CN108363836A - Multi-state adaptive industrial robot health degree appraisal procedure and system - Google Patents
Multi-state adaptive industrial robot health degree appraisal procedure and system Download PDFInfo
- Publication number
- CN108363836A CN108363836A CN201810045777.8A CN201810045777A CN108363836A CN 108363836 A CN108363836 A CN 108363836A CN 201810045777 A CN201810045777 A CN 201810045777A CN 108363836 A CN108363836 A CN 108363836A
- Authority
- CN
- China
- Prior art keywords
- health degree
- state
- robot
- characteristic
- data
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- 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
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Abstract
The present invention is a kind of industrial robot health degree appraisal procedure and system that multi-state is adaptive, including：Obtain each joint running state data of industrial robot, work information data and design parameter data；Running state data is handled, state property data base is obtained, matching work information data obtain work information database；State feature and duty parameter correlation model are established, various dimensions self-adaptive processing is carried out to state characteristic by correlation model, obtains the characteristic of operating mode desensitization；Each articulation health degree index is calculated based on operating mode desensitization characteristic；Robot complete machine health degree is calculated and assessed based on each articulation health degree index；Based on complete machine health degree, work information data and design parameter data prediction robot remaining life.By the method for the invention and system, the accuracy of the assessment of industrial robot health degree and remaining life prediction can be effectively improved, and model and system is continuously improved in automatic measure on line precision and fitness can be passed through.
Description
Technical field
The present invention relates to robot health control and life prediction fields, more particularly to a kind of work that multi-state is adaptive
Industry robot health degree appraisal procedure and system.
Background technology
In recent years, " machine substitute human labor " pattern at home manufacturing industry to modernization, automation, intelligent transition and upgrade process
In it is rapidly growing.Along with the rapid growth of robot usage quantity at home, thing followed operation, management, safety, dimension
The problems such as shield, repair, becomes increasingly serious.On the automatic production line based on robot, due to all kinds of failures of robot
Caused hang-up, or because performance degradation leads to problems such as product quality fluctuate, can all cause huge economic loss.Cause
This, how effectively assessing the health status of robot and capable of making effective preventive maintenance measure in time just seems heavy to closing
It wants.
Predictive maintenance is with health control (PHM) technology by monitoring the operation shape of robot and its component on production line
On the one hand state can accurately perceive the performance state of robot itself in conjunction with effective health degree appraisal procedure, be " condition maintenarnce "
Reliable characteristic evidences are provided, effectively solve the problems, such as to repair insufficient in traditional mode and are repaired excessive；On the other hand, it can be based on
The indexs such as health degree predict state Advanced management, avoid disorderly closedown, ensure the stable operation of production system, and then realize industry
The Value creation target of sheet, synergy, steady matter drops on production line.In this technology, the assessment of health degree is played to hold and be opened
Under key effect, so the validity of its appraisal procedure, accuracy and reliability are the most important thing.
However, robot application field is extensive, involved job task is also complicated various, it is caused often to be operated in
Under work condition environment complicated and changeable.The information such as the operating mode of variation such as homework type, the speed of service, load can be coupling in robot
In running state parameter, its true performance state and changing rule are covered.Traditional robot health degree appraisal procedure does not have
The operating condition data for considering robot, can not accurately reflect the true health status of robot itself, therefore obtained assessment
As a result differ larger with actual conditions.
Invention content
The present invention is directed to the problem of health degree accurate evaluation under industrial robot multi-state situation in the prior art, provides
A kind of industrial robot health degree appraisal procedure and system that multi-state is adaptive.
A kind of industrial robot health degree appraisal procedure that multi-state is adaptive, includes the following steps：
Obtain running state data, work information data and the robot relevant design parameter number in each joint of industrial robot
According to handling running state data, obtain state property data base, by giving state characteristic storehouse matching corresponding work
Condition information data obtains work information database；
Feature and duty parameter correlation model are established by work information database and state property data base, passes through pass
Gang mould type carries out various dimensions self-adaptive processing to the characteristic in the state property data base, obtains the feature of operating mode desensitization
Data；
The characteristic of operating mode desensitization is standardized, and obtains each articulation health degree in conjunction with Furthest Neighbor and refers to
Number；
According to each articulation health degree index and preset each articulation health degree threshold calculations robot complete machine health degree, and lead to
It crosses robot complete machine health degree and total evaluation is carried out to the health status of robot；
According to the complete machine health degree, work information data and design parameter data to the remaining life of robot into
Row prediction.
As a kind of embodiment, it is described to running state data carry out processing refer to the running state data into
Row feature extraction and feature selecting；
The feature extraction includes the essential characteristic extraction of signal time domain and frequency domain and, resolution process transformed through signal
Further Feature Extraction afterwards；
The feature selecting refers to will have the feature of clear symbolical meanings or tendency, in conjunction with its physical significance or demand
It is selected, the state characteristic needed；If the feature without clear symbolical meanings or tendency need to be selected, in conjunction with it is main at
Point analysis method dimensionality reduction is screened, the state characteristic needed.
As a kind of embodiment, it is described by work information database and state property data base establish feature with
It is adaptive to carry out various dimensions by correlation model to the characteristic in the state property data base for duty parameter correlation model
Processing, the characteristic specific steps for obtaining operating mode desensitization include：
Characteristic in work information database is resolved into mutually independent multiple duty parameter dimensions, and in difference
The correlation model that each feature and duty parameter in property data base are established under dimension, is denoted as the first correlation model；
Selected characteristic is matched under corresponding duty parameter dimension, and according to establishing the first correlation model in operating mode
The influence eliminated to operating mode is compensated under parameter dimensions, obtains the characteristic of operating mode desensitization.
As a kind of embodiment, the characteristic to operating mode desensitization is standardized, and combines
The specific steps that Furthest Neighbor obtains each articulation health degree index include：
By the Euclidean distance D of the selected feature vector of standardized characteristic calculating to healthy reference vector, here,
Health reference vector feature vector selected when being the manufacture in original steady-state when being dispatched from the factory by robot is corresponding
Standardized feature Value Data, which takes, to be worth to；
Normalized function is obtained by Sigmoid functional transformations, the Euclidean distance D is returned by normalized function
One change is handled, and obtains the health degree index in each joint.
It is described that normalized function, the normalization letter are obtained by Sigmoid functional transformations as a kind of embodiment
Number is expressed as：
Wherein, w is scale parameter, and t is smoothing parameter, and HI indicates the health degree index in each joint.
It is described according to each articulation health degree index and preset each articulation health degree threshold value meter as a kind of embodiment
Robot complete machine health degree is calculated, and the specific of total evaluation is carried out to the health status of robot by robot complete machine health degree
Operation is：
By each articulation health degree index construction health radar map, and is combined and preset according to the area of healthy radar map
Health degree threshold calculations complete machine health degree, complete machine health degree indicate it is as follows：
Wherein, HT is preset health degree threshold value, and A is the current state health degree radar area of pictural surface, A1It is strong for all joints
The area of health degree radar map when Kang Du is 1, j=1,2 ... indicate the joint number of robot.
It is described according to the complete machine health degree, work information data and design parameter data as a kind of embodiment
The specific steps predicted the remaining life of robot include：
It establishes between remaining life and complete machine health degree and closes according to complete machine health degree and the design parameter data
Gang mould type is denoted as the second correlation model；
In conjunction with all kinds of operating modes to the influence degree of service life and combine design parameter data, establish aging effects degree with
Correlation model between duty parameter is denoted as third correlation model；
The revised remaining life of duty parameter is established according to complete machine health degrees of data and in conjunction with two correlation models
Prediction model predicts the remaining life under actual condition.
It is further comprising the steps of as a kind of embodiment：
Update step：State property data base and matched work information database can be according to the operation shapes of robot
The increase of state data and work information data and constantly update；
Self study step：It can self study and the first correlation model of optimization, the second correlation model and third correlation model.
A kind of industrial robot health degree assessment system that multi-state is adaptive, comprises the following modules：
Data acquisition module：Obtain running state data, work information data and the robot in each joint of industrial robot
Relevant design supplemental characteristic；
Data processing module：Running state data is handled, state property data base is obtained, by giving state feature
The corresponding work information data of database matching, obtain work information database；
Operating mode self-adaptive processing module：Feature and operating mode are established by work information database and state property data base
Parameter association model carries out various dimensions to the characteristic in the state property data base by correlation model and adaptively locates
Reason obtains the characteristic of operating mode desensitization；
Each articulation health degree computing module：To the operating mode desensitization characteristic be standardized, and combine away from
Each articulation health degree index is obtained from method；
Complete machine health degree evaluation module：According to each articulation health degree index and preset each articulation health degree threshold calculations machine
Device people's complete machine health degree, and total evaluation is carried out to the health status of robot by robot complete machine health degree；
Remaining life prediction module：According to the complete machine health degree, work information data and design parameter data pair
The remaining life of robot is predicted；
Update self-learning module：State property data base and matched work information database can be according to robots
The increase of running state data and work information data and constantly update；It can self study and the first correlation model of optimization, the second pass
Gang mould type and third correlation model.
The present invention has significant technique effect as a result of above technical scheme：
A kind of industrial robot health degree appraisal procedure and system that multi-state is adaptive is provided through the invention, for work
Industry robot operating mode situation complicated and changeable under different application scene is associated model adaptation processing and realizes that operating mode is de-
It is quick, and the predicting residual useful life that the health degree based on Euclidean distance method calculates, assesses and combine operating mode is completed, to effectively improve
Industrial robot health degree is assessed and the accuracy of remaining life prediction, and model can be continuously improved by automatic measure on line
And the precision and fitness of system.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention without having to pay creative labor, may be used also for those of ordinary skill in the art
With obtain other attached drawings according to these attached drawings.
Fig. 1 is the overall flow schematic diagram of the present invention；
Fig. 2 is the idiographic flow schematic diagram of Fig. 1；
Fig. 3 is the overall system structure schematic diagram of the present invention；
Fig. 4 is the corresponding duty parameter variation of time series of the multi-state sample designed in the specific embodiment of the invention
Figure；
Before and after Fig. 5 is the RMS value operating mode self-adaptive processing in industrial robot third joint in the specific embodiment of the invention
Comparison diagram；
Fig. 6 a- Fig. 6 f are health of two industrial robots under multi-state to using duration different in specific embodiment
Spend Evaluated effect comparison；
Fig. 7 is the health degree radar map using two different industrial robots of duration under certain operating mode in specific embodiment
Comparison；
Fig. 8 is health degree radar of the industrial robot before and after certain operating mode Imitating failure in the specific embodiment of the invention
Figure comparison；
Fig. 9 is remaining life prediction result of the industrial robot under multi-state in the specific embodiment of the invention.
Specific implementation mode
With reference to embodiment, the present invention is described in further detail, following embodiment be explanation of the invention and
The invention is not limited in following embodiments.
Embodiment 1：
A kind of industrial robot health degree appraisal procedure that multi-state is adaptive, as shown in Figure 1, including the following steps：
S1, the running state data for obtaining each joint of industrial robot, work information data and robot relevant design ginseng
Number data, handle running state data, obtain state property data base, by giving state characteristic storehouse matching corresponding
Work information data, obtain work information database；
S2, feature and duty parameter correlation model are established by work information database and state property data base, led to
It crosses correlation model and various dimensions self-adaptive processing is carried out to the characteristic in the state property data base, obtain operating mode desensitization
Characteristic；
S3, the characteristic of operating mode desensitization is standardized, and each articulation health is obtained in conjunction with Furthest Neighbor
Spend index；
S4, according to each articulation health degree index and preset each articulation health degree threshold calculations robot complete machine health degree,
And total evaluation is carried out to the health status of robot by robot complete machine health degree；
S5, according to the complete machine health degree, work information data and design parameter data is used to the residue of robot the longevity
Life is predicted.
Further, in step sl, it is described to running state data carry out processing refer to the operating status number
According to progress feature extraction and feature selecting；
The feature extraction includes the essential characteristic extraction of signal time domain and frequency domain and, resolution process transformed through signal
Further Feature Extraction afterwards；
The feature selecting refers to will have the feature of clear symbolical meanings or tendency, in conjunction with its physical significance or demand
It is selected, the state characteristic needed；If the feature without clear symbolical meanings or tendency need to be selected, in conjunction with it is main at
Point analysis method dimensionality reduction is screened, the state characteristic needed.
In step s 2, described that feature and duty parameter are established by work information database and state property data base
Correlation model carries out various dimensions self-adaptive processing to the characteristic in the state property data base by correlation model, obtains
To operating mode desensitize characteristic specific steps include：
S21, the characteristic in work information database resolves into mutually independent multiple duty parameter dimensions, and
The correlation model that each feature and duty parameter in property data base are established under different dimensions, is denoted as the first correlation model, is
Convenient for description, here, by MijIt is expressed as the first correlation model, wherein i=1,2 ..., P is P selected features, j
=1,2 ..., K is K duty parameter dimension；
S22, selected characteristic is matched under corresponding duty parameter dimension, here, duty parameter dimension has K,
And the influence eliminated to operating mode is compensated under K duty parameter dimension according to the first correlation model is established, it is de- to obtain operating mode
Quick characteristic, is embodied as：
Wherein, i=1,2 ..., P is P selected features, FiFor i-th of selected characteristic value, VjIt ties up and corresponds to for jth
Duty parameter,For value of i-th of selected feature after removing all operating modes and influencing, here,Operating mode as to obtain takes off
Quick characteristic.
In step s3, the characteristic to operating mode desensitization is standardized, and is obtained in conjunction with Furthest Neighbor
Specific steps to each articulation health degree index include：
S31, the characteristic of operating mode desensitization is standardized, obtains standardized characteristic, here,
The purpose being standardized is to eliminate dimension and magnitude to influence, and obtains the standardized feature value S after operating mode influencesi,
In, (i=1,2 ..., P), further, standardization processing method has selected deviation standardization, makes standardized feature value
It falls in [0,1] section：
Wherein, max and min is respectively the maximum value and minimum value of sample number；SiIndicate standardized feature value,It is i-th
Value of the selected feature after removing all operating modes and influencing.
S32, by standardized characteristic calculate sample vector to health reference vector Euclidean distance D, here, institute
State healthy reference vector yBIt is the corresponding standardized feature value of sample set when being dispatched from the factory by robot in original steady-state
Data, which take, to be worth to, healthy reference vector yBIt is expressed as：
Wherein, the sample set of original steady-state is tieed up for n × P, and n representative sample numbers, P is characterized dimension；
Euclidean distance D is expressed as：
Wherein, y indicates sample vector, yBIndicate healthy reference vector；Here, sample vector be selected feature to
Amount, sample are combined into one group of feature vector selected when manufacture；
S33, normalized function is obtained by Sigmoid functional transformations, by normalized function to the Euclidean distance D into
Row normalized obtains the health degree index in each joint.
In the present embodiment, described to obtain normalized function by Sigmoid functional transformations, the normalized function indicates
For：
Wherein, w is scale parameter, and t is smoothing parameter, and HI indicates the health degree index in each joint.
In step s 4, described according to each articulation health degree index and preset each articulation health degree threshold calculations robot
Complete machine health degree, and be to the concrete operations of the health status progress total evaluation of robot by robot complete machine health degree：
By each articulation health degree index construction health radar map, and is combined and preset according to the area of healthy radar map
Health degree threshold calculations complete machine health degree, complete machine health degree indicate it is as follows：
Wherein, HT is preset health degree threshold value, and A is the current state health degree radar area of pictural surface, A1It is strong for all joints
The area of health degree radar map when Kang Du is 1, j=1,2 ... indicate the joint number of robot.
More specifically, in step s 5, it is described according to the complete machine health degree, work information data and design parameter data
The specific steps predicted the remaining life of robot include：
S51, established according to complete machine health degree and the design parameter data remaining life and complete machine health degree it
Between correlation model, be denoted as the second correlation model, use MRULIt indicates；
S52, to the influence degree of service life and design parameter data are combined in conjunction with all kinds of operating modes, establishes aging effects
Correlation model between degree and duty parameter, is denoted as third correlation model, uses IMjIt indicates, j=1,2 ..., K is K operating mode
Parameter dimensions；
S53, the revised remaining use of duty parameter is established according to complete machine health degrees of data and in conjunction with two correlation models
Life Prediction Model predicts the remaining life under actual condition, more specifically, the remaining longevity can carried out in turn
According to operating mode amendment when life prediction, keep interaction prediction value more accurate.It the remaining life interaction prediction and is repaiied according to operating mode
Correction method is as follows：
Wherein, RUL is the revised prediction remaining life of duty parameter, RUL0Not consider that operating mode influences basic
Predict remaining life, IFjDisturbance degree of the duty parameter dimension to service life is tieed up for jth, H is health obtained by step S4
Degree, ParmiFor robot relevant design parameter, VjCorresponding duty parameter is tieed up for jth, wherein j=1,2 ..., K are K work
Condition parameter dimensions, i=1,2 ..., be i-th selected by feature.
In this embodiment, further comprising the steps of：
Update step：State property data base and matched work information database can be according to the operation shapes of robot
The increase of state data and work information data and constantly update；
Self study step：It can self study and the first correlation model M of optimizationij, the second correlation model MRULAnd third is associated with mould
Type IMj。
A kind of industrial robot health degree assessment system that multi-state is adaptive, as shown in figure 3, comprising the following modules：
Data acquisition module 1：Obtain running state data, work information data and the robot in each joint of industrial robot
Relevant design supplemental characteristic；
Data processing module 2：Running state data is handled, state property data base is obtained, by giving state special
The corresponding work information data of database matching are levied, work information database is obtained；
Operating mode self-adaptive processing module 3：Feature and work are established by work information database and state property data base
Condition parameter association model carries out various dimensions to the characteristic in the state property data base by correlation model and adaptively locates
Reason obtains the characteristic of operating mode desensitization；
Each articulation health degree computing module 4：To the operating mode desensitization characteristic be standardized, and combine away from
Each articulation health degree index is obtained from method；
Complete machine health degree evaluation module 5：According to each articulation health degree index and preset each articulation health degree threshold calculations
Robot complete machine health degree, and total evaluation is carried out to the health status of robot by robot complete machine health degree；
Remaining life prediction module 6：According to the complete machine health degree, work information data and design parameter data pair
The remaining life of robot is predicted；
Update self-learning module 7：State property data base and matched work information database can be according to robots
Running state data and work information data increase and constantly update；It can self study and the first correlation model of optimization, second
Correlation model and third correlation model.
Specific embodiment
Below by way of in conjunction with real case come tell about the present invention technique effect：
Industrial robot described in this embodiment includes two with six axis joint type industrial robot of model, the two use
The time limit is different, other design parameters are considered as identical.
Referring to Figures 1 and 2, the relevant design supplemental characteristic of the industrial robot of acquisition includes its design service life, volume
Constant speed degree, nominal load, service life, theoretical Decline law parameter etc., the robot work information data include mainly
Its real-time overall operation speed and load, each joint running state data of robot include mainly torque and the electricity in each joint
Flow data；State property data base is extracted by the torque or current data in each joint through the extraction of signal temporal signatures and frequency domain character
And feature selecting obtains.Feature selecting is mainly selected according to signal intensity feature, mainly selects to be characterized as root mean square
It is worth (RMS), rectified mean value (ARV), peak-to-peak value (P-P) and standard deviation (STD), these are all the data handling procedures of early period,
Subsequently to model and prediction is prepared.
In order to verify the adaptive industrial robot health evaluating method of multi-state proposed by the invention, in the present embodiment
Various working caused by speed and load variation, the time series of the sample are devised for six axis joint type industrial robots
Corresponding duty parameter variation, as shown in Figure 4.
The RMS that Fig. 5 gives third joint in the time series of the sample carries out pair before and after multi-state self-adaptive processing
Than, it is seen then that before operating mode self-adaptive processing, temporal signatures RMS value by the speed of service and institute it is loaded influence it is bigger, it is whole
From the point of view of, it is all to show a rising trend with speed and load, and fluctuate and increase when speed is higher.Using institute's extracting method of the present invention
After carrying out adaptive operating mode processing, RMS value realizes the desensitization of operating mode substantially, but fluctuation letter when remaining high speed simultaneously
Breath, in Fig. 5, the diagram of top is indicated without passing through diagram when operating mode self-adaptive processing, it is shown that temporal signatures RMS value by
The loaded influence of the speed of service and institute, the diagram when diagram expression of lower section is by operating mode self-adaptive processing, RMS value are basic
Fluctuation information when realizing the desensitization of operating mode, but remaining high speed simultaneously.
Further to same model, same operating mode but service life different two six axis joint types industry in the present embodiment
Robot has carried out health degree assessment comparison, and the corresponding duty parameter variation of two sample time-series is as shown in figure 4, Fig. 6 a-
Fig. 6 e are the articulation health degree index contrast marked as J1-J6 through gained after step S1-S3, wherein new purchase is used only 6
Month each articulation health degree index of industrial robot substantially close to initial health a reference value, and the health degree when operating mode changes
It is worth basicly stable, is in a good state of health；The other is using 30 months robots in addition to the joint marked as J4 other close
The health degree of section has different degrees of decline, and health when it increases with service life can be intuitively embodied by these figures
Degree loss.
Fig. 7 be being calculated by health degree index marked as the joint of J1-J6 in two samples obtain later with it is each
The corresponding health degree radar map of state, health degree radar map, which can characterize method proposed by the invention, can effectively realize industry
The adaptive health degree assessment of multi-state of robot.
Further, exist to assess the adaptive industrial robot health degree appraisal procedure of multi-state proposed by the invention
Effect when incipient fault state, by the industrial robot to 6 months service lives marked as J3 in the present embodiment
Joint carry out strength binding, simulate easy malfunction.Fig. 8 is the health before and after simulated failure after step S1-S4
Spend contrast effect, it is seen then that compared to normal condition, marked as J1, J4, J5, J6 articulation health degree substantially without significant change, by mould
What quasi- failure directly affected is decreased obviously marked as J3 articulation health angle value, by simulated failure influence indirectly marked as the joints J2
Health degree value has to be declined by a small margin, and institute's extracting method of the present invention can be tracked effectively and identify the whole and local health of industrial robot
State.
To verify the adaptive industrial robot health evaluating method of multi-state proposed by the invention to industrial robot
The prediction effect of complete machine remaining life, it is shown in Fig. 4 for six axis multi-joint type industrial robots in the present embodiment
It is run under various working parameter, and through predicting that its remaining life under different operating modes, Fig. 9 are pair after step S1-S5
The prediction result answered, it is seen then that the six axis articulated industrial robots for being about 15 years for design service life, in low speed, low
Prediction remaining life under load operation conditions is almost the same with the difference between design service life and actual use duration,
Show the validity of its life prediction, meanwhile, when according to sequence variation operating mode shown in Fig. 4, the residue under corresponding operating mode uses the longevity
Life also can have different degrees of decaying on the basis of base regime according to operating mode, reflect the parameters such as operating mode in actual application environment
Influence to industrial robot remaining life, so as to more precisely predict work according to practical application operating mode and environment etc.
Industry robot remaining life, and carry out more efficiently safeguarding scheduling and predictive maintenance.
Furthermore, it is necessary to illustrate, the specific embodiment described in this specification, the shape of parts and components is named
Title etc. can be different.The equivalent or simple change that all structure, feature and principles according to described in inventional idea of the present invention are done, is wrapped
It includes in the protection domain of patent of the present invention.Those skilled in the art can be to described specific implementation
Example is done various modifications or additions or is substituted by a similar method, without departing from structure of the invention or surmounts this
Range as defined in the claims, is within the scope of protection of the invention.
Claims (9)
1. a kind of industrial robot health degree appraisal procedure that multi-state is adaptive, it is characterised in that include the following steps：
Running state data, work information data and the robot relevant design supplemental characteristic in each joint of industrial robot are obtained,
Running state data is handled, state property data base is obtained, by giving state characteristic storehouse matching corresponding operating mode
Information data obtains work information database；
Feature and duty parameter correlation model are established by work information database and state property data base, by being associated with mould
Type carries out various dimensions self-adaptive processing to the characteristic in the state property data base, obtains the characteristic of operating mode desensitization
According to；
The characteristic of operating mode desensitization is standardized, and each articulation health degree index is obtained in conjunction with Furthest Neighbor；
According to each articulation health degree index and preset each articulation health degree threshold calculations robot complete machine health degree, and pass through machine
Device people's complete machine health degree carries out total evaluation to the health status of robot；
The remaining life of robot is carried out according to the complete machine health degree, work information data and design parameter data pre-
It surveys.
2. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 1, which is characterized in that institute
It refers to carrying out feature extraction and feature selecting to the running state data to state and carry out processing to running state data；
The feature extraction include signal time domain and frequency domain essential characteristic extraction and, resolution process transformed through signal after
Further Feature Extraction；
The feature selecting refers to will have the feature of clear symbolical meanings or tendency, is carried out in conjunction with its physical significance or demand
Selection, the state characteristic needed；If the feature without clear symbolical meanings or tendency need to be selected, in conjunction with principal component point
Analysis method dimensionality reduction is screened, the state characteristic needed.
3. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 2, which is characterized in that institute
It states and feature and duty parameter correlation model is established by work information database and state property data base, pass through correlation model
Various dimensions self-adaptive processing is carried out to the characteristic in the state property data base, obtains the characteristic tool of operating mode desensitization
Body step includes：
Characteristic in work information database is resolved into mutually independent multiple duty parameter dimensions, and in different dimensions
The lower correlation model for establishing each feature and duty parameter in property data base, is denoted as the first correlation model；
Selected characteristic is matched under corresponding duty parameter dimension, and according to establishing the first correlation model in duty parameter
The influence eliminated to operating mode is compensated under dimension, obtains the characteristic of operating mode desensitization.
4. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 3, which is characterized in that institute
It states the characteristic for desensitizing to the operating mode to be standardized, and the tool of each articulation health degree index is obtained in conjunction with Furthest Neighbor
Body step includes：
The characteristic of operating mode desensitization is standardized, standardized characteristic is obtained；
By the Euclidean distance D of the selected feature vector of standardized characteristic calculating to healthy reference vector, here, described
Healthy reference vector is the corresponding standardized feature of feature vector selected under original steady-state when being dispatched from the factory by robot
Data, which take, to be worth to；
Normalized function is obtained by Sigmoid functional transformations, the Euclidean distance D is normalized by normalized function
Processing, obtains the health degree index in each joint.
5. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 4, which is characterized in that institute
It states and normalized function is obtained by Sigmoid functional transformations, the normalized function is expressed as：
Wherein, w is scale parameter, and t is smoothing parameter, and HI indicates the health degree index in each joint.
6. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 5, which is characterized in that institute
It states according to each articulation health degree index and preset each articulation health degree threshold calculations robot complete machine health degree, and passes through machine
The concrete operations that people's complete machine health degree carries out the health status of robot total evaluation are：
Preset be good for is combined by each articulation health degree index construction health radar map, and according to the area of healthy radar map
Kang Du threshold calculations complete machine health degrees, complete machine health degree indicate as follows：
Wherein, HT is preset health degree threshold value, and A is the current state health degree radar area of pictural surface, and A1 is all articulation health degree
The area of health degree radar map when being 1, j=1,2 ... indicate the joint number of robot.
7. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 6, which is characterized in that institute
It states and the remaining life of robot is predicted according to the complete machine health degree, work information data and design parameter data
Specific steps include：
It is established between remaining life and complete machine health degree according to complete machine health degree and the design parameter data and is associated with mould
Type is denoted as the second correlation model；
To the influence degree of service life and design parameter data are combined in conjunction with all kinds of operating modes, establish aging effects degree and operating mode
Correlation model between parameter is denoted as third correlation model；
The revised remaining life prediction of duty parameter is established according to complete machine health degrees of data and in conjunction with two correlation models
Model predicts the remaining life under actual condition.
8. the adaptive industrial robot health degree appraisal procedure of multi-state according to claim 1,3 or 7, feature exist
In further comprising the steps of：
Update step：State property data base and matched work information database can be according to the operating status numbers of robot
It is constantly updated according to the increase with work information data；
Self study step：It can self study and the first correlation model of optimization, the second correlation model and third correlation model.
9. a kind of industrial robot health degree assessment system that multi-state is adaptive, which is characterized in that comprise the following modules：
Data acquisition module：The running state data, work information data and robot for obtaining each joint of industrial robot are related
Design parameter data；
Data processing module：Running state data is handled, state property data base is obtained, by giving state characteristic
The corresponding work information data of storehouse matching, obtain work information database；
Operating mode self-adaptive processing module：Feature and duty parameter are established by work information database and state property data base
Correlation model carries out various dimensions self-adaptive processing to the characteristic in the state property data base by correlation model, obtains
The characteristic to desensitize to operating mode；
Each articulation health degree computing module：The characteristic of operating mode desensitization is standardized, and combines Furthest Neighbor
Obtain each articulation health degree index；
Complete machine health degree evaluation module：According to each articulation health degree index and preset each articulation health degree threshold calculations robot
Complete machine health degree, and total evaluation is carried out to the health status of robot by robot complete machine health degree；
Remaining life prediction module：According to the complete machine health degree, work information data and design parameter data to machine
The remaining life of people is predicted；
Update self-learning module：State property data base and matched work information database can be according to the operations of robot
The increase of status data and work information data and constantly update；It can self study and the first correlation model of optimization, the second association mould
Type and third correlation model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810045777.8A CN108363836B (en) | 2018-01-17 | 2018-01-17 | Multi-working-condition self-adaptive industrial robot health degree assessment method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810045777.8A CN108363836B (en) | 2018-01-17 | 2018-01-17 | Multi-working-condition self-adaptive industrial robot health degree assessment method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108363836A true CN108363836A (en) | 2018-08-03 |
| CN108363836B CN108363836B (en) | 2021-07-20 |
Family
ID=63006363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810045777.8A Active CN108363836B (en) | 2018-01-17 | 2018-01-17 | Multi-working-condition self-adaptive industrial robot health degree assessment method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108363836B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109239082A (en) * | 2018-09-21 | 2019-01-18 | 杭州安脉盛智能技术有限公司 | Tobacco structure quality online test method and system based on machine vision technique |
| CN109711453A (en) * | 2018-12-21 | 2019-05-03 | 广东工业大学 | A kind of equipment dynamical health state evaluating method based on multivariable |
| CN111126822A (en) * | 2019-12-19 | 2020-05-08 | 佛山科学技术学院 | Industrial robot health assessment method and device and storage medium |
| CN112115575A (en) * | 2020-07-29 | 2020-12-22 | 北京奔驰汽车有限公司 | Equipment lubricating oil state evaluation system and method |
| CN113032985A (en) * | 2021-03-11 | 2021-06-25 | 北京必创科技股份有限公司 | Intelligent service life assessment method and device for wireless sensing equipment |
| WO2022041064A1 (en) * | 2020-08-27 | 2022-03-03 | Rethink Robotics Gmbh | Method and apparatus for robot joint status monitoring |
| CN114418042A (en) * | 2021-12-30 | 2022-04-29 | 智昌科技集团股份有限公司 | Industrial robot operation trend diagnosis method based on cluster analysis |
| CN115139336A (en) * | 2021-03-31 | 2022-10-04 | 中国科学院沈阳自动化研究所 | Data acquisition and screening method for industrial robot health monitoring |
| CN117428760A (en) * | 2023-10-10 | 2024-01-23 | 无锡蔚动智能科技有限公司 | Joint module control system and method based on artificial intelligence |
| CN113032985B (en) * | 2021-03-11 | 2024-04-26 | 北京必创科技股份有限公司 | Intelligent evaluation method and device for service life of wireless sensing equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006021268A (en) * | 2004-07-07 | 2006-01-26 | Nissan Motor Co Ltd | Service life time predicting method of robot, operation pattern making method of robot and their program |
| CN102313577A (en) * | 2011-06-24 | 2012-01-11 | 上海大学 | Equipment health state evaluation and recession prediction method based on multi-channel sensing signals |
| CN103558565A (en) * | 2013-10-14 | 2014-02-05 | 中国科学院电工研究所 | Three-leg type magnetic field detection robot |
| CN107016235A (en) * | 2017-03-21 | 2017-08-04 | 西安交通大学 | The equipment running status health degree appraisal procedure adaptively merged based on multiple features |
-
2018
- 2018-01-17 CN CN201810045777.8A patent/CN108363836B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006021268A (en) * | 2004-07-07 | 2006-01-26 | Nissan Motor Co Ltd | Service life time predicting method of robot, operation pattern making method of robot and their program |
| CN102313577A (en) * | 2011-06-24 | 2012-01-11 | 上海大学 | Equipment health state evaluation and recession prediction method based on multi-channel sensing signals |
| CN103558565A (en) * | 2013-10-14 | 2014-02-05 | 中国科学院电工研究所 | Three-leg type magnetic field detection robot |
| CN107016235A (en) * | 2017-03-21 | 2017-08-04 | 西安交通大学 | The equipment running status health degree appraisal procedure adaptively merged based on multiple features |
Non-Patent Citations (2)
| Title |
|---|
| MARTIN FISCHER: ""Interdisciplinary technology assessment of service robots: the psychological/work science perspective"", 《POIESIS PRAX》 * |
| 陈胜军 等: ""工业机器人系统可靠性预测方法研究"", 《南京师范大学学报（工程技术版）》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109239082B (en) * | 2018-09-21 | 2021-01-26 | 杭州安脉盛智能技术有限公司 | Tobacco shred structure quality online detection method and system based on machine vision technology |
| CN109239082A (en) * | 2018-09-21 | 2019-01-18 | 杭州安脉盛智能技术有限公司 | Tobacco structure quality online test method and system based on machine vision technique |
| CN109711453B (en) * | 2018-12-21 | 2022-05-13 | 广东工业大学 | Multivariable-based equipment dynamic health state evaluation method |
| CN109711453A (en) * | 2018-12-21 | 2019-05-03 | 广东工业大学 | A kind of equipment dynamical health state evaluating method based on multivariable |
| CN111126822A (en) * | 2019-12-19 | 2020-05-08 | 佛山科学技术学院 | Industrial robot health assessment method and device and storage medium |
| CN111126822B (en) * | 2019-12-19 | 2023-04-28 | 佛山科学技术学院 | Industrial robot health assessment method, device and storage medium |
| CN112115575A (en) * | 2020-07-29 | 2020-12-22 | 北京奔驰汽车有限公司 | Equipment lubricating oil state evaluation system and method |
| WO2022041064A1 (en) * | 2020-08-27 | 2022-03-03 | Rethink Robotics Gmbh | Method and apparatus for robot joint status monitoring |
| CN113032985A (en) * | 2021-03-11 | 2021-06-25 | 北京必创科技股份有限公司 | Intelligent service life assessment method and device for wireless sensing equipment |
| CN113032985B (en) * | 2021-03-11 | 2024-04-26 | 北京必创科技股份有限公司 | Intelligent evaluation method and device for service life of wireless sensing equipment |
| CN115139336A (en) * | 2021-03-31 | 2022-10-04 | 中国科学院沈阳自动化研究所 | Data acquisition and screening method for industrial robot health monitoring |
| CN114418042A (en) * | 2021-12-30 | 2022-04-29 | 智昌科技集团股份有限公司 | Industrial robot operation trend diagnosis method based on cluster analysis |
| CN114418042B (en) * | 2021-12-30 | 2022-07-22 | 智昌科技集团股份有限公司 | Industrial robot operation trend diagnosis method based on cluster analysis |
| CN117428760A (en) * | 2023-10-10 | 2024-01-23 | 无锡蔚动智能科技有限公司 | Joint module control system and method based on artificial intelligence |
| CN117428760B (en) * | 2023-10-10 | 2024-03-29 | 无锡蔚动智能科技有限公司 | Joint module control system and method based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108363836B (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108363836A (en) | Multi-state adaptive industrial robot health degree appraisal procedure and system | |
| CN108280543A (en) | The adaptive equipment health degree appraisal procedure of operating mode based on classification Regression Model | |
| CN104992270B (en) | Power transmission and transformation equipment state overhauling aid decision-making system and method | |
| JP6598872B2 (en) | System and method for tuning a process model | |
| US10839115B2 (en) | Cleansing system for a feed composition based on environmental factors | |
| CN106598791B (en) | Industrial equipment fault preventive identification method based on machine learning | |
| CN100470416C (en) | Power plant thermal equipment intelligent state diagnosing and analyzing system | |
| CN104390657B (en) | A kind of Generator Unit Operating Parameters measurement sensor fault diagnosis method and system | |
| EP3278278A1 (en) | Advanced data cleansing system and method | |
| US9261862B2 (en) | Automation management system and method | |
| DE102017000536A1 (en) | Cell controller for detecting a cause of an abnormality in a manufacturing machine | |
| EP3278277A1 (en) | Data cleansing system and method for inferring a feed composition | |
| CN105793789B (en) | The computer implemented method and system that automatic monitoring and state for all processes section in process unit determine | |
| WO2005054968A1 (en) | Intelligent system for detection of process status, process fault and preventive maintenance | |
| CN107291830A (en) | A kind of creation method of equipment knowledge base | |
| CN109155018A (en) | There is the MPC without constraint dependent variable for KPI performance evaluation | |
| CN113256102B (en) | High-risk technological process risk control method and system | |
| TWI738016B (en) | Intelligent learning mechanism and adaptive regulation system for liquid gas filling | |
| CN115422635A (en) | Steel structure engineering monitoring method, system, equipment and storage medium based on BIM | |
| CN106567964A (en) | Flow control valve monitoring method based on DCS data | |
| CN111553581B (en) | Equipment maintainability evaluation model based on entropy value | |
| CN102419593A (en) | Sensor fault diagnosis method based on data mining | |
| CN116037705A (en) | Real-time monitoring system for working state of cold stamping die | |
| CN113487149B (en) | Welding spot abnormity identification system and method based on Catboost K-fold cross verification | |
| CN112288116A (en) | Industrial big data system and method for production and manufacturing process optimization management |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |