing system, generator, gear-box, rotor
And hydraulic system, and several other subsystems.Under the background of medical machine, the example of subsystem 202 can include scanning system
System, motor, coil and/or magnet system, signal processing system, rotor and electrical system, and several other subsystems.
As indicated above, assets 200 can be equipped with:Various sensors 204, it is configured to the behaviour for monitoring assets 200
Make situation;And various actuators 205, it is configured to assets 200 or its component interaction and monitors the operation shape of assets 200
Condition.In some cases, some in sensor 204 and/or actuator 205 can be grouped based on particular subsystem 202.With this
The group of mode, sensor 204 and/or actuator 205 can be configured to monitor the operating conditions of particular subsystem 202, and come
It is can be configured from the actuator of the group to be interacted in a manner with particular subsystem 202, the mode can be based on that
A little operating conditions and change the behavior of subsystem.
Generally, sensor 204 can be configured to detect the physical of one or more operating conditions for may indicate that assets 200
Matter, and the instruction of detected physical property is provided, such as electric signal.In operation, sensor 204 can be configured with continuous
Ground, periodically (for example, based on sample frequency) and/or obtain measured value in response to some trigger event.In some examples
In, sensor 204 can be provided with the operating parameter for performing measurement in advance and/or can be carried according to by CPU 206
The operating parameter (for example, indication sensor 204 obtains sampled signal of measured value) of confession performs measurement.In instances, it is different
Sensor 204 can have different operating parameter (for example, some sensors can be sampled based on first frequency, and other sensings
Device is sampled based on the second different frequency).Anyway, sensor 204 can be configured to launch to CPU 206
Indicate the electric signal of measured physical property.Such signal is continuously or periodically supplied to center by sensor 204
Processing unit 206.
For example, sensor 204 can be configured to measure the physical property of assets 200, for example, the position of assets 200 and/or
Mobile, in said case, sensor can take GNSS sensors, the sensor based on dead reckoning, accelerometer, gyro
The forms such as instrument, pedometer, magnetometer.
In addition, various sensors 204 can be configured to measure other operating conditions of assets 200, the example can include temperature
Degree, pressure, speed, rate of acceleration or the rate of deceleration, friction, power usage amount, fuel usage amount, liquid level, run time, voltage and electricity
Stream, magnetic field, electric field, the existence or non-existence of object, the position of component and generating etc..One of ordinary skill in the art will
Understand, these are only that sensor can be configured some example operation situations with measurement.Depending on sector application or specific money
Production, can be used more or less sensors.
As indicated above, the configuration of actuator 205 can be similar to sensor 204 in some aspects.Specifically, activate
Device 205 can be configured to detect instruction assets 200 operating conditions physical property, and with the similar mode of sensor 204
Its instruction is provided.
In addition, actuator 205 can be configured to be handed over assets 200, one or more subsystems 202 and/or some components
Mutually.Thus, actuator 205, which can include, to be configured to perform mechanically actuated (for example, mobile) or otherwise control assembly, son
Motor of system or system etc..In particular instances, actuator can be configured to measure fuel flow rate and change fuel flow rate
(such as limiting fuel flow rate), or actuator can be configured to measure hydraulic pressure and change hydraulic pressure (for example, increasing or subtracting
Small hydraulic pressure).Several other example interactions of actuator are also possible and paid attention to herein.
Generally, CPU 206 can include one or more processors and/or controller, and it can take general or special
With the form of processor or controller.Specifically, in example implementation, CPU 206 can be or comprising micro-
Processor, microcontroller, application specific integrated circuit, digital signal processor etc..And then data storage device 208 can be or comprising
One or more non-transitory computer-readable storage mediums, such as optics, magnetic, organic or flash memory etc..
CPU 206 can be configured to store, access and perform the calculating being stored in data storage device 208
Machine readable program instructions, to perform the operation of assets described herein.For example, as indicated above, CPU 206
It can be configured to receive corresponding sensor signal from sensor 204 and/or actuator 205.CPU 206 can through with
Put and accessed so that sensor and/or actuator data are stored in data storage device 208 and then from data storage device 208
The data.
CPU 206 also can be configured to determine whether received sensor and/or actuator signal trigger
Any unusual condition designator, such as failure code.For example, CPU 206 can be configured with data storage device
Unusual condition rule is stored in 208, each of which rule includes the given unusual condition designator for representing specific exceptions situation
And the corresponding triggering criterion of triggering unusual condition designator.That is, each unusual condition designator corresponds to be indicated in abnormality
Accord with one or more sensors and/or actuator measured value that is triggered and must be satisfied for before.In fact, assets 200 can be pre-
It is programmed with unusual condition rule and/or can to receive new unusual condition from computing system (such as analysis system 108) regular or right
Existing well-regulated renewal.
Anyway, CPU 206 can be configured is with the received sensor of determination and/or actuator signal
No any unusual condition designator of triggering.That is, CPU 206 can determine that received sensor and/or actuator letter
Number whether meet any triggering criterion.When this is defined as certainly, CPU 206 can produce abnormality data, and
The instruction of the output abnormality situation of user interface 212 of assets, such as vision and/or audible alarms can also be caused.In addition, centre
Reason unit 206 can may record the generation of triggered unusual condition designator with time stab in data storage device 208.
Fig. 3 depicts the example unusual condition designator of assets and the conceptual illustration of corresponding triggering criterion.In particular, scheme
3 depict the conceptual illustration of Instance failure code.As indicated, table 300, which includes, corresponds respectively to sensors A, actuator B and sensing
Device C row 302,304 and 306 and the row 308,310 and 312 for corresponding respectively to failure code 1,2 and 3.Entry 314 is then specified
Sensor criterion (for example, sensor value threshold) corresponding with given failure code.
For example, when sensors A detects that wheel measuring value and sensor C more than 135 rpms (RPM) detect greatly
When the measured temperature of 65 degrees Celsius (C), failure code 1 will be triggered.When actuator B detects the electricity more than 1000 volts (V)
Pressure measured value and sensor C is when detecting the measured temperature less than 55 DEG C, will trigger failure code 2.When sensors A detects
Wheel measuring value, actuator B more than 100RPM detect the voltage measuring value more than 750V and sensor C is detected more than 60
DEG C measured temperature when, will trigger failure code 3.One of ordinary skill in the art are readily apparent that Fig. 3 offer is only
For example and task of explanation, and many other failure codes and/or triggering criterion are herein defined as possible and paid attention to.
Referring back to Fig. 2, CPU 206 also can be configured to be used to managing and/or controlling assets 200 to carry out
The various additional functionalities of operation.For example, CPU 206 can be configured to be carried to subsystem 202 and/or actuator 205
For command signal, the command signal causes subsystem 202 and/or actuator 205 to perform some operations (such as to change air throttle
Position).In addition, CPU 206 can be configured handles the number from sensor 204 and/or actuator 205 to change it
According to speed, or CPU 206 can be configured to provide command signal, institute to sensor 204 and/or actuator 205
Stating command signal causes sensor 204 and/or actuator 205 for example to change sampling rate.In addition, CPU 206 can
It is configured to from subsystem 202, sensor 204, actuator 205, network interface 210 and/or the reception signal of user interface 212,
And operation is caused based on such signal.In addition, CPU 206 can be configured by terms of such as diagnostic device
Device reception signal is calculated, the signal causes CPU 206 to be advised according to the diagnosis being stored in data storage device 208
Then perform one or more diagnostic tools.Other features of CPU 206 are discussed below.
Network interface 210 can be configured to provide assets 200 and be connected between the various networking components of communication network 106
Communication.For example, network interface 210 can be configured to promote whereabouts and radio communication from communication network 106, and therefore may be used
Take for launching and receiving the antenna structure of various wireless signals (over-the-air signal) and the shape of associated devices
Formula.Other examples are also possible.In fact, network interface 210 can be according to communication protocol (such as, but not limited to above-mentioned communication protocols
Any one of view) configure.
User interface 212 can be configured to promote interacting for user and assets 200, and also can be configured to promote to cause to provide
Production 200 performs operation in response to user mutual.The example of user interface 212 includes touch sensitive interface, mechanical interface (for example, thick stick
Bar, button, roller, dial, keyboard etc.) and other input interfaces (such as microphone) etc..In some cases, user
Interface 212 can include or provide the connectivity of output precision (such as display screen, loudspeaker, earphone jack etc.).
Local analytics device 220 is generally configured to receive and analyzes the data related to assets 200, and based on this point
Analysis can cause one or more operations occur at assets 200.For example, local analytics device 220 can receive the operand of assets 200
According to (for example, the data caused by sensor 204 and/or actuator 205) and based on this data can to CPU 206,
Sensor 204 and/or actuator 205 provide the instruction for causing assets 200 to perform operation.
In order to promote this operation, local analytics device 220 can include one or more asset interfaces, and it is configured to local
Analytical equipment 220 is coupled to one or more of mobile system of assets.For example, as shown in Figure 2, local analytics device 220
There can be the interface to the CPU 206 of assets, it may be such that local analytics device 220 can be from CPU
206 receive operation data (for example, being produced by sensor 204 and/or actuator 205 and being sent to the behaviour of CPU 206
Make data), and then provide instruction to CPU 206.In this way, local analytics device 220 can be via centre
Reason unit 206 interfaced with indirectly with other mobile systems (for example, sensor 204 and/or actuator 205) of assets 200 and from
Wherein receive data.Additionally or alternatively, as shown in Figure 2, local analytics device 220, which can have, arrives one or more sensors
204 and/or the interface of actuator 205, it may be such that local analytics device 220 can be with sensor 204 and/or actuator 205
Direct communication.Local analytics device 220 can also other manner and the mobile system of assets 200 interface with, comprising illustrated in fig. 2
The possibility that is promoted by one or more intermediate systems for not showing of interface.
In fact, local analytics device 220 may be such that assets 200 can be performed locally advanced analysis and associated behaviour
Make, such as perform prediction model and corresponding workflow, the operation may not use component on other assets to perform.Cause
And local analytics device 220 can help to provide extra disposal ability and/or intelligence to assets 200.
It should be appreciated that local analytics device 220 also can be configured to cause the execution of assets 200 unrelated with forecast model
Operation.For example, local analytics device 220 can receive data from remote source (such as analysis system 108 or output system 110),
And assets 200 are caused to perform one or more operations based on received data.One particular instance can relate to local analytics device
220 receive the firmware renewal of assets 200 from remote source, and then cause assets 200 to update its firmware.Another particular instance can relate to
And local analytics device 220 receives diagnostic instruction from remote source, and assets 200 are caused to perform sheet then according to the instruction received
Ground diagnostic tool.Several other examples are also possible.
As indicated, in addition to one or more asset interfaces discussed above, local analytics device 220 can also include processing
Unit 222, data storage device 224 and network interface 226, it all can pass through system bus, network or other connection machines
System communicatedly links.Processing unit 222 can include any one of component discussed above for CPU 206.Enter
And data storage device 224 can be or can include one or more non-transitory computer-readable storage mediums, it can use above
Any one of form of computer-readable storage medium of discussion.
Processing unit 222 can be configured to store, access and perform the computer being stored in data storage device 224 can
Reader instructs, to perform the operation of local analytics device described herein.For example, processing unit 222 can be configured to connect
Corresponding sensor and/or actuator signal as caused by sensor 204 and/or actuator 205 are received, and such signal can be based on
Perform prediction model-workflow pair.Other functions are described by below.
Network interface 226 can be same or like with above-mentioned network interface.In fact, network interface 226 can promote local point
Communication between analysis apparatus 220 and analysis system 108.
In some example implementations, the user that local analytics device 220 can include can be similar with user interface 212 connects
Mouthful and/or communicated with the user interface.In fact, user interface can be located remotely from local analytics device 220 (and assets 200)
Position.Other examples are also possible.
Physically and it is communicably coupled to via one or more asset interfaces although Fig. 2 illustrates local analytics device 220
Its associated assets (for example, assets 200), it should also be understood that, situation may not such was the case with.For example,
In some embodiments, local analytics device 220 can not be physically coupled to its associated assets, but can be located remotely from money
The position of production 220.In the example of this embodiment, local analytics device 220 can wirelessly be communicably coupled to assets 200.
Other arrangements and configuration are also possible.
One of ordinary skill in the art are readily apparent that the assets 200 shown in Fig. 2 are only that the simplified of assets represents
One example, and many other examples are also possible.For example, other assets can include the additional assemblies do not described and/or more
More or less components described.In addition, given assets can include coherency operation to perform the multiple of the operation of given assets
Individual asset.Other examples are also possible.
III. instance analysis system
Turning now to Fig. 4, the simplified block diagram of instance analysis system 400 is depicted.As indicated above, analysis system 400
It can include and communicatedly link and be arranged to carry out one or more computing systems of various operations described herein.It is specific next
Say, as indicated, analysis system 400 can include data shooting system 402, data science system 404 and one or more databases
406.These system components can and/or wired connection wireless via one or more be communicatively coupled, it is described connection can be configured with
Promote secure communication.
Data shooting system 402 is generally used for reception and processing data and outputs data to data science system 404.
Thus, data shooting system 402 can include one or more network interfaces, and it is configured to the various networks from network configuration 100
Component (such as assets 102 and 104, output system 110 and/or data source 112) receives data.Specifically, data intake system
System 402 can be configured to receive analog signal, data flow and/or network packet etc..Thus, network interface can include one or more
Individual wired network interface (such as port etc.) and/or radio network interface, similar to above-described radio network interface.One
In a little examples, data shooting system 402 can be or comprising the component configured according to data-oriented Flow Technique, such as NiFi receivers
Deng.
Data shooting system 402 can include one or more processing components for being configured to perform one or more operations.Example
Operation can include compression and/or decompression, encryption and/or decryption, analog/digital conversion and/or D/A switch, screening and amplification etc.
Other operations.In addition, data shooting system 402 can be configured with the data type based on data and/or data characteristic to parse,
Classification, tissue and/or route data.In some instances, data shooting system 402 can be configured with based on data science system
404 one or more characteristics or operating parameter are formatted, encapsulated and/or route data.
Generally, the data received by data shooting system 402 can take various forms.For example, the payload of data can
Include single sensor or actuator measured value, multiple sensors and/or actuator measured value and/or one or more unusual conditions
Data.Other examples are also possible.
In addition, received data can include some characteristics, such as source identifier and time stab (for example, obtaining information
Date and/or time).For example, can be by unique identifier (for example, letter, numeral, alphanumeric or class caused by computer
Like identifier) it is assigned to each assets, thereby increases and it is possible to it is assigned to each sensor and actuator.This class identifier is operable with identification
Assets, sensor or the actuator that data are derived from.In some cases, another characteristic can include the position for obtaining information
(for example, gps coordinate).Data characteristic can occur in the form of signal signature or metadata etc..
Data science system 404 is generally used for (for example, from data shooting system 402) and receives data and analyze data,
And one or more operations are caused to occur based on this analysis.Thus, data science system 404 can include one or more network interfaces
408th, processing unit 410 and data storage device 412, it all can be led to by system bus, network or other connection mechanisms
The link of letter ground.In some cases, data science system 404 can be configured to store and/or access promotion implementation and takes off herein
One or more application programming interfaces (API) of some in the feature shown.
Network interface 408 can be same or like with any of the above described network interface.In fact, network interface 408 can promote number
According to scientific system 404 and various other entities (such as data shooting system 402, database 406, assets 102, output system 110
Deng) between communication (for example, there is certain level of security).
Processing unit 410 can include one or more processors, and it can use any one of above-mentioned processor form.Enter
And data storage device 412 can be or can include one or more non-transitory computer-readable storage mediums, it can use above
Any one of form of computer-readable storage medium of discussion.Processing unit 410 can be configured to store, accesses and perform
The computer-readable program instructions being stored in data storage device 412, to perform the operation of analysis system described herein.
Generally, processing unit 410 can be configured to perform analysis to the data received from data shooting system 402.Therefore,
Processing unit 410 can be configured to perform one or more modules, and it can each take one be stored in data storage device 412
Or the form of multiple program instruction sets.Module can be configured to promote to cause result based on the execution that corresponding program instructs to be gone out
It is existing.Sample result from given module, which can include, outputs data to another module, the given module of renewal and/or another module
Programmed instruction, and output data to network interface 408 to be transmitted into assets and/or output system 110 etc..
Database 406 is generally used for (for example, from data science system 404) and receives data and data storage.Thus, often
One database 406 can include one or more non-momentary computer-readable storage mediums, such as any in examples provided above
Person.In fact, database 406 can be separated with data storage device 412 or integrated with data storage device 412.
Database 406 can be configured to store the data of a few types, and some of which is discussed below.In fact, deposit
Some in the data of storage in database 406 can include time stab, and it indicates that data are generated or be added to database
Date and time.In addition, data can be stored in database 406 with several means.In addition to other examples, for example, data can press
According to time sequencing, forms mode storage, and/or based on data source types (for example, based on assets, Asset Type, sensor, biography
Sensor type, actuator or actuator types) or unusual condition designator carry out tissue.
IV. example operation
The operation for the example network configuration 100 described in Fig. 1 will be discussed in further detail below now.In order to help to retouch
Some in this generic operation are stated, refer to flow chart to describe the combination of executable operation.In some cases, each frame can
The module of representation program code or part, it is included can be by computing device with the specific logical function or step in implementation process
Instruction.Program code is storable on any kind of computer-readable media, such as non-transitory computer-readable media
On.In other cases, each frame can represent the circuit of the routed specific logical function with implementation procedure or step.Separately
Outside, the frame shown in flow chart can be rearranged into different order, be combined into less frame, be divided into extra frame, and/or be based on
Specific embodiment and remove.
Description refers to individual data source wherein such as assets 102 to point for then performing one or more functions below
Analysis system 108 provides the example of data.It should be appreciated that what this was carried out just for the sake of clear and explanation, and simultaneously unexpectedly
Taste limitation.In fact, analysis system 108 generally simultaneously from multiple sources receive data, and based on this polymerization reception data come
Perform operation.
A. the collection of operation data
As mentioned above, typical asset 102 can take various forms and can be configured to perform multiple operations.In non-limit
In property example processed, in the form of assets 102 can take the operable locomotive by U.S. various regions transshipment cargo.When transporting, assets
102 sensor and/or actuator can obtain the data of one or more operating conditions of reflection assets 102.Sensor and/or cause
Dynamic device can launch data to the processing unit of assets 102.
Processing unit can be configured to receive data from sensor and/or actuator.In fact, processing unit can simultaneously or
The sensing data from multiple sensors and/or the actuator data from multiple actuators are received in order.Such as institute above
Discuss, when receiving this data, processing unit also can be configured to determine whether data meet that triggering any unusual condition refers to
Show the triggering criterion of symbol (such as failure code).Determine to trigger the situation of one or more abnormality designators in processing unit
Under, processing unit can be configured to perform one or more local operations, such as export the designator that is triggered via user interface
Instruction.
Assets 102 can then launch via the network interface and communication network 106 of assets 102 to analysis system 108 to be operated
Data.In operation, assets 102 are continuously, periodically and/or in response to trigger event (for example, unusual condition) to behaviour
The system that performs an analysis 108 launches operation data.Specifically, assets 102 can be based on specific frequency (for example, daily, per hour, often
15 minutes, it is per minute once, once per second etc.) periodically launch operation data, or assets 102 can be configured to launch behaviour
Make the continuous feeding in real time of data.Additionally or alternatively, such as when sensor and/or actuator measured value meet for any
During the triggering criterion of unusual condition designator, assets 102 can be configured to launch operation data based on some triggerings.Assets
102 can also other manner transmitting operation data.
In fact, the operation data of assets 102 can include sensing data, actuator data and/or unusual condition data.
In some embodiments, assets 102 can be configured to provide operation data in individual traffic, and in other embodiments
In, assets 102 can be configured to provide operation data in multiple different data flows.For example, assets 102 can be to analysis system
108 provide sensor and/or the first data flow of actuator data and the second data flow of unusual condition data.Other possibilities
There is also.
Sensor and actuator data may take various forms.For example, sometimes, sensing data (or actuator data)
The measured value obtained by each of sensor (or actuator) of assets 102 can be included.And at other times, sensor number
The measured value obtained by the subset of the sensor (or actuator) of assets 102 can be included according to (or actuator data).
Specifically, sensor and/or actuator data can be included by with giving the unusual condition designator phase being triggered
The measured value that the sensor and/or actuator of association obtain.For example, if the failure code being triggered is the failure from Fig. 3
Code 1, then sensing data can include the original measurement value obtained by sensors A and C.Additionally or alternatively, data can wrap
The measured value obtained containing one or more sensors or actuator by being not directly relevant to connection with the failure code being triggered.Continue most
Example afterwards, data can additionally comprise the measured value by actuator B and/or the acquisition of other sensors or actuator.In some realities
In example, assets 102 can be included specific based on the failure code rule provided by analysis system 108 or instruction in operation data
Sensing data, it for example can determine that the measured value just measured in actuator B with originally causing what failure code 1 was triggered
Correlation between measured value be present.Other examples are also possible.
In addition, data can include based on the specific time of interest from each sensor and/or actuator of interest
One or more sensors and/or actuator measured value, the time can be selected based on several factors.In some instances,
Special time of interest can be based on sampling rate.In other examples, special time of interest can be based on triggering abnormal shape
The time of condition designator.
In particular, the time based on triggering unusual condition designator, data can be included from each sensing of interest
Device and/or actuator (for example, with the designator that is triggered directly and indirect correlation connection sensor and/or actuator) one or
Multiple corresponding sensors and/actuator measured value.One or more measured values can be based on the unusual condition designator in triggering
The specific times or specific duration of measurement around time.
For example, if the failure code of triggering is the failure code 2 from Fig. 3, then sensor of interest and actuating
Device can include actuator B and sensor C.One or more measured values can be included in triggering failure code and it (measured) for example, triggering
The preceding nearest respective measurement values obtained by actuator B and sensor C or before triggering measures, afterwards or neighbouring corresponding survey
Value set.For example, five times before or after one group of five measurement can include triggering measurement measure (for example, not comprising triggering
Measurement), triggering measurement before or after four times measurement and triggering measurement, or triggering measurement before measurement twice and triggering survey
Measurement twice and triggering measurement and other possibilities after amount.
Similar with sensor and actuator data, unusual condition data can take various forms.Generally, abnormal conditions data
It can include or take the form of designator, the designator is operable with all other different from what may be occurred at assets 102
The specific exceptions situation occurred at assets 102 is uniquely identified in normal situation.Unusual condition designator can take letter, numeral
Or the form of alpha numeric identifier etc..In addition, unusual condition designator can take the form of the word string of description unusual condition,
Such as " engine overheat " or " fuel shortage " etc..
Analysis system 108, and the data shooting system of particularly analysis system 108 can be configured with from one or more assets
And/or data sources operation data.Data shooting system can be configured to perform one or more operations to received data,
And data are then relayed to the data science system of analysis system 108.And then data science system can analyze received number
One or more operations are performed according to and based on this analysis.
B. forecast model and workflow are defined
As an example, analysis system 108 can be configured with the operation data received based on one or more assets
And/or the external data that is received related to one or more assets defines forecast model and corresponding workflow.Analysis system
108 may be based on various other data definition model-workflows pair.
Generally, model-workflow pair can include program instruction set, and it causes assets to monitor some operating conditions and carries out certain
A little operations, the operation help to promote to prevent to monitor the particular event indicated by operating conditions.Specifically, forecast model
One or more algorithms can be included, the input of the algorithm is the sensing of one or more sensors and/or actuator from assets
Device and/or actuator data, and it is exported for determining that particular event may occur at assets in following special time period
Probability.And then the correspondence that workflow can include one or more triggerings (for example, model output valve) and assets are carried out based on triggering
Operation.
As indicated above, analysis system 108 can be configured to define polymerization and/or personalized forecast model and/or work
Flow." polymerization " model/workflow may refer to such as drag/workflow:It is general to assets group and do not considering to be deployed with mould
Model/the workflow being defined in the case of the particular characteristics of the assets of type/workflow.On the other hand, " personalization " model/
Workflow may refer to such as drag/workflow:For single assets or the assets subgroup from assets group specifically customization and
Model/workflow that particular characteristics based on single assets or the assets subgroup for being deployed with model/workflow are defined.Under
Text is discussed in further detail these different types of model/workflows and performed by analysis system 108 with Definition Model/workflow
Operation.
1. polymerization model and workflow
In example implementation, analysis system 108 can be configured poly- to be defined based on the aggregated data of multiple assets
Matched moulds type-workflow pair.Defining polymerization model-workflow pair can perform in a variety of ways.
Fig. 5 is the flow chart 500 for a possible example for describing the definition phase available for Definition Model-workflow pair.
For purpose of explanation, the example definition stage is described as being carried out by analysis system 108, but this definition phase also can be by other systems
System is carried out.One of ordinary skill in the art are readily apparent that, flow chart 500 is provided for clear and task of explanation, and available
Several other combinations of operation carry out Definition Model-workflow pair.
As shown in Figure 5, at frame 502, analysis system 108 can begin at definition and form the basic of given prediction model
Data acquisition system (for example, data of interest).Data of interest may originate from several sources, such as assets 102 and 104 and data
Source 112, and can be stored in the database of analysis system 108.
Data of interest can include the special assets set or all in assets group in assets group
The historical data of assets (for example, assets of interest).In addition, data of interest can include it is every in assets of interest
The particular sensor and/or collective actuators of one or all the sensors from each of assets of interest and/or
The measured value of actuator.In addition, data of interest can include the data from past special time period, such as the history of two weeks
Data.
Data of interest can include various types of data, and this is likely to be dependent on given forecast model.In some feelings
Under condition, data of interest can comprise at least the operation data of the operating conditions of instruction assets, and wherein operation data as existed above
Discussed in the collection paragraph of operation data.In addition, data of interest can include the ring that instruction assets are generally operated wherein
The environmental data in border and/or instruction during it assets by the date of enterprise for carrying out some tasks and the scheduling data of time.Its
The data of its type can be also comprised in data of interest.
In fact, data of interest can define in many ways.In an example, data of interest can be to use
What family defined.In particular, the operable output system 110 of user, it receives the use for the selection for indicating some data of interest
Family inputs, and output system 110 can provide the data for indicating such selection to analysis system 108.Based on received data, divide
The then definable data of interest of analysis system 108.
In another example, data of interest can be machine definitions.In particular, analysis system 108 is executable each
Kind operation, such as simulate, to determine to produce the data of interest of most accurate forecast model.Other examples are also possible.
Fig. 5 is returned to, at frame 504, analysis system 108 can be configured to be grasped based on data definition of interest and assets
Make related polymerization forecast model.Generally, polymerization, the operating conditions of forecast model definable assets at assets with occurring event
Possibility between relation.Specifically, polymerization, forecast model can receive the sensing data of the sensor from assets
And/or the actuator data of the actuator from assets is used as input, and output will be at assets in some following time quantum
The probability of generation event.
The event of forecast model prediction may depend on particular implementation situation and different.For example, event is probably failure,
And therefore, forecast model is probably that the fault model whether prediction will break down within some following period (hereafter exists
Fault model is discussed in detail in health status Fraction Model and workflow paragraph).In another example, event can be that assets are completed
Task, and therefore, assets can be predicted by the possibility for the task that is timely completed in forecast model.In other examples, event can be stream
Body or unit replacement, and therefore, forecast model can be predicted special assets fluid or component need to be replaced before time quantum.
Again in other examples, event can be the change of the productivity of assets, and therefore, it is domestic-investment that following special time period can be predicted in forecast model
The productivity ratio of production.In another example, event can be the generation of " leading designator " event, and the event may indicate that to be provided with expected
The different assets behavior of production behavior, and therefore, forecast model can be predicted in following one or more leading designator events of generation
Possibility.Other examples of forecast model are also possible.
Anyway, analysis system 108 can define polymerization forecast model in a variety of ways.Generally, this operation can relate to profit
The model of probability of the return between 0 and 1, such as random forest technology, logic are produced with one or more modeling techniques
Regression technique or other regression techniques and other modeling techniques.In particular instance embodiment, analysis system 108 can basis
Polymerization forecast model is defined below with reference to Fig. 7 discussion.Analysis system 108 can also other manner define polymerization model.
At frame 506, analysis system 108 can be configured to define the polymerization work corresponding to the Definition Model from frame 504
Flow.Generally, workflow can be taken based on the form of the action of the specific output implementation of forecast model.In example implementation
In, workflow can include one or more operations that output of the assets based on defined forecast model performs.It can be workflow
The example of partial operation includes assets according to specific data acquisition scheme gathered data, according to specific data launch scenario to point
Analysis system 108 launches data, performs local diagnostic tool and/or changes the workflow operations such as the operating conditions of assets.
Specific data acquisition plan may indicate that assets how gathered data.In particular, data acquisition plan may indicate that
Assets obtain some sensors and/or actuator of data therefrom, for example, assets multiple sensors and actuator (for example,
Sensor/actuators of interest) in sensor and/or actuator subset.In addition, data acquisition plan may indicate that assets
The amount and/or assets of the data obtained from sensor/actuators of interest gather the sample frequency of such data.Data acquisition
Scheme can also include various other attributes.In specific example implementation, specific data acquisition scheme may correspond to assets
The forecast model of health status, and assets health status can be based on reduce and adjust with (for example, from particular sensor) collection more
More data and/or specific data.Or specific data acquisition plan may correspond to leading designator forecast model, and can be based on
The possibility increase of leading designator event occurs and is adjusted to the modification number by asset sensor and/or actuator collection
According to the leading designator event can be shown that subsystem may break down.
Specific data emission case may indicate that how assets to analysis system 108 launch data.Specifically, data are sent out
The scheme of penetrating may indicate that the data type (and may further indicate that the form and/or structure of data) that assets should launch, such as from certain
Multiple data samples that data, the assets of a little sensors or actuator should launch, tranmitting frequency, and/or assets should be in its number
According to the precedence scheme of the data included in transmitting.In some cases, specific data acquisition plan can include data and launch
Scheme or data acquisition plan can match with data emission case.In some example implementations, specific data emission
Case may correspond to the forecast model of assets health status, and less frequency can be adjusted to based on the assets health status higher than threshold value
Launch data numerously.Other examples are also possible.
As indicated above, local diagnostic tool can be to be locally stored in collection of programs at assets etc..Local diagnosis work
Tool can generally promote to diagnose the reason for mistake or failure at assets.In some cases, when implemented, local diagnostic tool can
Input will be tested to be delivered in subsystem or part thereof of assets to obtain test result, this can promote diagnostic error or failure
Reason.These local diagnostic tool dormancy generally in assets, and unless otherwise asset acceptance will not to specific diagnostic instruction
It can be performed.Other local diagnostic tools are also possible.In an example implementation, specific local diagnostic tool can
Corresponding to the forecast model of the health status of the subsystem of assets, and can the subsystem health status based on equal to or less than threshold value
To perform.
Finally, workflow can relate to change the operating conditions of assets.For example, one or more actuators of controllable assets with
Promote the operating conditions of modification assets.Can change various operating conditions, for example, speed, temperature, pressure, liquid level, current drain and
Power distribution etc..In specific example implementation, operating conditions modification stream may correspond to be used to predict that assets are
The no forecast model by the task that is timely completed, and can be based on causing assets to improve its row less than the prediction Percent Complete of threshold value
Enter speed.
Anyway, overall work stream can define in a variety of ways.In an example, it can be user to polymerize workflow
Definition.Specifically, the computing device of the operable user's input for receiving the selection for indicating some workflow operations of user, and
Computing device can provide the data for indicating such selection to analysis system 108.Based on this data, analysis system 108 can then determine
Justice polymerization workflow.
In another example, it can be machine definitions to polymerize workflow.In particular, analysis system 108 can perform various
Operate the reason for (such as simulation) is to determine the probability that can promote to determine to be exported by forecast model and/or prevent by model prediction
The workflow of the generation of event.Other examples of definition polymerization workflow are also possible.
In workflow of the definition corresponding to forecast model, the triggering of the definable workflow of analysis system 108.It is real in example
Apply in scheme, workflow triggering can be the scope by the value of the probability of forecast model output or the value by forecast model output.
Under certain situation, workflow can have multiple triggerings, and each of which triggering can cause different operations or multiple operations occurs.
In order to illustrate, Fig. 6 A be polymerization model-workflow to 600 conceptual illustration.As indicated, polymerization model-workflow
Explanation 600 is included and calculates 604, model output area 606 for mode input 602, model and corresponds to workflow operation 608
Row.In this example, forecast model has the single input data from sensors A, and has two calculated values:Calculated value I
And II.The output of this forecast model influences performed workflow operation.If output probability is less than or equal to 80%, then holds
Row workflow operation 1.Otherwise, workflow operation 2 is performed.Other instance model-workflows are to being herein defined as possible and giving
To consider.
2. personalized model and workflow
In another aspect, analysis system 108 can be configured to define the personalized forecast model of assets and/or work
Stream, this can relate to by the use of polymerization model-workflow to being used as benchmark.Personalization can some characteristics based on assets.In this way,
Analysis system 108 can be provided with polymerization model-workflow model-workflow more accurate and sane to compared with given assets
It is right.
In particular, Fig. 5 is returned to, at frame 508, analysis system 108 can be configured to decide whether personalization in frame
The polymerization model defined at 504 for given assets (for example, assets 102).Analysis system 108 can carry out this in many ways and determine
It is fixed.
In some cases, analysis system 108 can be configured acquiescently to define the forecast model of personalization.In other feelings
Under condition, analysis system 108 can be configured to decide whether to define personalized forecast model based on some characteristics of assets 102.
For example, in some cases, only some types or classification or operate in specific environment or commented with some health status
The assets divided may receive the forecast model of personalization.In the case of again other, whether user's definable is fixed for assets 102
Adopted personalized model.Other examples are also possible.
Anyway, if analysis system 108 determines to define the personalized forecast model of assets 102, then analysis system
108 can so do at frame 510.Otherwise, analysis system 108 may proceed to frame 512.
At frame 510, analysis system 108 can be configured to define personalized forecast model in various ways.In example
In embodiment, one or more characteristics that analysis system 108 can be based at least partially on assets 102 predict mould to define personalization
Type.
Before the personalized forecast model of assets 102 is defined, analysis system 108 may have determined that to form personalization
One or more basic of model asset character of interest.In fact, different forecast models may have different correspondences
Characteristic of interest.
Generally, characteristic of interest can be to related characteristic to polymerization model-workflow.For example, characteristic of interest
It can be the characteristic for the accuracy that analysis system 108 has determined influence polymerization model-workflow pair.The example of this class feature can wrap
The time limit containing assets, assets behaviour in service, assets ability, assets load, assets health status (may be good for by the assets being discussed below
Health status index indicates), class of assets (for example, brand and/or model) and the characteristic such as environment for operating assets.
Analysis system 108 may determine characteristic of interest in many ways.In an example, analysis system
108 can be completed by performing one or more modeling Simulations for promoting to identify characteristic of interest.In another example, it is of interest
Characteristic may predefine and be stored in the data storage device of analysis system 108.It is of interest in another example again
Characteristic may be defined by the user and be provided to analysis system 108 via output system 110.Other examples are also possible
's.
Anyway, it is determined that after characteristic of interest, analysis system 108 can determine that and identified spy of interest
The characteristic of assets 102 corresponding to property.That is, analysis system 108 can determine that the characteristic of assets 102 corresponding with characteristic of interest
Type, value, its presence or shortage etc..Analysis system 108 can perform this operation in many ways.
For example, analysis system 108 can be configured to perform this based on the data from assets 102 and/or data source 112
Operation.In particular, the operation data of the available assets 102 of analysis system 108 and/or the external data from data source 112
To determine one or more characteristics of assets 102.Other examples are also possible.
Based on one or more characteristics of identified assets 102, analysis system 108 can be defined by changing polymerization model
Personalized forecast model.Polymerization model can modify in many ways.For example, can by change (for example, increase, remove, again
New sort etc.) one or more mode inputs, change one or more sensors corresponding with the Asset operation limit and/or actuator survey
Amount scope (for example, changing operating limit corresponding with " leading designator " event), change one or more model calculation values, to meter
The variable of calculation or output weighting (or changing its weight), utilize the modeling skill different from the modeling technique for defining polymerization model
Art and/or response variable different from the response variable for defining polymerization model etc. is utilized to change polymerization model.
In order to illustrate, Fig. 6 B be personalized model-workflow to 610 conceptual illustration.- specifically, personalized model-
Workflow is to illustrating that 610 be the revision of polymerization model-workflow pair from Fig. 6 A.As indicated, personalized model-work
Flow to include explanation 610 and arranged for the modification of mode input 612 and model calculating 614, and include the model output from Fig. 6 A
The original row of scope 606 and workflow operation 608.In this example, personalized model has two inputs, from sensors A and
Actuator B data, and there are two calculated values:Calculated value II and III.Output area and corresponding workflow operation are with Fig. 6 A's
Output area and corresponding workflow operation are identical.Analysis system 108 may have been based on determining assets 102 it is for example relatively old and
The reasons such as health status is relatively poor define personalized model in this way.
In fact, personalized polymerization model may depend on one or more characteristics of given assets.In particular, some characteristics
The modification of polymerization model may be influenceed in a manner of different from other characteristics.In addition, the type of characteristic, being worth, existing
It can influence to change.For example, the assets time limit may influence the Part I of polymerization model, and class of assets may influence to polymerize
Second different piece of model.And the assets time limit in the range of first time limit may influence the of polymerization model in the first way
A part, and the assets time limit in the range of second time limit different from the first scope may influence polymerization model in a second manner
Part I.Other examples are also possible.
In some embodiments, personalized polymerization model may depend on as the supplement of asset character or examining for replacement
Consider.For example, when known assets are in relatively good mode of operation (for example, as defined in machinist etc.), money can be based on
The sensor and/or actuator reading of production carry out personalized to polymerization model.More particularly, in leading designator forecast model
Example in, analysis system 108 can be configured to receive assets and be in the instruction of excellent operation state (for example, coming from machinist
The computing device of operation) and data of the operation from assets.Operation data is at least based on, analysis system 108 can then pass through
Modification carrys out the leading designator forecast model of personalised assets with the corresponding corresponding operating limit of " leading designator " event.Its
Its example is also possible.
Fig. 5 is returned to, at frame 512, analysis system 108 also can be configured to decide whether the work of personalised assets 102
Flow.Analysis system 108 can carry out this decision in many ways.In some embodiments, analysis system 108 can be according to frame
508 perform this operation.In other embodiments, analysis system 108 can be decided whether fixed based on personalized forecast model
Adopted individualized work stream.In another embodiment again, if defining personalized forecast model, then analysis system 108 can
Determine to define individualized work stream.Other examples are also possible.
Anyway, if analysis system 108 determines to define the individualized work stream of assets 102, then analysis system 108
Can so it be done at frame 514.Otherwise, analysis system 108 can terminate the definition phase.
At frame 514, analysis system 108 can be configured to define individualized work stream in various ways.It is real in example
Apply in scheme, analysis system 108 can be based at least partially on one or more characteristics of assets 102 to define individualized work stream.
Before the individualized work stream of assets 102 is defined, similar to the personalized forecast model of definition, analysis system 108
It may have determined that to form one or more basic of individualized work stream asset character of interest, it may basis
The discussion of frame 510 is determined.Generally, these characteristics of interest can be the characteristic for the effect for influenceing polymerization workflow.This
Class feature can include any one of examples discussed above characteristic.Other characteristics are also possible.
Be similar to frame 510 again, analysis system 108 can determine that with individualized work stream determined by characteristic of interest
The characteristic of corresponding assets 102.In example implementation, characteristic that analytic approach system 108 can be discussed with reference block 510
Similar mode is determined to determine the characteristic of assets 102, and actually using some or all of described determination.
Anyway, one or more characteristics based on determined by assets 102, analysis system 108 can be polymerize by changing
Workflow carrys out the workflow of personalised assets 102.Polymerization workflow can modify in many ways.For example, change can be passed through
(for example, increase, removal, rearrangement, replacement etc.) one or more workflow operations from the first data acquisition plan (for example, change
It is changed into alternative plan or changes into specific local diagnostic tool from specific data acquisition scheme) and/or change (for example, increase,
Reduce, increase, removing etc.) corresponding model output valve or value scope of particular workflow operation etc. is triggered to change polymerization work
Stream.In fact, the mode that the modification to polymerization workflow can be similar to the modification to polymerization model depends on the one of assets 102
Or multiple characteristics.
In order to illustrate, Fig. 6 C be personalized model-workflow to 620 conceptual illustration.Specifically, personalized model-
Workflow is to illustrating that 620 be the revision of polymerization model-workflow pair from Fig. 6 A.As indicated, personalized model-work
Flow the original row that 604 and model output area 606 are calculated explanation 620 comprising the mode input 602 from Fig. 6 A, model, but
It is to be arranged comprising the modification for workflow operation 628.In this example, personalized model-workflow pair and polymerizeing in Fig. 6 A
Model-workflow is to similar, except when when the output of polymerization model is more than 80%, workflow operation 3 is triggered rather than operated 1
Outside.In addition to other reasons, analysis system 108 can for example increase the generation of asset failures in history based on assets 102 are determined
Environment in operation define this individual work stream.
After individualized work stream is defined, analysis system 108 can terminate the definition phase.At that time, analysis system 108 can connect
With personalized model-workflow pair for assets 102.
In some example implementations, analysis system 108 can be configured is used for the personalization for giving assets in advance to define
Model and/or corresponding workflow are surveyed, without defining polymerization forecast model and/or corresponding workflow first.Other examples are also can
Can.
Although discussed above is 108 personalized forecast model of analysis system and/or workflow, other devices and/or
The executable personalization of system.For example, the local analytics device personalizable forecast model and/or workflow of assets 102, or can be with
Analysis system 108 works to perform this generic operation together.Discuss that local analytics device performs this generic operation in further detail below.
3. health status Fraction Model and workflow
In specific embodiments, as mentioned above, analysis system 108 can be configured to define the health status with assets
Associated forecast model and corresponding workflow.In example implementation, for monitor assets health status it is one or more
Individual forecast model can be used for the health indicator (for example, " health status fraction ") of output assets, the health indicator
It is single polymerization index, whether it is indicated will be in given money (for example, ensuing two weeks) in the range of the preset time in future
Broken down at production.In particular, health indicator may indicate that will not be at assets in the range of the preset time in future
The possibility of any failure to break down in group, or may indicate that will in the range of the preset time in future for health indicator
The possibility of at least one failure to be broken down at assets in group.
In fact, it can be defined according to the above discussion for the forecast model and corresponding workflow for exporting health indicator
For polymerization or personalized model and/or workflow.
In addition, the expectation granularity depending on health indicator, analysis system 108 can be configured to define the different water of output
The different forecast models of flat health indicator and define different corresponding workflows.For example, the definable of analysis system 108
Export the forecast model of the health indicator (that is, Assets Levels health indicator) of whole assets.As another example, divide
Corresponding health indicator (that is, the subsystem irrespective of size health status of one or more subsystems of the definable of analysis system 108 output assets
Index) corresponding forecast model.In some cases, the output of each subsystem irrespective of size forecast model may be combined to generate assets
Level health status index.Other examples are also possible.
Generally, the forecast model of definition output health indicator can perform in a variety of ways.Fig. 7 is to describe to can be used for determining
The flow chart 700 of the possible example of the modelling phase of the model of justice output health indicator.For purpose of explanation, example
Modelling phase is described as being carried out by analysis system 108, but this modelling phase can also be carried out by other systems.Art
Those skilled in the art are readily apparent that, flow chart 700 is provided for clear and task of explanation, and using several other combinations of operation
To determine health indicator.
As shown in Figure 7, at frame 702, analysis system 108 can form basic one of health indicator by definition
Or multiple failures (that is, failure of interest) set and start.In fact, one or more failures can be in the case of generation
The inoperable failure of assets (or its subsystem) may be made.Based on defined failure collection, analysis system 108 can take step
It is rapid to define for predicting what is any one of broken down in the range of the preset time in future (for example, ensuing two weeks)
The model of possibility.
In particular, at frame 704, analysis system 108 can analyze the historical operating data of one or more assets groups, with
Identify that the past of given failure occurs from the failure collection.At frame 706, analysis system 108 can recognize that and given failure
Past of each identification associated corresponding operating data acquisition system occurs (for example, before given failure occurs, from given
The sensor and/or actuator data of time range).At frame 708, analysis system 108 can analyze the past with given failure
The associated operation data set identified of generation, to define (1) for the value of given operation index set and (2) not
Relation (for example, fault model) between the possibility that (for example, following two weeks) breaks down in the range of the preset time come.Most
Afterwards, at frame 710, by each failure in institute's definition set define relation (for example, individual failure model) then can group
Synthesize the model of the overall possibility occurred for predicting failure.
When analysis system 108 continues to the operation data of the renewal of one or more assets groups, analysis system 108 is also
It can be improved by the operation data repeat step 704 to 710 to renewal come institute's definition set continuing with one or more failures
Forecast model.
The function of example modelling phase illustrated in fig. 7 will be described in further detail now.Since frame 702, such as
Upper described, analysis system 108 can be started by defining the set for one or more the basic failures for forming health indicator.Point
Analysis system 108 can perform this function in a variety of ways.
In an example, the set of one or more failures can be based on one or more users input.Specifically, analysis system
System 108 can receive the user's selection for indicating one or more failures from the computing system (such as output system 110) operated by user
Input data.Thus, the set of one or more failures can be user-defined.
In other examples, the set of one or more failures can be based on the determination made by analysis system 108 (for example, machine
What device defined).In particular, analysis system 108 can be configured one or more failures that may occur in many ways to define
Set.
For example, analysis system 108 can be configured with based on one or more characteristics of assets 102 come failure definition set.That is,
Some failures likely correspond to some characteristics of assets, such as Asset Type, classification etc..For example, each type and/or classification
Assets may have corresponding failure of interest.
In another example, analysis system 108 can be configured with based on being stored in going through in the database of analysis system 108
History data and/or the external data that is provided by data source 112 are come failure definition set.For example, analysis system 108 can utilize this number
According to determining which failure result in most long maintenance time and/or additional faults etc. followed by occurs in which failure in history
Deng.
In other examples again, one can be defined based on user's input and the determined combination made by analysis system 108
Or the set of multiple failures.Other examples are also possible.
At frame 704, for each of failure in failure collection, analysis system 108 can analyze one or more moneys
The historical operating data (for example, abnormal behaviour data) for producing group was occurred with the past of the given failure of identification.One or more assets
Group can include single assets (such as assets 102), or include the multiple assets of same or like type, such as include assets
The troop of 102 and 104 asset groups.Analysis system 108 can analyze certain amount of historical operating data, for example, it is a certain amount of when
Between the data (for example, one month data value) that are worth or a number of data point (for example, 1,000 nearest data
Point) etc..
In fact, the past of the given failure of identification can relate to the operation of the given failure of the identification instruction of analysis system 108
The type of data (such as unusual condition data).Generally, given failure can be with one or more unusual condition designators (such as failure
Code) it is associated.That is, when given failure occurs, one or more abnormality designators may be triggered.Thus, unusual condition
Designator can reflect the potential sign of given failure.
After the type that the operation data of given failure is indicated in identification, analysis system 108 can identify in many ways to
The past for determining failure occurs.For example, analysis system 108 can be according to the historical operation being stored in the database of analysis system 108
Data are come the corresponding unusual condition data of the unusual condition designator that positions be associated with given failure.Each exception after positioning
Status data will indicate that given failure occurs.Based on this unusual condition data after positioning, analysis system 108 is recognizable to be occurred
The time of past failure.
At frame 706, the recognizable operation associated with the past generation of the given failure of each identification of analysis system 108
The corresponding set of data.In particular, around the time of the given generation of the recognizable next comfortable given failure of analysis system 108
The set of sensor and/or actuator data in some time range.For example, the data acquisition system may be from given event
Barrier occur before, afterwards or surrounding particular time range (for example, two weeks).In other cases, can be from given failure
Before generation, afterwards or a number of data point of surrounding identifies the data acquisition system.
In example implementation, the operation data set can include some or all sensors from assets 102
And the sensor and/or actuator data of actuator.For example, the operation data set can include come from corresponding to given therefore
The data of the associated sensor of the unusual condition designator of barrier and/or actuator.
In order to illustrate, Fig. 8 depicts analysis system 108 and can analyzed to promote the concept of the historical operating data of Definition Model
Explanation.Curve map 800 may correspond to some in sensor and actuator from assets 102 (for example, sensors A and actuating
Device B) or the owner historical data fragment.As indicated, curve map 800 includes the time in x-axis 802, the measurement in y-axis 804
Value and sensing data 806 corresponding with sensors A and actuator data 808 corresponding with actuator B, it is every in the data
One, which includes, represents particular point in time TiThe various data points of the measured value at place.In addition, curve map 800 is included in time in the past Tf
The instruction of the generation for the failure 810 that (for example, " fault time ") occurs and the instruction Δ T of the time quantum 812 before breaking down,
From the instruction identification operation data set.Thus, Tf- Δ T defines the time range 814 of data point of interest.
Fig. 7 is returned to, identifies the given generation of given failure (for example, T in analysis system 108fThe generation at place) operation
After data acquisition system, analysis system 108 can determine whether that its any remaining generation for identifying operation data set should be directed to.
, will be to each remaining generation repeat block 706 in the case where remaining occur be present.
Hereafter, at frame 708, analysis system 108 can analyze associated with the generation of the past of given failure identified
Operation data set, to define the given set of (1) operation index (for example, the given collection of sensor and/or actuator measured value
Close) relation between (for example, future two weeks) will break down in the range of the preset time in future with (2) possibility (for example,
Fault model).That is, given fault model can be by sensor and/or actuator from one or more sensors and/or actuator
Measured value exports as input and the probability of given failure will occur in the range of the preset time in future.
Generally, the relation between the operating conditions of fault model definable assets 102 and the possibility to break down.One
In a little embodiments, in addition to sensor and/or the original data signal of actuator from assets 102, fault model can
Receive a number of other data inputs, also referred to as feature derived from sensor and/or actuator signal.This category feature can include
The average value or scope of the value measured in history when breaking down, the value gradient measured in history before breaking down it is flat
Duration between average or scope (for example, rate of change of measured value), feature (for example, breaking down and for the first time
It is secondary break down between time quantum or data count out) and/or instruction break down the sensor and/or actuator of surrounding
One or more fault modes of measured value trend.One of ordinary skill in the art are readily apparent that these are only can be from sensing
Several example aspects derived from device and/or actuator signal, and many other features are possible.
In fact, fault model can define in many ways.In example implementation, analysis system 108 can pass through
Using one or more modeling techniques come failure definition model, the modeling technique returns to the probability between 0 and 1, described to build
Mould technology can use the form of any of the above described modeling technique.
In particular instances, failure definition model can relate to analysis system 108 based on the historical operation number identified in frame 706
According to producing response variable.Specifically, analysis system 108 can determine that the sensor received in particular point in time and/or cause
The associated response variable of each set of dynamic device measured value.Thus, response variable can take the number associated with fault model
According to the form of set.
Whether response variable may indicate that given set of measurements in any one of the time range determined at the frame 706.
That is, response variable can reflect data-oriented set whether the time of interest from the surrounding that breaks down.Response variable can be two
Hex value response variable so that if given set of measurements is in any one of identified time range, then related
Connection response variable is assigned value 1, is otherwise associated response variable and is assigned value 0.
Fig. 8 is returned to, response variable vector Y is illustrated on curve map 800resConceptual illustration.As indicated, the and time
The response variable that set of measurements in scope 814 is associated has value 1 (for example, in time Ti+3To Ti+8The Y at placeres), and with
The response variable that set of measurements outside time range 814 is associated has value 0 (for example, in time TiTo Ti+2And Ti+9Arrive
Ti+10The Y at placeres).Other response variables are also possible.
Continue in the particular instance based on response variable failure definition model, analysis system 108 is available to be known in frame 706
Other historical operating data and caused response variable train fault model.Based on this training process, analysis system 108 connects
Definable fault model, it receives various sensors and/or actuator data as input, and export with for producing sound
The probability between 0 and 1 that will be broken down in the dependent variable equivalent period.
In some cases, the historical operating data identified at frame 706 and the training of caused response variable are utilized
The variable importance statistic for each sensor and/or actuator can be caused.Variable importance statistic is given to may indicate that
Sensor or actuator within the period in future to will occur the relative effect of the probability of given failure.
Additionally or alternatively, analysis system 108 can be configured with based on one or more survival analysis technologies (such as Cox
Proportional hazards technology) carry out failure definition model.Analysis system 108 can be similar to modeling technique discussed above in some aspects
Mode utilize survival analysis technology, but analysis system 108 can determine that life span response variable, its indicated from last time therefore
Hinder the time quantum of next expected event.Next expected event is probably the measured value for receiving sensor and/or actuator
Or the generation of failure, it is defined by first sending out survivor.This response variable can be comprising each in the particular point in time with receiving measured value
A pair of values that person is associated.Then, can determine to break down in the range of the preset time in future using response variable
Probability.
In some example implementations, may be based partly on the external data such as weather data and " hot tank " data with
And other data definition fault models.For example, being based on this data, fault model can increase or reduce output probability of malfunction.
In fact, it can be seen at the time point for the Time Inconsistency for obtaining measured value with asset sensor and/or actuator
Examine external data.For example, the time of " hot tank " data is collected (for example, locomotive edge is equipped with the railroad track section row of hot tank sensor
The time entered) can be inconsistent with sensor and/or actuator time of measuring.In such cases, analysis system 108 can be configured
To perform one or more operations to determine the external data that will be observed originally at the time corresponding with sensor time of measuring
Observed value.
Specifically, analysis system 108 can using the time of external data observed value and the time of measured value come interpolation outside
Portion's data observed value is to produce the external data value of time corresponding with time of measuring.The interpolation of external data can allow outside
Data observed value or the feature being derived from are included in fault model as input.In fact, in addition to other examples, may be used also
Utilize sensor and/or actuator data using various technologies come interpolation external data, for example, arest neighbors interpolation, it is linear in
Slotting, polynomial interpolation and spline interpolation.
Fig. 7 is returned to, the given failure in the failure collection defined at the next comfortable frame 702 of the determination of analysis system 108
After fault model, analysis system 108 can determine whether that any remaining failure of its determination fault model should be directed to.Still
Right presence should be directed in the case of the failure of its determination fault model, and analysis system 108 repeats the loop of frame 704 to 708.One
In a little embodiments, analysis system 108 can determine that the single fault model of the owner of the failure defined at frame 702.
In other embodiments, analysis system 108 can determine that the fault model of each subsystem of assets 102, then using institute
Fault model is stated to determine Assets Levels fault model.Other examples are also possible.
Finally, at frame 710, then each failure in institute's definition set can be defined relation (for example, indivedual events
Barrier model) it is combined into for predicting that the totality to be broken down in following preset time scope (for example, ensuing two weeks) can
The model (for example, health indicator model) of energy property.That is, described model can be received from one or more sensors and/or cause
The sensor and/or actuator measured value of dynamic device are exported and will broken down in the range of the preset time in future as input
The individual probability of at least one failure in set.
Analysis system 108 can define health indicator model in many ways, and this may depend on health indicator
It is expected granularity.That is, in the case where multiple fault models be present, can be obtained in many ways using the result of fault model strong
The output of health status index model.For example, analysis system 108 can from multiple fault models determining maximum, intermediate value or average
It is worth, and the determination value is used as to the output of health indicator model.
In other examples, determine that health indicator model can relate to analysis system 108 and belong to weight by indivedual
The individual probability of fault model output.For example, each failure from failure collection can be considered as same undesirable, therefore
Each probability equally can be it is determined that carry out identical weighting during health indicator model.In other cases, some failures may
It is considered as more being out of favour (for example, more calamity or needs longer maintenance time etc.) than other failures, therefore those are right
The probability answered may more be weighted than other probability.
In other examples again, determine that health indicator model can relate to analysis system 108 and be built using one or more
Mould technology, such as regression technique.Aggregate response variable can take the response variable (example from each of individual failure model
Such as, the Y in Fig. 8res) logical separation (logic or) form.For example, with determined in frame 706 any time scope (for example,
Fig. 8 time range 814) in the associated aggregate response variable of any set of measurements for occurring can have a value 1, and with when
Between the associated aggregate response variable of the set of measurements that occurs outside any one of scope can have null value.Define healthy shape
The other manner of condition index model is also possible.
In some embodiments, frame 710 is probably unnecessary.For example, as discussed above, analysis system 108 can
Single fault model is determined, in this case, health indicator model can be single fault model.
In fact, analysis system 108 can be configured to update single fault model and/or general health index mould
Type.Analysis system 108 can daily, weekly, the more new model such as monthly, and being based on from assets 102 or from other assets (example
Such as, from the other assets with the identical group of assets 102) historical operating data new portion.Other examples are also possible.
C. deployment model and workflow
In 108 Definition Models of analysis system-workflow to afterwards, analysis system 108 can be by model-workflow pair of definition
It is deployed to one or more assets.Specifically, analysis system 108 can launch at least one assets (such as assets 102) and define
Forecast model and/or corresponding workflow.Analysis system 108 can be periodically or based on trigger event (such as to setting models-
Any modification or renewal of workflow pair) carry out Launching Model-workflow pair.
In some cases, analysis system 108 can only launch one of personalized model or individualized work stream.Example
Such as, in the case where analysis system 108 only defines personalized model or workflow, analysis system 108 can launch workflow or mould
The polymerization version and personalized model or workflow of type, or if polymerization version is stored in data storage dress by assets 102
In putting, then analysis system 108 may not be needed transmitting polymerization version.In a word, analysis system 108 can launch (1) personalized mould
Type and/or individualized work stream, (2) personalized model and polymerization workflow, (3) polymerization model and individualized work stream or (4)
Polymerization model and polymerization workflow.
In fact, in the operation for the frame 702 to 710 that analysis system 108 may perform Fig. 7 for multiple assets
Some or all define model-workflow pair of each assets.For example, analysis system 108 can define the mould of assets 104 in addition
Type-workflow pair.Analysis system 108 can be configured simultaneously or sequentially to launch corresponding model-work to assets 102 and 104
Make stream pair.
D. assets locally execute
Such as the given assets of the grade of assets 102 can be configured to receive model-workflow to or part thereof, and according to being received
Model-workflow to operating.That is, assets 102 can in data storage device storage model-workflow pair, and will be by providing
The data input that the sensor and/or actuator of production 102 obtain is into forecast model, and the output for being sometimes based upon forecast model is held
The corresponding workflow of row.
In fact, the various assemblies of assets 102 can perform forecast model and/or corresponding workflow.For example, as begged for above
By each assets can include local analytics device, and it is configured to store and runs model-work for being provided by analysis system 108
Make stream pair.When local analytics device receives particular sensor and/or actuator data, it can input received data
Into forecast model, and depending on one or more operations of the executable corresponding workflow of the output of model.
In another example, the CPU of the assets 102 separated with local analytics device can perform forecast model
And/or corresponding workflow.In other examples again, the local analytics device and CPU of assets 102 can be performed in unison with
Model-workflow pair.For example, local analytics device can perform forecast model, and CPU can perform workflow, on the contrary
It is as the same.
In example implementation, model-workflow is performed locally to before (or model may be being locally executed first
During workflow), the forecast model of local analytics device personalizable assets 102 and/or corresponding workflow.No matter model-work
Flow to being the form using polymerization model-workflow pair or personalized model-workflow pair, possible this thing happens.
As indicated above, analysis system 108 can be based on some predictions on assets group or special assets, hypothesis
And/or vague generalization carrys out Definition Model-workflow pair.For example, in Definition Model-workflow pair, analysis system 108 is predictable,
Assuming that and/or vague generalization assets relevant characteristic and/or assets operating conditions and other considerations.
Anyway, local analytics device personalization forecast model and/or corresponding workflow can relate to local analytics device
Confirm or overthrow one or more predictions, hypothesis and/or vague generalization that analysis system 108 is made in Definition Model-workflow pair
One or more of.According to local analytics device to predicting, assuming and/or hereafter general assessment, local analytics device may be used
Modification (or further being changed in the case of personalized model and/or workflow) forecast model and/or workflow.With
This mode, local analytics device can help to define more real and/or accurate model-workflow pair, and this may cause more to have
The assets monitoring of effect.
In fact, local analytics device can be based on many considerations come personalized forecast model and/or workflow.For example, this
Ground analytical equipment can so be done based on operation data caused by one or more sensors and/or actuator by assets 102.Tool
For body, local analytics device can be by following item come personalized:(1) obtain by one or more sensors and/or actuator
Operation data caused by particular demographic (for example, this data is obtained by the CPU via assets indirectly, or may
Directly from sensor and/or actuator in itself in some directly obtain this data;(2) assessed based on the operation data obtained
With model-workflow to associated one or more predictions, hypothesis and/or vague generalization;And (3) if it is described assess instruction appoint
What prediction, hypothesis and/or vague generalization are incorrect, then correspondingly change model and/or workflow.This generic operation can be each
Kind mode performs.
In an example, local analytics device (for example, CPU via assets) obtain by sensor and/
Or operation data caused by the particular demographic of actuator can based on the part as model-workflow pair or therewith by comprising
Instruction.In particular, one or more tests for instructing recognizable local analytics device to perform, the test, which is assessed, defines mould
Some or all predictions, hypothesis and/or the vague generalization being related to during type-workflow pair.Each recognizable local analytics dress of test
One or more sensors of interest of interest of operation data and/or actuator, the behaviour to be obtained will be obtained for it by putting
The amount and/or other tests for making data consider.Therefore, local analytics device obtains the particular cluster by sensor and/or actuator
Operation data caused by group may relate to local analytics device and obtain this operation data according to test instruction etc..Local analytics device
It is also possible to obtain other examples of the operation data for personalized model-workflow pair.
As described above, after operation data is obtained, local analytics device availability data assesses Definition Model-work
Some or all predictions, hypothesis and/or the vague generalization being related to during stream pair.This operation can perform in a variety of ways.In a reality
In example, local analytics device can be by the operation data obtained and one or more threshold values (for example, the threshold value model of threshold value and/or value
Enclose) it is compared.Generally, given threshold value or scope may correspond to for Definition Model-workflow pair one or more prediction,
Assuming that and/or vague generalization.Specifically, each sensor or actuator (or sensor and/or the actuating identified in test instruction
The combination of device) can have corresponding threshold value or scope.Local analytics device then can determine that by given sensor or actuator
Caused operation data is above or below corresponding threshold value or scope.Local analytics device assessment prediction, hypothesis and/or one
As other examples for changing be also possible.
Hereafter, local analytics device can change (or not changing) forecast model and/or workflow based on assessing.I.e., such as
It is incorrect that fruit, which assesses any prediction of instruction, hypothesis and/or vague generalization, then local analytics device can correspondingly change prediction
Model and/or workflow.Otherwise, local analytics device can perform model-workflow pair in the case of no modification.
In fact, local analytics device can change forecast model and/or workflow in many ways.For example, local point
Analysis apparatus can (for example, value or scope for passing through modified values) modification forecast model and/or workflow one or more parameters and/or
Trigger point of forecast model and/or workflow etc..
As a non-limiting examples, it is assumed that the engine operating temperature of assets 102 is no more than specified temp, analysis system
System 108 may have been defined for model-workflow pair of assets 102.As a result, the part of the forecast model of assets 102 can relate to
Determine the first calculated value and then only the first calculated value exceed based on the assumption that engine operating temperature determine threshold value when
Determine the second calculated value.When personalized model-workflow pair, local analytics device can obtain to be started by measurement assets 102
Data caused by one or more sensors and/or actuator of the operation data of machine.Then, this number can be used in local analytics device
According to determining whether on assuming that in fact that for engine operating temperature be true (for example, whether engine operating temperature exceedes threshold
Value).If data instruction engine operating temperature has more than the value for the specified temp assumed or more than the specified temp threshold
Value amount, then local analytics device can for example change the threshold value that triggering determines the second calculated value.Local analytics device is personalized pre-
It is also possible to survey other examples of model and/or workflow.
Local analytics device can be based on extra or substitute consideration come personalized model-workflow pair.For example, local analytics fill
One or more asset characters (such as any one of asset character discussed above) can be based on so to do by putting, and the characteristic can
Local analytics device is determined or is supplied to by local analytics device.Other examples are also possible.
In example implementation, after local analytics device personalization forecast model and/or workflow, local analytics
Device can provide forecast model and/or workflow personalized instruction to analysis system 108.This instruction can take various shapes
Formula.For example, the forecast model and/or the aspect of workflow of the recognizable local analytics device modification of instruction or part (for example,
What the parameter and/or parameter changed be modified to) and/or can recognize that the reason for changing (for example, causing local analytics device
The description of the fundamental operation data or other assets data and/or reason modified).Other examples are also possible.
In some example implementations, local analytics device and the both of which of analysis system 108 can be in personalized model-works
Flow to being involved, personalized model-workflow pair can perform in a variety of ways.For example, analysis system 108 can be to local point
Analysis apparatus provides the instruction of some situations and/or characteristic of test assets 102.Based on the instruction, local analytics device can be
Test is performed at assets 102.For example, local analytics device can obtain the behaviour as caused by special assets sensor and/or actuator
Make data.Hereafter, local analytics device can provide the result from test status to analysis system 108.Based on such result, divide
Analysis system 108 can correspondingly define assets 102 forecast model and/or workflow and be transmitted to local analytics device with
In locally executing.
In other examples, the same or similar test in the executable part with performing workflow of local analytics device is grasped
Make.That is, particular workflow corresponding with forecast model can cause local analytics device to perform some tests and to analysis system 108
Emission results.
In example implementation, in local analytics device personalization forecast model and/or workflow (or and analysis system
108 work with personalized forecast model and/or workflow together) after, local analytics device can perform personalized forecast model
And/or workflow rather than archetype and/or workflow are (for example, what local analytics device initially received from analysis system 108
Model and/or workflow).In some cases, although local analytics device performs personalized version, local analytics device
The prototype version of model and/or workflow can be retained in data storage device.
Generally, assets perform prediction model and based on gained export perform workflow operation can promote determine it is defeated by model
The reason for possibility that the particular event that goes out occurs and/or can promote, prevents from following particular event occurring.When performing workflow,
Assets can determine locally and take action to help prevent event, and this makes such determination in dependency analysis system 108
And the action that recommendation is provided be it is invalid or it is infeasible (for example, when network delay be present, when network connection is bad, work as money
Production remove communication network 106 coverage when etc.) in the case of be beneficial.
In fact, assets can perform prediction model in a variety of ways, this is likely to be dependent on particular prediction model.Fig. 9 is to retouch
Paint the flow chart 900 available for the possible example for locally executing the stage for locally executing forecast model.Will be in output assets
Health indicator health indicator model background in discuss example locally execute the stage it should be appreciated that
It is, it is same or similar to locally execute the stage and can be used for other types of forecast model.In addition, for purpose of explanation, example sheet
The ground execution stage is described as the local analytics device implementation by assets 102, but this stage also by other devices and/or can be
System is carried out.One of ordinary skill in the art are readily apparent that, flow chart 900 is provided for clear and task of explanation, and available
Operation and several other combinations of function locally execute forecast model.
As shown in Figure 9, at frame 902, local analytics device can receive the number of the current operating situation of reflection assets 102
According to.At frame 904, local analytics device can identify from received data will be input into the mould provided by analysis system 108
Operation data set in type.At frame 906, the operation data set identified then can be input to mould by local analytics device
In type and moving model is to obtain the health indicator of assets 102.
When local analytics device continues to the operation data of the renewal of assets 102, local analytics device can also pass through
The operation of operation data repeat block 902 to 906 based on renewal and continue the health indicator of more new assets 102.At some
In the case of, can when each local analytics device receives new data from the sensor and/or actuator of assets 102 or periodically
The operation of the repeat block 902 to 906 such as (for example, per hour, daily, weekly, monthly).In this way, money is used when in operation
During production, local analytics device can be configured to dynamically update health indicator, possible real-time update.
The function that example illustrated in fig. 9 locally executes the stage will be described in further detail now.At frame 902,
Local analytics device can receive the data of the current operating situation of reflection assets 102.This data can include the biography from assets 102
The actuator data of the sensing data of one or more of sensor, one or more actuators from assets 102, and/or its
Unusual condition data and other types of data can be included.
At frame 904, local analytics device can be identified to be input into and provided by analysis system 108 from received data
Health indicator model in operation data set.This operation can perform in many ways.
Characteristic that in an example, local analytics device can be based on assets 102 (such as just it is directed to its determination health status
The Asset Type or class of assets of index) identify that the operation data input for the model is gathered (for example, coming from specific sensing
The data of device and/or actuator of interest).In some cases, the operation data input set identified can be from assets
Some or all of persons in 102 sensor in the sensing data of some or all of persons and/or actuator from assets 102
Actuator data.
In another example, local analytics device can identify operation based on the forecast model provided by analysis system 108
Data input set.That is, analysis system 108 can be provided to assets 102 the specific input for model some instruction (for example,
In forecast model or in the transmitting of single data).Identify that other examples of operation data input set are also possible.
At frame 906, the operation of local analytics device can then run health indicator model.Specifically, local point
The operation data set identified can be input in model by analysis apparatus, the model and then determined and exported given in future
The overall possibility of at least one failure occurs for (for example, ensuing two weeks) in time range.
In some embodiments, this operation can relate to local analytics device by certain operational data (for example, sensor
And/or actuator data) be input in one or more individual failure models of health indicator model, each individual failure mould
The exportable individual probability of type.Then, these individual probabilities can be used in local analytics device, may be according to health indicator model
Some probability weights must be more than other probability, it is overall possible to determine to break down in the range of the preset time in future
Property.
It is determined that after the overall possibility to break down, the probability to break down can be converted into being good for by local analytics device
Health status index, the health indicator can be taken to be reflected in the time range in future in (for example, two weeks) and will not be in
The form of the single polymerization parameter of the possibility to be broken down at assets 102.In example implementation, probability of malfunction is changed
The benefit that local analytics device determines probability of malfunction is can relate into health indicator.Specifically, total breakdown probability can be taken
The form of value from 0 to 1;Health indicator can be by subtracting the numeral to determine from 1.Probability of malfunction is converted into health
Other examples of status index are also possible.
After assets locally execute forecast model, assets then can based on performed forecast model gained export come
Perform corresponding workflow.Generally, assets execution workflow, which can relate to local analytics device, causes to perform operation (example at assets
Such as, by one or more of mobile system to assets send instruct) and/or local analytics device cause such as analysis system
The computing system such as 108 and/or output system 110 performs the operation away from assets.As mentioned above, workflow can take various shapes
Formula, therefore workflow can perform in a variety of ways.
For example, assets 102 can be made internally to perform one or more operations for some behaviors for changing assets 102, such as repair
Change data acquisition and/or launch scenario, perform local diagnostic tool, change assets 102 operating conditions (for example, modification speed,
Acceleration, fan speed, spiral propeller angle, air inlet etc. perform other machinery via one or more actuators of assets 102
Operation), or output be probably the relatively low index of health status or the preventive actions of recommendation instruction, the preventive actions
Assets 102 should be performed at the user interface of assets 102 or outside computing system is performed.
In another example, assets 102 can cause to system (such as output system 110) transmitting on communication network 106
System carries out the instruction of operation, such as produces work order or order specific component for repairing assets 102.In another example again
In, assets 102 can be communicated with remote system (such as analysis system 108), and the remote system then promotes to cause to operate
Occur away from assets 102.Other examples that assets 102 locally execute workflow are also possible.
E. model/workflow modification stage
In another aspect, 108 practicable modification stage of analysis system, during the modification stage, analysis system 108
Model and/or workflow based on new asset data modification deployment.Polymerization and both personalized model and workflow can be held
This stage of row.
In particular, when given assets (for example, assets 102) according to model-workflow to operation when, assets 102 can be to
Analysis system 108 provides operation data and/or data source 112 can provide the external number related to assets 102 to analysis system 108
According to.At least be based on this data, analysis system 108 can change assets 102 model and/or workflow and/or other assets (such as
Assets 104) model and/or workflow.When changing the model and/or workflow of other assets, analysis system 108 can be shared
The information obtained from the action learning of assets 102.
In fact, analysis system 108 can modify in many ways.Figure 10 is to describe to can be used for modification model-work
The flow chart 1000 of one possible example of the modification stage of stream pair.For purpose of explanation, the example modifications stage be described as by
Analysis system 108 is carried out, but this modification stage can also be carried out by other systems.One of ordinary skill in the art will be bright
In vain, flow chart 1000 is provided for clear and task of explanation, and model-work can be changed using several other combinations of operation
Stream pair.
As shown in Figure 10, at frame 1002, analysis system 108 can receive analysis system 108 and identify particular event from it
Generation data.Data can be operation data from assets 102 or from data source 112 it is related to assets 102 outside
The data such as portion's data.The event can take the failure at the form of any event discussed above, such as assets 102.
In other example implementations, event can take new component or subsystem to be added to the form of assets 102.
Another event can take the form of " leading designator " event, and this may relate to the sensor of assets 102 and/or actuator produces
From the data of the data different (may differ by threshold difference) identified during the model definition stage at Fig. 7 frame 706.This difference can
Indicate that assets 102 have the operating conditions of the normal operation conditions higher or lower than the assets similar with assets 102.Another thing again
Part can take the form for the event for being followed by one or more leading designator events.
Generation and/or basic data based on the particular event identified are (for example, the operation data related to assets 102
And/or external data), analysis system 108 can then change polymerization, forecast model and/or workflow and/or one or more individual characteies
Change forecast model and/or workflow.In particular, at frame 1004, analysis system 108 can be determined whether modification polymerization prediction mould
Type.Analysis system 108 can determine modification polymerization forecast model due to many reasons.
If for example, the particular event identified is the first time hair comprising the multiple assets including assets 102
Raw (such as specific fault occurs for the first time at the assets from assets troop or is added to by specific New Parent for the first time
From the assets of assets troop), then analysis system 108 can change polymerization forecast model.
In another example, if the associated data of the generation of the particular event with being identified be different from being used for it is initially fixed
The data of adopted polymerization model, then analysis system 108 can modify.For example, the generation of the particular event identified may be
Do not occurred under the operating conditions associated with the generation of particular event (for example, specific fault may be previously not described previously
Occur under the associated sensor value measured under specific fault).The other reasons for changing polymerization model are also possible.
If analysis system 108 determines modification polymerization forecast model, then analysis system 108 can be at frame 1006 so
Do.Otherwise, analysis system 108 may proceed to frame 1008.
At frame 1006, analysis system 108 can be based at least partially on received at frame 1002 it is related to assets 102
Data change polymerization model.In example implementation, (such as it can be discussed in a variety of ways above with reference to Fig. 5 frame 510
Any mode) modification polymerization model.In other embodiments, polymerization model can also other manner modification.
At frame 1008, analysis system 108 can next determine whether modification polymerization workflow.Analysis system 108 can be due to
Many reasons and change polymerization workflow.
For example, analysis system 108 can be based on polymerization model whether being changed at frame 1004 and/or at analysis system 108
Polymerization workflow is changed with the presence or absence of other changes.In other examples, although assets 102 perform polymerization workflow, but such as
Identified event occurs at frame 1002 and occurs for fruit, then analysis system 108 can change polymerization workflow.If for example, work
It is intended to that help promotes to prevent the generation of event (for example, failure) and workflow is executed correctly as stream, but event still occurs,
So analysis system 108 can change polymerization workflow.Other reasons of modification polymerization workflow are also possible.
If analysis system 108 determines modification polymerization workflow, then analysis system 108 can so be done at frame 1010.
Otherwise, analysis system 108 may proceed to frame 1012.
At frame 1010, analysis system 108 can be based at least partially on received at frame 1002 it is related to assets 102
Data change polymerization workflow.In example implementation, (such as it can be begged in a variety of ways above with reference to Fig. 5 frame 514
Any mode of opinion) modification polymerization workflow.In other embodiments, polymerization model can also other manner modification.
At frame 1012 to frame 1018, analysis system 108 can be configured to be received with being based at least partially at frame 1002
Data related to assets 102 (for example, for each of assets 102 and 104) change one or more personalized models
And/or (for example, for one in assets 102 or assets 104) change one or more individualized work streams.Analysis system 108
The mode that frame 1004 to 1010 can be similar to so is done.
However, the reason for modification personalized model or workflow, is different the reason for may be from polymerization situation.For example, analysis system
System 108 is it is further contemplated that be initially used for defining the underlying assets characteristic of personalized model and/or workflow.In particular instance
In, if the particular event identified is the first of this particular event of the assets of the asset character with assets 102
Secondary generation, then analysis system 108 can change personalized model and/or workflow.Change personalized model and/or workflow
Other reasons are also possible.
In order to illustrate, Fig. 6 D be modification model-workflow to 630 conceptual illustration.Specifically, model-workflow
It is the revision of polymerization model-workflow pair from Fig. 6 A to explanation 630.As indicated, model-workflow of modification is to saying
Bright 630 include the original row of the mode input 602 from Fig. 6 A, and comprising for model calculating 634, model output area 636
And the modification row of workflow operation 638.In this example, the forecast model of modification has the single input number from sensors A
According to, and there are two calculated values:Calculated value I and III.If the output probability for changing model is less than 75%, then performs work
Stream operation 1.If output probability is between 75% and 85%, then performs workflow operation 2.And if output probability is more than
85%, then perform workflow operation 3.Other example modifications model-workflows are to being herein defined as possible and paying attention to.
Return to Figure 10, at frame 1020, analysis system 108 can then launch to one or more assets any model and/
Or workflow modification.For example, analysis system 108 can launch of modification to assets 102 (for example, data cause the assets of modification)
Property model-workflow pair and to assets 104 launch modification polymerization model.In this way, analysis system 108 can be based on and money
Such modification is distributed to multiple moneys by the data that the operation of production 102 is associated dynamically to change model and/or workflow
Production, such as the troop belonging to assets 102.Therefore, other assets can be benefited from the data from assets 102, because other moneys
Local model-workflow pair of production can be modified based on this data, thereby aided in and created more accurate and sane model-work
Stream pair.
Although above-mentioned modification stage is discussed as being performed by analysis system 108, in example implementation, assets
102 local analytics device additionally or alternatively can carry out modification stage with similar mode discussed above.For example, one
In individual example, when assets 102 by using operation data is to operate caused by one or more sensors and/or actuator when,
Local analytics device can change model-workflow pair.Therefore, the local analytics device of assets 102, analysis system 108 or one
A little combinations can change forecast model and/or workflow in the conditions associated change of assets.In this way, local analytics device and/
Or analysis system 108 can continuously adjust model-workflow pair based on its available nearest data.
F. dynamic execution model/workflow
In another aspect, assets 102 and/or analysis system 108 can be configured dynamically to adjust execution model-workflow
It is right.In particular, assets 102 and/or analysis system 108 can be configured to detect triggering on assets 102 and/or analysis system
Whether 108 answer some events of the change of the responsibility of perform prediction model and/or workflow.
In operation, assets 102 and the both of which of the analysis system 108 executable model-workflow pair for representing assets 102
It is all or part of.For example, receive model-workflow to afterwards from analysis system 108 in assets 102, assets 102 can by model-
Workflow then can be dependent on the concentrative implementation model of analysis system 108-work centering to being stored in data storage device
It is part or all of.In particular, assets 102 can provide at least sensor and/or actuator data to analysis system 108, point
Then this data can be used to carry out the forecast models of concentrative implementation assets 102 for analysis system 108.Output based on model, analysis system
108 then can perform corresponding workflow, or output that analysis system 108 can be to the Launching Model of assets 102 or make assets 102 local
Perform the instruction of workflow.
In other examples, analysis system 108 can be dependent on assets 102 to locally execute the portion of model-workflow centering
Divide or whole.Specifically, assets 102 can locally execute part or all of in forecast model, and transmit the result to analysis
System 108, the result can then cause the concentrative implementation of analysis system 108 to correspond to workflow.Or assets 102 can also locally execute
Corresponding workflow.
In other examples, analysis system 108 and assets 102 can share the responsibility of execution model-workflow pair again.Example
Such as, analysis system 108 can concentrative implementation model and/or workflow part, and assets 102 locally execute model and/or work
The other parts of stream.Assets 102 and analysis system 108 can launch the result of the responsibility from their corresponding executeds.It is other
Example is also possible.
At some time point, assets 102 and/or analysis system 108, which can determine that, should adjust model-workflow to holding
OK.That is, one or both can determine that performing responsibility should be changed.This operation can occur in a variety of ways.
Figure 11 is the flow of the possible example for the adjusting stage for describing the execution that can be used for adjustment model-workflow pair
Figure 110 0.For purpose of explanation, the example adjusting stage is described as being carried out by assets 102 and/or analysis system 108, but this
Modification stage can also be carried out by other systems.One of ordinary skill in the art are readily apparent that, are carried for clear and task of explanation
For flow chart 1100, and the execution of model-workflow pair can be adjusted using several other combinations of operation.
At frame 1102, assets 102 and/or analysis system 108 can detect Dynamic gene (or potentially multiple adjustment because
Son), its instruction needs the situation that the execution to model-workflow pair is adjusted.The example of such situation includes communication network
106 network condition or the treatment situation of assets 102 and/or analysis system 108 etc..Example network situation can prolong comprising network
Late, the signal intensity of the link between network bandwidth, assets 102 and communication network 106, or some other instructions of network performance
Etc..Instance processes situation can include processing capacity (for example, available disposal ability), processing usage amount (for example, being consumed
The amount of disposal ability) or disposal ability some it is other instruction etc..
In fact, detection Dynamic gene can perform in a variety of ways.For example, this operation can relate to determine network (or processing)
Whether situation reaches one or more threshold values or whether situation changes in some way.Other examples of detection Dynamic gene are also can
Can.
In particular, in some cases, detection Dynamic gene can relate to assets 102 and/or analysis system 108 detects
The signal intensity of communication link between assets 102 and analysis system 108 becomes less than the instruction of threshold signal strength or with some
The instruction that rate reduces.In this example, Dynamic gene may indicate that assets 102 will " offline ".
In another case, detection Dynamic gene can additionally or alternatively be related to assets 102 and/or analysis system
108 detect network delay higher than threshold delay or with the increased instruction of some rate of change.Or the instruction can be network bandwidth
Decline less than threshold value bandwidth or with some rate of change.In these examples, Dynamic gene may indicate that communication network 106 lags.
In the case of again other, detection Dynamic gene can additionally or alternatively be related to assets 102 and/or analysis system 108
Detect that processing capacity is reduced and/or handled usage amount less than specific threshold or with some rate of change and become higher than threshold value or with some
The increased instruction of rate.In such example, Dynamic gene may indicate that the disposal ability of assets 102 (and/or analysis system 108)
It is relatively low.It is also possible to detect other examples of Dynamic gene.
At frame 1104, based on the Dynamic gene detected, adjustable to locally execute responsibility, this can send out in many ways
It is raw.For example, assets 102 may have detected that Dynamic gene, and then determine to locally execute model-workflow pair or its portion
Point.In some cases, assets 102 then can to analysis system 108 launch assets 102 be performed locally forecast model and/or
The notice of workflow.
In another example, analysis system 108 may have detected that Dynamic gene, and then refer to the transmitting of assets 102
Order with cause assets 102 locally execute model-workflow to or part thereof.Based on the instruction, assets 102 then can be in local
Perform model-workflow pair.
At frame 1106, concentrative implementation responsibility is can adjust, this can occur in many ways.For example, analysis system can be based on
108 detect that assets 102 are performed locally the instruction of forecast model and/or workflow to adjust concentrative implementation responsibility.Analysis system
System 108 can detect this instruction in many ways.
In some instances, analysis system 108 can be performed locally forecast model by receiving assets 102 from assets 102
And/or the notice of workflow detects the instruction.Notice can take various forms, such as binary system or text, and recognizable
The particular prediction model and/or workflow that assets are performed locally.
In other examples, analysis system 108 can detect instruction based on the operation data received of assets 102.Tool
For body, detection instruction can relate to analysis system 108 and receive the operation data of assets 102, and then detect received data
One or more characteristics.According to received data one or more characteristics for detecting, analysis system 108 is it can be inferred that assets
102 are performed locally forecast model and/or workflow.
Indeed, it is possible to various modes carry out one or more characteristics of perform detection received data.For example, analysis system
The type of 108 detectable received datas.In particular, analysis system 108 can detect data source, for example, produce sensor or
The particular sensor or actuator of actuator data.Type based on received data, analysis system 108 is it can be inferred that assets
102 are performed locally forecast model and/or workflow.For example, based on the sensor id for detecting particular sensor, analysis
For system 108 it can be inferred that assets 102 are performed locally forecast model and corresponding workflow, it causes assets 102 from specific sensing
Device gathered data simultaneously launches the data to analysis system 108.
In another situation, analysis system 108 can detect the amount of received data.Analysis system 108 can be by the amount
Compared with some data threshold amount.Threshold quantity is reached based on the amount, analysis system 108 it can be inferred that assets 102
Locally executing causes the collection of assets 102 equal to or more than the forecast model and/or workflow of the data volume of threshold quantity.Other examples
It is and possible.
In example implementation, one or more characteristics for detecting received data can relate to analysis system 108 and detect
Specific change in one or more characteristics of received data, for example, the type of received data change, received
The change of data volume or the change of the frequency of reception data.In particular instances, the change of the type of received data can relate to
And the change in its sensing data source received of the detection of analysis system 108 (is provided to analysis system for example, producing
The change of the sensor and/or actuator of 108 data).
In some cases, the change detected in received data can relate to the number that analysis system 108 will receive in the recent period
Compared with the data received with the past (for example, one hour, one day, one week etc. before current time).Anyway, base
In the change for one or more characteristics for detecting received data, analysis system 108 is it can be inferred that assets 102 are performed locally
Forecast model and/or workflow, it causes this change of the data to providing analysis system 108 by assets 102.
In addition, analysis system 108 can be performed locally based on Dynamic gene is detected at frame 1102 to detect assets 102
The instruction of forecast model and/or workflow.For example, if analysis system 108 detects Dynamic gene at frame 1102, then point
Analysis system 108 can cause the adjustment of assets 102 to the transmitting of assets 102, and it locally executes the instruction of responsibility, and correspondingly, analysis system
108 can adjust the concentrative implementation responsibility of oneself.Other examples of detection instruction are also possible.
In example implementation, concentrative implementation responsibility can be adjusted according to the adjustment for locally executing responsibility.For example, such as
Fruit assets 102 are performed locally forecast model now, then analysis system 108 can correspondingly stop concentrative implementation forecast model
(and may or may not stop to perform corresponding workflow).In addition, if assets 102 are performed locally corresponding workflow, then analysis
System 108 can correspondingly stop performing workflow (and may or may not stop concentrative implementation forecast model).Other examples are also can
Can.
In fact, assets 102 and/or analysis system 108 can be consecutively carried out the operation of frame 1102 to 1106.Sometimes, may be used
Local and concentrative implementation responsibility are adjusted to promote the execution of Optimized model-workflow pair.
In addition, in some embodiments, assets 102 and/or analysis system 108 can be held based on Dynamic gene is detected
The other operations of row.For example, based on communication network 106 situation (for example, bandwidth, delay, signal intensity or network quality it is another
Instruction), assets 102 can be performed locally particular workflow.Analysis system 108 can be based on analysis system 108 and detect communication network
Situation particular workflow is provided, the particular workflow may have stored in assets 102, or can be to have stored in money
The revision (for example, assets 102 can be in locally modified workflow) of workflow in production 102.In some cases, except other
Outside possible workflow operation, particular workflow can include data acquisition plan and/or the increase for increasing or decreasing sampling rate
Or reduce the emissivity of data or the data emission case of quantity for being launched into analysis system 108.
In particular instances, assets 102 can determine that one or more situations detected of communication network have reached corresponding
Threshold value (for example, instruction network quality bad).Based on this determination, assets 102 can be performed locally workflow, the workflow
Comprising data are launched according to data emission case, the data emission case reduces the number that assets 102 are transmitted into analysis system 108
According to amount and/or frequency.Other examples are also possible.
V. case method
Turning now to Figure 12, depict explanation and be used to define and dispose and can predict mould by the polymerization that analysis system 108 performs
The flow chart of the case method 1200 of type and corresponding workflow.For method 1200 and the other methods being discussed below, by flow
The operation illustrated by frame in figure can perform according to the above discussion.In addition, one or more operations discussed above can be added
To given flow chart.
At frame 1202, method 1200 can relate to analysis system 108 and receive multiple assets (such as assets 102 and 104)
Corresponding operating data.At frame 1204, method 1200 can relate to analysis system 108 based on the operation data received come define with
The forecast model of the operation correlation of multiple assets and corresponding workflow (for example, fault model and corresponding workflow).In frame 1206
Place, method 1200 can relate at least one assets (for example, assets 102) transmitting prediction of the analysis system 108 into multiple assets
Model and correspondingly workflow are locally executed at least one assets.
Figure 13 is depicted for define and dispose can by personalized forecast model that analysis system 108 performs and/or correspondingly
The flow chart of the case method 1300 of workflow.At frame 1302, method 1300 can relate to analysis system 108 and receive multiple assets
Operation data, plurality of assets comprise at least the first assets (such as assets 102).At frame 1304, method 1300 can relate to
And analysis system 108 is defined the polymerization forecast model related to the operation of multiple assets and gathered based on the operation data received
Close corresponding workflow.At frame 1306, method 1300 can relate to one or more characteristics that analysis system 108 determines the first assets.
At frame 1308, method 1300 can relate to analysis system 108 one or more characteristics and polymerization prediction mould based on the first assets
Type and the corresponding workflow of polymerization work to define the personalized forecast model related to the operation of the first assets or personalized correspond to
At least one of stream.At frame 1310, method 1300 can relate to defined in analysis system 108 to the first assets transmitting at least
Property forecast model or the personalized workflow that corresponds to are so that the first assets are locally executed one by one.
Figure 14 is depicted for dynamically change can be by the reality of the execution for model-workflow pair that analysis system 108 performs
The flow chart of example method 1400.At frame 1402, method 1400 can relate to analysis system 108 and be sent out to assets (for example, assets 102)
The forecast model related to the operation of assets and corresponding workflow are penetrated so that assets are locally executed.At frame 1404, method
1400, which can relate to analysis system 108, detects the finger that assets are performed locally at least one of forecast model or corresponding workflow
Show.At frame 1406, method 1400 can relate to analysis system 108 based on the instruction detected come by forecast model or corresponding work
The computing system that at least one of flows changes concentrative implementation.
, can be by assets (example for dynamically changing the another method of execution of model-workflow pair similar to method 1400
Such as assets 102) perform.For example, the method can relate to assets 102 from central computer system (for example, analysis system 108) receive with
The related forecast model of operation of assets 102 and corresponding workflow.Methods described may also refer to the detection of assets 102 instruction and adjustment
The Dynamic gene for one or more situations that the execution of forecast model and corresponding workflow is associated.Methods described can relate to based on inspection
The Dynamic gene (i) measured changes locally executing for the assets 102 of at least one of forecast model or corresponding workflow, and
(ii) instruction of at least one of forecast model or corresponding workflow is performed locally to central computer system transmitting assets 102
To promote central computer system to change concentrative implementation by the computing system of at least one of forecast model or corresponding workflow.
Figure 15 is depicted for example locally executes the case method of model-workflow pair by the local analytics device of assets 102
1500 flow chart.At frame 1502, method 1500 can relate to local analytics device and be received via network interface and via local
The asset interface of analytical equipment is coupled to the forecast model of the operation correlation of the assets (for example, assets 102) of local analytics device,
The computing system (for example, analysis system 108) that wherein forecast model is positioned by remote local analytics device is based on multiple assets
Operation data defines.At frame 1504, method 1500 can relate to local analytics device and be received via asset interface for assets
102 operation data by one or more sensors and/or actuator (for example, produced and can be via the CPU of assets
The operation data received indirectly or directly from one or more sensors and/or actuator).At frame 1506, method 1500 can
It is related at least part of local analytics device based on the operation data received for assets 102 come perform prediction model.
At frame 1508, method 1500 can relate to local analytics device based on perform prediction model to perform the work corresponding to forecast model
Stream, assets 102 are caused to perform operation via asset interface wherein performing workflow packages and containing.
VI. conclusion
The example embodiment of disclosed innovation is hereinbefore described.However, those skilled in the art will manage
Solution, described embodiment can be changed and be changed without departing from the true scope of the invention being defined by the claims and
Spirit.
In addition, example described herein is related to by the row such as such as " people ", " operator ", " user " or other entities
For the operation that dynamic person performs or initiated, this is merely for the sake of example and the purpose of explanation.Claim is not necessarily to be construed as will
Such actor is asked to take action, unless clearly being described in claim language.
Claims (60)
1. a kind of computing system, it includes:
At least one processor;
Non-transitory computer-readable media;And
The programmed instruction being stored in the non-transitory computer-readable media, described program instruction can be by described at least one
Computing device is to cause the computing system:
Receive the corresponding operating data of multiple assets;
Based on the received operation data, the forecast model related to the operation of the multiple assets and corresponding work are defined
Stream;And
At least one assets into the multiple assets launch the forecast model and the corresponding workflow for it is described extremely
Few assets locally execute.
2. computing system according to claim 1, wherein corresponding operation data includes (i) and existed in special time
The unusual condition data that the failure occurred at given assets is associated, and (ii) indicate the given assets in the special time
At least one operating conditions sensor or at least one of actuator data.
3. computing system according to claim 1, wherein the forecast model is defined to export in future time section
The probability of particular event will occur at given assets.
4. computing system according to claim 3, wherein the corresponding workflow includes holding based on identified probability
One or more capable operations.
5. computing system according to claim 1, wherein the corresponding workflow, which includes given assets, controls described give
One or more actuators of assets are to promote to change the operating conditions of the given assets.
6. computing system according to claim 1, wherein the corresponding workflow includes being locally executed by given assets
One or more diagnostic tools.
7. computing system according to claim 1, wherein the corresponding workflow includes being adopted according to data acquisition plan
Collect sensing data.
8. computing system according to claim 7, wherein gathered data is given therefrom for data acquisition plan instruction
Determine one or more sensors of assets.
9. computing system according to claim 8, wherein the data acquisition plan further indicates the given assets
The data volume that will be gathered from each of one or more described sensors.
10. computing system according to claim 1, wherein the corresponding workflow is included according to data emission case to institute
State computing system transmitting data.
11. computing system according to claim 10, wherein data emission case instruction gives assets to the meter
Calculation system launches the frequency of data.
12. computing system according to claim 1, wherein the computing system is the first computing system, and it is wherein described
Corresponding workflow include given assets to the second computing system firing order with promote to cause second computing system carry out with
The related operation of the given assets.
13. computing system according to claim 1, wherein at least one assets in the multiple assets include the
One assets and the second assets, and wherein launch the forecast model and the corresponding workflow and include to first assets and institute
State the second assets and launch the forecast model and the corresponding workflow.
14. a kind of non-transitory computer-readable media for being stored with instruction above, the instruction is executable to be to cause to calculate
System performs following operate:
Receive the corresponding operating data of multiple assets;
Based on the received operation data, the forecast model related to the operation of the multiple assets and corresponding work are defined
Stream;And
At least one assets into the multiple assets launch the forecast model and corresponding workflow for described at least one
Individual assets locally execute.
15. non-transitory computer-readable media according to claim 14, wherein the forecast model is defined with defeated
Go out at given assets to occur in future time section the probability of particular event.
16. non-transitory computer-readable media according to claim 14, wherein the corresponding workflow includes giving
Assets control one or more actuators of the given assets to promote to change the operating conditions of the given assets.
17. non-transitory computer-readable media according to claim 14, wherein include will be by for the corresponding workflow
One or more diagnostic tools that given assets locally execute.
18. non-transitory computer-readable media according to claim 14, wherein the computing system is the first calculating
System, and wherein described corresponding workflow includes given assets to the second computing system firing order to promote to cause described second
Computing system performs the operation related to the given assets.
19. a kind of computer implemented method, it includes:
Receive the corresponding operating data of multiple assets;
Based on the received operation data, the forecast model related to the operation of the multiple assets and corresponding work are defined
Stream;And
At least one assets into the multiple assets launch the forecast model and corresponding workflow for described at least one
Individual assets locally execute.
20. computer implemented method according to claim 19, wherein the corresponding workflow is included according to data acquisition
Scheme gathers sensing data, wherein one or more of data acquisition plan instruction given assets of gathered data therefrom
Sensor.
21. a kind of computing system, it includes:
At least one processor;
Non-transitory computer-readable media;And
The programmed instruction being stored in the non-transitory computer-readable media, described program instruction can be by described at least one
Computing device is so that the computing system performs following operate:
The operation data of multiple assets is received, wherein the multiple assets include the first assets;
Based on the received operation data, the polymerization forecast model related to the operation of the multiple assets and polymerization are defined
Corresponding workflow;
Determine one or more characteristics of first assets;
Based on one or more characteristics described in first assets and the polymerization forecast model and the corresponding work of the polymerization
Stream, define in the personalized forecast model related to the operation of first assets or personalized corresponding workflow at least
One;And
To first assets transmitting it is described defined at least one personalized forecast model or personalized corresponding workflow with
Locally executed for first assets.
22. computing system according to claim 21, wherein one or more described characteristics of first assets include money
Produce at least one of the time limit or assets health status.
23. computing system according to claim 21, wherein determining one or more described characteristic bags of first assets
The operation data received based on first assets is included to determine one or more characteristics described in first assets.
24. computing system according to claim 21, personalized forecast model or personalized corresponding workflow defined in it
At least one of include defining the personalized forecast model and the personalized corresponding workflow, and wherein described in transmitting extremely
Few property forecast model one by one or personalized corresponding workflow include launching the personalized forecast model and the personalization
Corresponding workflow.
25. computing system according to claim 21, personalized forecast model or personalized corresponding workflow defined in it
At least one of include defining the personalized corresponding workflow, and wherein launch at least one personalized forecast model
Or personalized corresponding workflow includes launching the polymerization forecast model and the personalized corresponding workflow.
26. computing system according to claim 25, wherein the corresponding workflow of the polymerization includes the first operation, and wherein
The personalized corresponding workflow includes second operation different from the described first operation.
27. computing system according to claim 26, wherein first operation includes being gathered according to the first acquisition scheme
Data, and wherein described second operation is included according to the second acquisition scheme gathered data.
28. computing system according to claim 26, wherein first operation is included according to acquisition scheme gathered data,
And wherein described second operation includes performing one or more diagnostic tools.
29. computing system according to claim 21, wherein the multiple assets further comprise the second assets, and wherein
The instruction that described program instruction further comprises can perform causing the computing system to perform following operation:
After at least one personalized forecast model or personalized corresponding workflow is launched, instruction second money is received
The operation data of second assets of generation event at production;
The operation data received based on the second assets, change at least one personalized forecast model or personalized corresponding
Workflow;And
Launch changed at least one personalized forecast model or personalized corresponding workflow to first assets.
30. a kind of non-transitory computer-readable media for being stored with instruction above, the instruction is executable to be to cause to calculate
System performs following operate:
The operation data of multiple assets is received, wherein the multiple assets include the first assets;
Based on the received operation data, the polymerization forecast model related to the operation of the multiple assets and polymerization are defined
Corresponding workflow;
Determine one or more characteristics of first assets;
It is fixed based on one or more characteristics described in first assets and the polymerization forecast model and the polymerization workflow
At least one of the adopted personalized forecast model related to the operation of first assets or personalized corresponding workflow;
And
To first assets transmitting it is described defined at least one personalized forecast model or personalized corresponding workflow with
Locally executed for first assets.
31. non-transitory computer-readable media according to claim 30, personalized forecast model or individual defined in it
At least one of corresponding workflow of propertyization includes defining the personalized forecast model and the personalized corresponding workflow, and
Wherein launching at least one personalized forecast model or personalized corresponding workflow includes launching the personalized prediction mould
Type and the personalized corresponding workflow.
32. non-transitory computer-readable media according to claim 30, personalized forecast model or individual defined in it
At least one of corresponding workflow of propertyization includes defining the personalized corresponding workflow, and wherein launches described at least one
Personalized forecast model or personalized corresponding workflow include launching the polymerization forecast model and the personalized corresponding work
Stream.
33. non-transitory computer-readable media according to claim 32, wherein the corresponding workflow of the polymerization includes
First operation, and wherein described personalized corresponding workflow includes second operation different from the described first operation.
34. non-transitory computer-readable media according to claim 33, wherein first operation is included according to the
One acquisition scheme gathered data, and wherein described second operation is included according to the second acquisition scheme gathered data.
35. non-transitory computer-readable media according to claim 33, wherein first operation includes basis and adopted
Collection scheme gathered data, and wherein described second operation includes performing one or more diagnostic tools.
36. non-transitory computer-readable media according to claim 30, wherein the multiple assets further comprise
Second assets, and the finger that wherein described program instruction further comprises can perform causing the computing system to perform following operation
Order:
After at least one personalized forecast model or personalized corresponding workflow is launched, instruction is received described second
The operation data of second assets of event occurs at assets;
The operation data received based on the second assets, change at least one personalized forecast model or personalized corresponding
Workflow;And
At least one personalized forecast model changed to first assets transmitting or personalized corresponding workflow.
37. a kind of computer implemented method, it includes:
The operation data of multiple assets is received, wherein the multiple assets include the first assets;
Based on the received operation data, the polymerization forecast model related to the operation of the multiple assets and polymerization are defined
Corresponding workflow;
Determine one or more characteristics of first assets;
Based on one or more characteristics described in first assets and the polymerization forecast model and the corresponding work of the polymerization
Stream, define in the personalized forecast model related to the operation of first assets or personalized corresponding workflow at least
One;And
To first assets transmitting it is described defined at least one personalized forecast model or personalized corresponding workflow with
Locally executed for first assets.
38. the computer implemented method according to claim 37, personalized forecast model or personalized corresponding defined in it
At least one of workflow includes defining the personalized corresponding workflow, and wherein launches described at least one personalized pre-
Surveying model or personalized corresponding workflow includes launching the polymerization forecast model and the personalized corresponding workflow.
39. the computer implemented method according to claim 38, wherein the corresponding workflow of the polymerization includes the first operation,
And wherein described personalized corresponding workflow includes second operation different from the described first operation.
40. the computer implemented method according to claim 39, wherein in first operation or second operation
One includes performing one or more diagnostic tools.
41. a kind of computing device, it includes:
Asset interface, it is configured to the computing device being coupled to assets;
Network interface, it is configured to promote between the computing device and the computing system positioned away from the computing device
Communication;
At least one processor;
Non-transitory computer-readable media;And
The programmed instruction being stored in the non-transitory computer-readable media, described program instruction can be by described at least one
Computing device is operated below with causing the computing device to perform:
The forecast model related to the operation of the assets is received via the network interface, wherein the forecast model is by described
Computing system is defined based on the operation data of multiple assets;
The operation data of the assets is received via the asset interface;
At least part of the operation data received based on the assets performs the forecast model;And
Based on the forecast model is performed, the workflow corresponding to the forecast model is performed, wherein performing the workflow packages
Include causes the assets to perform operation via the asset interface.
42. computing device according to claim 41, wherein the computing device is communicatively coupled by the asset interface
Computer on to the assets of the assets.
43. computing device according to claim 41, wherein the assets include actuator, and wherein perform the work
Stream includes causing the actuator to perform mechanically actuated.
44. computing device according to claim 41, wherein performing the workflow includes causing the assets execution to examine
Disconnected instrument.
45. computing device according to claim 41, further included wherein performing the workflow via the network
Interface causes to perform the operation away from the assets.
46. computing device according to claim 45, wherein causing to perform the operation away from the assets including indicating institute
State computing system and perform the operation away from the assets.
47. computing device according to claim 41, wherein being stored in the non-transitory computer-readable media
Described program instruction further can cause the computing device by least one computing device:
Before the forecast model is performed, the personalized forecast model.
48. computing device according to claim 47, wherein the personalized forecast model includes at least being based on the money
The operation data received of production changes one or more parameters of the forecast model.
49. computing device according to claim 47, wherein being stored in the non-transitory computer-readable media
Described program instruction can further be caused the computing device to perform following operate by least one computing device:
After the personalized forecast model, the forecast model is launched to the computing system via the network interface
The instruction being personalized.
50. computing device according to claim 41, wherein the forecast model is the first forecast model, and wherein store
Described program instruction in the non-transitory computer-readable media further can be by least one computing device
To cause the computing device to perform following operate:
Before first forecast model is performed, launch the institute of the assets to the computing system via the network interface
The given subset of received operation data is stated, wherein the given subset of the operation data received is included by one or more
Operation data caused by the given group of sensor.
51. computing device according to claim 50, wherein being stored in the non-transitory computer-readable media
Described program instruction can further be caused the computing device to perform following operate by least one computing device:
After the given subset of the received operation data of the assets is launched, the institute with the assets is received
The second related forecast model of operation is stated, wherein second forecast model is as described in the computing system based on the assets
The given subset definitions of asset of the operation data of reception;And
Perform second forecast model rather than first forecast model.
52. a kind of non-transitory computer-readable media for being stored with instruction above, the instruction is executable to cause in terms of
The asset interface for calculating device is coupled to the following operation of computing device execution of assets:
The forecast model related to the operation of the assets, the computing device are received via the network interface of the computing device
The network interface be configured to promote between the computing device and computing system away from computing device positioning
Communication, wherein the forecast model is defined by the computing system based on the operation data of multiple assets;
The operation data of the assets is received via the asset interface;
At least part of the operation data received based on the assets performs the forecast model;And
Based on the forecast model is performed, the workflow corresponding to the forecast model is performed, wherein performing the workflow packages
Include causes the assets to perform operation via the asset interface.
53. non-transitory computer-readable media according to claim 52, calculated wherein being stored in the non-transitory
Described program instruction on machine readable media further can perform to cause the computing device:
Before the forecast model is performed, the personalized forecast model.
54. non-transitory computer-readable media according to claim 53, wherein the personalized forecast model includes
One or more parameters of the forecast model are at least changed based on the operation data received of the assets.
55. non-transitory computer-readable media according to claim 52, wherein the forecast model is the first prediction
Model, and the described program instruction being wherein stored in the non-transitory computer-readable media further can perform to cause
The computing device:
Before first forecast model is performed, launch the institute of the assets to the computing system via the network interface
The given subset of received operation data is stated, wherein the given subset of the operation data received is included by one or more
Operation data caused by the given group of sensor.
56. non-transitory computer-readable media according to claim 55, calculated wherein being stored in the non-transitory
Described program instruction on machine readable media further can perform so that the computing device performs following operate:
After the operation data is launched from the particular demographic of one or more sensors, receive described with the assets
The second related forecast model is operated, wherein second forecast model is connect as described in the computing system based on the assets
The given subset definitions of asset of the operation data of receipts;And
Perform second forecast model rather than first model.
57. a kind of computer implemented method, methods described include:
The forecast model related to the operation of assets, the net of the computing device are received via the network interface of computing device
Network interface is coupled to the assets via the asset interface of the computing device, wherein the forecast model is by away from the calculating
The computing system of device positioning is defined based on the operation data of multiple assets;
The operation data of the assets is received via the asset interface by the computing device;
The forecast model is performed by least part of the operation data that is received of the computing device based on the assets;And
Based on the forecast model is performed, the workflow corresponding to the forecast model is performed by the computing device, wherein holding
The row workflow includes causing the assets to perform operation via the asset interface.
58. computer implemented method according to claim 57, methods described further comprise:
Before the forecast model is performed, by the personalized forecast model of the computing device.
59. computer implemented method according to claim 58, wherein the personalized forecast model includes at least being based on
The operation data received of the assets changes one or more parameters of the forecast model.
60. computer implemented method according to claim 57, further included wherein performing the workflow via institute
Stating network interface causes to perform the operation away from the assets.
Applications Claiming Priority (7)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/744,369 US20160371616A1 (en) | 2014-12-01 | 2015-06-19 | Individualized Predictive Model & Workflow for an Asset |
US14/744,352 | 2015-06-19 | ||
US14/744,352 US10261850B2 (en) | 2014-12-01 | 2015-06-19 | Aggregate predictive model and workflow for local execution |
US14/744,369 | 2015-06-19 | ||
US14/963,207 US10254751B2 (en) | 2015-06-05 | 2015-12-08 | Local analytics at an asset |
US14/963,207 | 2015-12-08 | ||
PCT/US2016/037247 WO2016205132A1 (en) | 2015-06-19 | 2016-06-13 | Local analytics at an asset |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107851233A true CN107851233A (en) | 2018-03-27 |
Family
ID=60989383
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201680043854.5A Pending CN107851233A (en) | 2015-06-19 | 2016-06-13 | Local analytics at assets |
Country Status (7)
Country | Link |
---|---|
EP (1) | EP3311345A4 (en) |
JP (1) | JP2018519594A (en) |
KR (1) | KR20180011333A (en) |
CN (1) | CN107851233A (en) |
AU (1) | AU2016277850A1 (en) |
CA (1) | CA2989806A1 (en) |
HK (1) | HK1251701A1 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110427992A (en) * | 2019-07-23 | 2019-11-08 | 杭州城市大数据运营有限公司 | Data matching method, device, computer equipment and storage medium |
CN112711609A (en) * | 2019-10-24 | 2021-04-27 | 霍尼韦尔国际公司 | Method, apparatus and system for integrating and managing an automated data flow system |
CN114821856A (en) * | 2022-04-18 | 2022-07-29 | 大连理工大学 | Intelligent auxiliary device connected in parallel to automobile traveling computer for rapid automobile maintenance |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP7105213B2 (en) * | 2019-04-09 | 2022-07-22 | ソフトバンク株式会社 | Communication terminal and program |
JP7096201B2 (en) | 2019-05-29 | 2022-07-05 | 株式会社 ミックウェア | Seat terminals, programs, and information processing methods |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002033544A1 (en) * | 2000-10-17 | 2002-04-25 | Pdf Solutions, Incorporated | Method for optimizing the characteristics of integrated circuits components from circuit speficications |
JP3636883B2 (en) * | 1998-03-20 | 2005-04-06 | 富士通株式会社 | Simulation device, simulation method, and computer-readable recording medium on which simulation program is recorded |
CN101943911A (en) * | 2008-12-23 | 2011-01-12 | 巴西航空有限公司 | Be used for the health monitoring and the prediction of Mechatronic Systems and assembly |
CN101981989A (en) * | 2008-03-31 | 2011-02-23 | 高通股份有限公司 | Methods of reliably sending control signal |
CN103502899A (en) * | 2011-01-26 | 2014-01-08 | 谷歌公司 | Dynamic predictive modeling platform |
CN103516563A (en) * | 2013-10-18 | 2014-01-15 | 北京奇虎科技有限公司 | Equipment and method for monitoring abnormal or normal command |
CN103532760A (en) * | 2013-10-18 | 2014-01-22 | 北京奇虎科技有限公司 | Equipment, system and method for analyzing commands executed on hosts |
US20140208163A1 (en) * | 2013-01-22 | 2014-07-24 | General Electric Company | Systems and methods for analyzing data in a non-destructive testing system |
WO2014145977A1 (en) * | 2013-03-15 | 2014-09-18 | Bates Alexander B | System and methods for automated plant asset failure detection |
CN104392752A (en) * | 2014-10-13 | 2015-03-04 | 中国科学院合肥物质科学研究院 | Real-time on-line nuclear reactor fault diagnosis and monitoring system |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004214785A (en) * | 2002-12-27 | 2004-07-29 | Matsushita Electric Ind Co Ltd | Apparatus management system, apparatus management method, management apparatus, apparatus to be managed, apparatus management program for management apparatus, and apparatus management program for apparatus to be managed |
JP2007164724A (en) * | 2005-12-16 | 2007-06-28 | Ricoh Co Ltd | System, method and program for supporting delivery of required component |
JP2007172131A (en) * | 2005-12-20 | 2007-07-05 | Nec Fielding Ltd | Failure prediction system, failure prediction method and failure prediction program |
JP2009206850A (en) * | 2008-02-28 | 2009-09-10 | Fuji Xerox Co Ltd | Failure diagnosis device and program |
US20090327325A1 (en) * | 2008-06-30 | 2009-12-31 | Honeywell International Inc., | Meta modeling in decision support system |
JP5129725B2 (en) * | 2008-11-19 | 2013-01-30 | 株式会社日立製作所 | Device abnormality diagnosis method and system |
US20140188778A1 (en) * | 2012-12-27 | 2014-07-03 | General Electric Company | Computer-Implemented System for Detecting Anomaly Conditions in a Fleet of Assets and Method of Using the Same |
-
2016
- 2016-06-13 CN CN201680043854.5A patent/CN107851233A/en active Pending
- 2016-06-13 KR KR1020187001578A patent/KR20180011333A/en unknown
- 2016-06-13 AU AU2016277850A patent/AU2016277850A1/en not_active Abandoned
- 2016-06-13 JP JP2017565106A patent/JP2018519594A/en active Pending
- 2016-06-13 EP EP16812206.7A patent/EP3311345A4/en not_active Withdrawn
- 2016-06-13 CA CA2989806A patent/CA2989806A1/en not_active Abandoned
-
2018
- 2018-08-30 HK HK18111155.8A patent/HK1251701A1/en unknown
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3636883B2 (en) * | 1998-03-20 | 2005-04-06 | 富士通株式会社 | Simulation device, simulation method, and computer-readable recording medium on which simulation program is recorded |
WO2002033544A1 (en) * | 2000-10-17 | 2002-04-25 | Pdf Solutions, Incorporated | Method for optimizing the characteristics of integrated circuits components from circuit speficications |
CN101981989A (en) * | 2008-03-31 | 2011-02-23 | 高通股份有限公司 | Methods of reliably sending control signal |
CN101943911A (en) * | 2008-12-23 | 2011-01-12 | 巴西航空有限公司 | Be used for the health monitoring and the prediction of Mechatronic Systems and assembly |
CN103502899A (en) * | 2011-01-26 | 2014-01-08 | 谷歌公司 | Dynamic predictive modeling platform |
US20140208163A1 (en) * | 2013-01-22 | 2014-07-24 | General Electric Company | Systems and methods for analyzing data in a non-destructive testing system |
WO2014145977A1 (en) * | 2013-03-15 | 2014-09-18 | Bates Alexander B | System and methods for automated plant asset failure detection |
CN103516563A (en) * | 2013-10-18 | 2014-01-15 | 北京奇虎科技有限公司 | Equipment and method for monitoring abnormal or normal command |
CN103532760A (en) * | 2013-10-18 | 2014-01-22 | 北京奇虎科技有限公司 | Equipment, system and method for analyzing commands executed on hosts |
CN104392752A (en) * | 2014-10-13 | 2015-03-04 | 中国科学院合肥物质科学研究院 | Real-time on-line nuclear reactor fault diagnosis and monitoring system |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110427992A (en) * | 2019-07-23 | 2019-11-08 | 杭州城市大数据运营有限公司 | Data matching method, device, computer equipment and storage medium |
CN112711609A (en) * | 2019-10-24 | 2021-04-27 | 霍尼韦尔国际公司 | Method, apparatus and system for integrating and managing an automated data flow system |
CN114821856A (en) * | 2022-04-18 | 2022-07-29 | 大连理工大学 | Intelligent auxiliary device connected in parallel to automobile traveling computer for rapid automobile maintenance |
CN114821856B (en) * | 2022-04-18 | 2023-04-07 | 大连理工大学 | Intelligent auxiliary device connected in parallel to automobile traveling computer for rapid automobile maintenance |
Also Published As
Publication number | Publication date |
---|---|
HK1251701A1 (en) | 2019-02-01 |
CA2989806A1 (en) | 2016-12-22 |
JP2018519594A (en) | 2018-07-19 |
EP3311345A1 (en) | 2018-04-25 |
KR20180011333A (en) | 2018-01-31 |
EP3311345A4 (en) | 2019-03-20 |
AU2016277850A1 (en) | 2018-02-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107851057A (en) | The Dynamic Execution of forecast model | |
US10754721B2 (en) | Computer system and method for defining and using a predictive model configured to predict asset failures | |
US10878385B2 (en) | Computer system and method for distributing execution of a predictive model | |
US11036902B2 (en) | Dynamic execution of predictive models and workflows | |
CN107851233A (en) | Local analytics at assets | |
CN108780526B (en) | Disposal of asset localization-based predictive models | |
CN109716356A (en) | Abnormality detection in multivariate data | |
US10254751B2 (en) | Local analytics at an asset | |
CA3032946A1 (en) | Computer architecture and method for recommending asset repairs | |
CN108463806A (en) | Computer Architecture and method for changing data acquisition parameters based on prediction model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1251701 Country of ref document: HK |
|
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180327 |
|
WD01 | Invention patent application deemed withdrawn after publication |