CN114881269B - Abnormity detection method and device for material conveying pipeline - Google Patents
Abnormity detection method and device for material conveying pipeline Download PDFInfo
- Publication number
- CN114881269B CN114881269B CN202210817516.XA CN202210817516A CN114881269B CN 114881269 B CN114881269 B CN 114881269B CN 202210817516 A CN202210817516 A CN 202210817516A CN 114881269 B CN114881269 B CN 114881269B
- Authority
- CN
- China
- Prior art keywords
- set time
- time period
- material conveying
- conveying pipeline
- pipeline
- 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.)
- Active
Links
- 239000000463 material Substances 0.000 title claims abstract description 332
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000011156 evaluation Methods 0.000 claims abstract description 122
- 230000002159 abnormal effect Effects 0.000 claims abstract description 94
- 230000005856 abnormality Effects 0.000 claims abstract description 49
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 141
- 238000005299 abrasion Methods 0.000 claims description 108
- 239000013598 vector Substances 0.000 claims description 53
- 238000000034 method Methods 0.000 claims description 30
- 238000004364 calculation method Methods 0.000 claims description 21
- 239000000126 substance Substances 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 4
- 230000000875 corresponding effect Effects 0.000 description 90
- 230000008569 process Effects 0.000 description 23
- 230000006870 function Effects 0.000 description 21
- 238000012546 transfer Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000006229 carbon black Substances 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000000746 purification Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000013024 troubleshooting Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2462—Approximate or statistical queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Remote Sensing (AREA)
- Pipeline Systems (AREA)
Abstract
The invention relates to the technical field of data identification, in particular to an anomaly detection method and device for a material conveying pipeline, which comprises the following steps: by acquiring the wear information of each position of each connecting part of the material conveying pipeline to be detected in the current first set time period and in the current frontNAnd carrying out data identification and processing on the basis of the position wear information in each first set time period, so as to obtain a predicted value of the connection part wear degree evaluation value and an actual evaluation value of the connection part wear degree of the material conveying pipeline to be detected in the current first set time period, and finally determining whether the connection part wear abnormality occurs in the material conveying pipeline to be detected. The abnormity detection method and the abnormity detection device provided by the invention adopt a data identification and processing mode, can accurately determine whether the material conveying pipeline is abnormal in advance, and effectively solve the problem that the existing discovery of the abnormity of the material conveying pipeline is lagged.
Description
Technical Field
The invention relates to the technical field of data identification, in particular to an anomaly detection method and device for a material conveying pipeline.
Background
Pneumatic transport has been a material transportation means for over one hundred years and has the following unique advantages compared with traditional mechanical transportation and vehicle transportation: air is used as a medium, a pipeline transportation mode is totally closed, secondary pollution is avoided, and the layout of a transportation route can be flexibly changed according to the site. However, in the pneumatic conveying process, the phenomena of pipe blockage and even material leakage can be caused inside the conveying pipeline under the influence of some factors. Due to the lack of reliable monitoring means, in many cases, the pipeline is found to be abnormal only after the pipeline is blocked or leaked, and the pipeline has certain hysteresis, which affects the working efficiency of the pipeline and causes great difficulty in maintaining the pipeline.
Disclosure of Invention
The invention aims to provide an anomaly detection method and device for a material conveying pipeline, which are used for solving the problem that the existing discovery of anomaly of the material conveying pipeline is lagged.
In order to solve the technical problem, the invention provides an anomaly detection method for a material conveying pipeline, which comprises the following steps:
acquiring the wear information of each position of each connecting part of a material conveying pipeline to be detected in a current first set time period and the current front position of each connecting partNWear information of each position within a first set time period;
according to the wear information of each connecting part at each position in the current first set time period and the current previous position of each connecting partNThe wear degree evaluation value of each connecting part in the current first set time period and the wear degree evaluation value before the current time period are calculated according to the wear information of each position in the first set time periodNEvaluating the wear degree of the material conveying pipeline to be detected in the first set time period so as to obtain the actual evaluation value of the wear degree of the connecting part of the material conveying pipeline to be detected in the current first set time period and the current wear degree of the connecting part of the material conveying pipeline to be detected in the previous set time periodNActually evaluating the abrasion degree of the connecting part in a first set time period;
the material conveying pipeline to be detected is arranged in front of the current positionNInputting the actual evaluation value of the abrasion degree of the connecting part in the first set time period into a material conveying pipeline connecting part abrasion degree prediction network to obtain a predicted value of the evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period;
and comparing the predicted value of the connection part abrasion degree evaluation value of the material conveying pipeline to be detected in the current first set time period with the actual evaluation value of the connection part abrasion degree, and determining whether the connection part abrasion abnormality occurs in the material conveying pipeline to be detected.
Further, the wear degree evaluation value of each connecting part in the current first set time period and before the current time are calculatedNThe calculation formula corresponding to the wear degree evaluation value in the first set time period is as follows:
wherein the content of the first and second substances,
is as followsjThe wear degree evaluation value of each connecting part in the current first set time period or the wear degree evaluation value in each previous first set time period,
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Average value of all elements in (1) orjThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The average value of all the elements in (a),
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Median or second of all elements injThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The median value of all the elements in (a),
the function is a hyperbolic tangent function.
Further, further obtainThe actual evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period and before the current timeNThe step of actually evaluating the degree of wear of the connecting portion for each first set period of time includes:
according to the abrasion degree evaluation value of each connecting part in the current first set time period and before the current timeNThe average value of the wear degree evaluation values of all the connecting parts in the current first set time period and the average value of the wear degree evaluation values in each previous first set time period are respectively calculated, so that the actual evaluation value of the wear degree of the connecting parts of the material conveying pipeline to be detected in the current first set time period and the actual evaluation value of the wear degree of the connecting parts in the current previous first set time period of the material conveying pipeline to be detected are correspondingly obtainedNAnd actually evaluating the abrasion degree of the connecting part in the first set time period.
Further, the step of obtaining the prediction network of the degree of wear of the material conveying pipeline connecting part includes:
acquiring abrasion information of each position of each connecting part of each first material conveying pipeline sample in each historical first set time period, wherein each first material conveying pipeline sample is the same as the material conveying pipeline to be detected in configuration;
determining a wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period according to the wear information of each position of each connecting part of each first material conveying pipeline sample in each historical first set time period, and further obtaining an actual wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period;
screening each first material conveying pipeline sample according to the actual evaluation value of the abrasion degree of the connecting part of each first material conveying pipeline sample in each historical first set time period, so as to obtain each screened first material conveying pipeline sample;
constructing a material conveying pipeline connecting part abrasion degree prediction network, and determining training samples of the network and a weight value corresponding to each training sample according to an actual evaluation value of the abrasion degree of the connecting part of each screened first material conveying pipeline sample in each historical first set time period;
and constructing a loss function of the network according to the training samples of the network and the weight values corresponding to each training sample, and training the network according to the training samples and the loss function of the network, so as to obtain a trained prediction network of the abrasion degree of the connecting part of the material conveying pipeline.
Further, the step of determining the weight value corresponding to each training sample includes:
determining each relevant training sample of each training sample, wherein the historical first set time period corresponding to the relevant training sample is the historical first set time period adjacent to the historical first set time period corresponding to the corresponding training sample;
respectively calculating the absolute value of the difference between the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample and the actual evaluation value of the abrasion degree of the connecting part corresponding to each relevant training sample;
and calculating the average value of the absolute values of all the difference values corresponding to each training sample so as to determine the weight value corresponding to each training sample.
Further, the calculation formula corresponding to the loss function of the network is as follows:
wherein the content of the first and second substances,
as a function of the losses of the network,
is as followstThe predicted value of the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample,
is as followstThe label value of the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample,
is as followstThe weight value corresponding to each training sample,Nthe number of training samples that each time the network is trained.
Further, the method further comprises:
acquiring each position air pressure value and each position humidity value of a material conveying pipeline to be detected at each set moment in a current second set time period and each position air pressure value and each position humidity value of each second material conveying pipeline sample at each set moment in a historical second set time period, wherein the configuration of each second material conveying pipeline sample is the same as that of the material conveying pipeline to be detected, and the pipeline work state abnormality of each second material conveying pipeline sample does not occur in the historical second set time period;
determining the pipeline abnormal index at each set time in the current second set time period and the historical second set time period according to each position air pressure value and each position humidity value at each set time in the current second set time period and the historical second set time period, so as to obtain the corresponding pipeline abnormal index vector in the current second set time period and the historical second set time period;
determining the radius of the hyperplane of the abnormal index vector of the pipeline and the distance from the abnormal index vector of the pipeline corresponding to the current second set time period to the radius of the hyperplane of the abnormal index vector of the pipeline according to the abnormal index vector of the pipeline corresponding to the current second set time period and the historical second set time period, and further prejudging whether the pipeline work state abnormality occurs to the material conveying pipeline to be detected in the current second set time period;
if the pipeline is judged to be abnormal in working state in advance, the material conveying pipeline to be detected is continuously acquired in the futureGEach set time in the second set time periodDetermining the air pressure value and the humidity value of each position in the futureGThe abnormal index of the pipeline at each set time in the second set time period is obtained in the futureGCorresponding pipeline abnormal index vectors in a second set time period;
according to the futureGDetermining the corresponding pipeline abnormal index vector in the second set time period and the historical second set time period in the futureGDetermining the distance from the corresponding pipeline abnormal index vector to the pipeline abnormal index vector hyperplane radius in a second set time period, thereby determining the comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected;
and actually judging whether the pipeline work state abnormity occurs in the material conveying pipeline to be detected according to the comprehensive pipeline abnormity index corresponding to the material conveying pipeline to be detected.
Further, the calculation formula corresponding to the pipeline abnormality index at each set time in the current second set time period is determined as follows:
wherein the content of the first and second substances,
is at the first within the current second set period of timeiThe abnormal index of the pipeline at each set moment,
a sequence formed by the air pressure values of the positions at each set time in the current second set time period
The standard deviation of all the elements in (A),
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The average value of all the elements in (A),
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The maximum position humidity value among all the elements in (c),
the function is a hyperbolic tangent function.
Further, the calculation formula corresponding to the comprehensive pipeline abnormity index corresponding to the material conveying pipeline to be detected is determined as follows:
wherein the content of the first and second substances,
is a comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected,
is the pipe abnormal index vector hyperplane radius,
to be in the futuregAnd G is the total number of the future second set time periods.
The invention also provides an abnormality detection device for the material conveying pipeline, which comprises a processing unit and a storage unit, wherein the storage unit is used for storing instructions, and the processing unit is used for executing the instructions stored in the storage unit so as to realize the abnormality detection method for the material conveying pipeline.
The invention has the following beneficial effects: by acquiring the wear information of each position of each connecting part of the material conveying pipeline to be detected in the current first set time period and in the current frontNAnd carrying out data identification and processing on the basis of the position wear information in each first set time period so as to obtain a predicted value of the connection part wear degree evaluation value and an actual evaluation value of the connection part wear degree of the material conveying pipeline to be detected in the current first set time period, comparing the predicted value of the connection part wear degree evaluation value and the actual evaluation value of the connection part wear degree, and finally determining whether the material conveying pipeline to be detected is abnormal in connection part wear. According to the invention, based on the abrasion information of each position of each connecting part of the material conveying pipeline to be detected, by adopting a data identification and processing mode, before the pipeline is not subjected to obvious abnormal abrasion of the connecting part, whether the material conveying pipeline is subjected to abnormal abrasion can be accurately determined in advance, the problem that the existing discovery of the abnormal condition of the material conveying pipeline is lagged is effectively solved, and the maintenance work of the pipeline is facilitated.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flowchart illustrating an embodiment of the present invention for determining whether a wear anomaly occurs in a connection portion of a material conveying pipeline;
fig. 2 is a flowchart illustrating an embodiment of determining whether a pipeline work state abnormality occurs in a material conveying pipeline.
Detailed Description
To further explain the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the embodiments, structures, features and effects of the technical solutions according to the present invention will be given with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
In order to find the abnormality of the material conveying pipeline in the pneumatic conveying process in advance and avoid accidents, the embodiment provides an abnormality detection method for the material conveying pipeline, which mainly comprises two parts: the first part is to determine whether the material conveying pipeline to be detected has abnormal abrasion of the connecting part, and a corresponding flow chart is shown in figure 1; the second part is to determine whether the material conveying pipeline to be detected has pipeline working state abnormity, and a corresponding flow chart is shown in fig. 2. The abnormality detection method for the material conveying line will be described in detail below.
Step S1: acquiring the wear information of each position of each connecting part of a material conveying pipeline to be detected in a current first set time period and the current front position of each connecting partNAnd each position wear information in the first set time period.
In the pneumatic conveying process, this embodiment specifically means that in the pneumatic conveying process of automobile tire carbon black, the joint of adjacent pipeline sections in the material conveying pipeline, such as the joint of the horizontal section and the vertical section of the pipeline, is a part which is prone to failure. In the material conveying process, the solid air flow moves at a high speed to impact the pipeline connecting part, the connecting part is abraded to leak materials under long-term impact, and the subsequent pipeline is blocked due to pressure reduction.
According to the condition of arranging of material conveying pipeline, along with conveying pipeline's extension, the pressure in the pipeline can attenuate gradually, and the pressure reduces also along with it to the wearing and tearing of connecting portion and reduce, so in pneumatic conveying process, this embodiment selects all connecting portions of pipeline to carry out wearing and tearing information acquisition. For the connecting part, the abrasion amount of each position of the connecting part may be different due to the mode of pipeline connection and the transmission characteristic of the solid air flow, so that for each connecting part of the pipeline, the abrasion information of each position of the connecting part is collected by adopting the distributed installation infrared distance sensor, and each position of the connecting part can be determined according to the actual abrasion condition. Under the fixed condition of distance sensor position, when the wearing and tearing condition of connecting portion corresponding position department changed, then the distance value that this distance sensor gathered will change correspondingly, consequently can adopt the distance value that distance sensor gathered to come the wearing and tearing condition of representation connecting portion corresponding position department, when the distance value grow, explain the wearing and tearing degree grow of connecting portion corresponding position department.
For each connecting part of the material conveying pipeline, the infrared distance sensors installed in a distributed manner acquire and record position wear information of corresponding positions of the connecting part every 1 first set time period, and considering that the wear of each position of the connecting part is a slow process, in this embodiment, the first set time period is set to be 1 day, and the position wear information acquired by all the infrared distance sensors every 1 day forms a sequence which is recorded as
Wherein, in the step (A),
is the first of a material conveying pipelinejA sequence of wear information of each position corresponding to each connection part,
is the first of a material conveying pipelinejThe first connecting part corresponds toeThe information of the wear of the individual positions,Eis the first of a material conveying pipelinejThe total number of the position wear information corresponding to each connection portion.
In the process of pneumatic conveying, when a certain material conveying pipeline needs to be monitored, the material conveying pipeline is monitoredAccording to the mode, the abrasion information of each connecting part of the material conveying pipeline at each position in the current 1 day and the abrasion information of each connecting part in the current front are obtainedNWear information of each position of the day is obtained, so that the wear information of each position of each connecting part of the material conveying pipeline in the current first set time period and the wear information of each connecting part in the current front of the connecting part are obtainedNAnd each position wear information in the first set time period. It should be noted that, hereNIs determined according to the number of training samples participated each time when the material conveying pipeline connecting part abrasion degree prediction network is trained in the subsequent step S3.
Step S2: according to the wear information of each connecting part at each position within the current first set time period and the current previous position of each connecting partNThe wear degree evaluation value of each connecting part in the current first set time period and the wear degree evaluation value before the current time period are calculated according to the wear information of each position in the first set time periodNThe abrasion degree evaluation value in the first set time period is obtained, and then the actual evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period and the actual evaluation value before the current abrasion degree are obtainedNAnd actually evaluating the abrasion degree of the connecting part in the first set time period.
On the basis of the above-described step S1, based on the respective positional wear information of each connecting portion, the wear degree evaluation value of each connecting portion in the current first set time period and before the current is calculatedNThe abrasion degree evaluation value in the first set time period corresponds to a calculation formula as follows:
wherein the content of the first and second substances,
is as followsjThe wear degree evaluation value of each connecting part in the current first set time period or the wear degree evaluation value in each previous first set time period,
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Average value of all elements in (1) orjThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The average value of all the elements in (a),
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Median or second of all elements injThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The median value of all the elements in (a),
the function is hyperbolic tangent function for
And (6) carrying out normalization.
Evaluation value of degree of wear described above
In the formula for calculating (a) of (b),
is shown asjAverage wear of the respective jointsThe larger the average wear amount is, the larger the degree of wear is;
is shown asjThe median value of the wear amounts of the respective connecting portions is larger, and the larger the median value is, the larger the degree of wear is. When it comes tojWhen the positions of the connecting parts are not worn, the connecting parts are not worn
And
are all relatively small, the calculated wear level evaluation value
Will be smaller; when it comes tojWhen the individual positions of the respective connecting parts are worn, at the same time
Is relatively small, and
then it will increase and the calculated wear rating will be evaluated
Will increase; when it comes tojWhen a plurality of positions of the connecting parts are worn, at the moment
And
all increase, the calculated wear degree evaluation value
It will be larger. Therefore, by setting the wear degree evaluation value in this way
The calculated wear degree evaluation value can be obtained by the calculation formula
Accurately characterize thejThe degree of wear of the respective connection portions.
Determining the wear degree evaluation value of each connecting part of the material conveying pipeline to be detected in the current first set time period and before the current time period through the formulaNAfter the evaluation value of the wear degree in the first set time period, the evaluation value of the wear degree of each connecting part in the current first set time period and before the current time period are usedNCalculating the average value of the wear degree evaluation values of all the connecting parts in the current first set time period so as to obtain the actual evaluation value of the wear degree of the connecting parts of the material conveying pipeline to be detected in the current first set time period; meanwhile, calculating the average value of the wear degree evaluation values of all the connecting parts in each first preset time period before the current time, thereby obtaining the material conveying pipeline to be detected before the current timeNAnd actually evaluating the abrasion degree of the connecting part in the first set time period.
Step S3: the material conveying pipeline to be detected is arranged in front of the current positionNAnd inputting the actual evaluation value of the abrasion degree of the connecting part in the first set time period into a material conveying pipeline connecting part abrasion degree prediction network to obtain a predicted value of the evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period.
On the basis of the step S2, the material conveying pipeline to be detected is arranged before the current stepNInputting the actual evaluation value of the abrasion degree of the connecting part in the first set time period into a pre-acquired prediction network of the abrasion degree of the connecting part of the material conveying pipeline, predicting the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period under the normal condition by the prediction network, and subsequently, predicting the predicted abrasion degree of the connecting part in the current first set time period and the actual abrasion degree of the connecting part in the current first set time periodAnd comparing the abrasion degrees of the connecting parts in the time period, so that whether the abrasion abnormality of the connecting parts occurs in the material conveying pipeline to be detected can be determined. The key for predicting the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period is to obtain a material conveying pipeline connecting part abrasion degree prediction network in advance, so the obtaining process of the network is described in detail below.
Step S31: and acquiring the wear information of each position of each connecting part of each first material conveying pipeline sample in each historical first set time period, wherein each first material conveying pipeline sample is the same as the material conveying pipeline to be detected in configuration.
Selecting all first material conveying pipeline samples with the same configuration as the material conveying pipeline to be detected in the step S1, wherein the first material conveying pipeline samples have the same use time and synchronous working time, and then acquiring all position wear information of all connecting parts of all first material conveying pipeline samples in historical all first set time periods according to the mode of acquiring all position wear information of all connecting parts of the material conveying pipeline to be detected in the step S1. It should be emphasized that, the historical first set time periods are continuous, and since the first set time period is set to be 1 day in this embodiment, as a specific embodiment, in the historical working process of the first material conveying pipeline samples, the wear information of each position of each connecting part of the first material conveying pipeline samples for 90 days or 120 days can be obtained at the same time, where 90 days or 120 days are the above-mentioned historical first set time periods.
Step S32: and determining the wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period according to the wear information of each position of each connecting part of each first material conveying pipeline sample in each historical first set time period, and further obtaining the actual wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period.
In addition to the first sub-step, with reference to step S2, based on the wear information of each connecting portion of each first material conveying pipeline sample at each position within each historical first set time period, an actual evaluation value of the wear degree of the connecting portion of each first material conveying pipeline sample within each historical first set time period can be determined. Then, for each first material conveying pipeline sample in each first material conveying pipeline sample, forming a sequence of corresponding actual evaluation values of the abrasion degrees of the connecting parts in each historical first set time period according to the time sequence, and recording the sequence as the sequence
Wherein
The connection part abrasion degree actual evaluation values of the xth first material conveying pipeline sample in each first set time period in history are formed into a sequence,
and actually evaluating the connection part abrasion degree of the xth first material conveying pipeline sample in the history pth first set time period, wherein P is the total number of the history first set time periods.
Step S33: and screening each first material conveying pipeline sample according to the actual evaluation value of the abrasion degree of the connecting part of each first material conveying pipeline sample in each historical first set time period, so as to obtain each screened first material conveying pipeline sample.
Considering that the actual evaluation values of the wear degrees of the connecting parts of the first material conveying pipeline samples in the historical first set time periods are relatively close under the normal wear condition, but due to some special reasons, such as the original quality problem of the connecting parts, the difference between the actual evaluation values of the wear degrees of the connecting parts of some first material conveying pipeline samples and the actual evaluation values of the wear degrees of the connecting parts of other first material conveying pipeline samples is relatively large, if the difference between the actual evaluation values of the wear degrees of the connecting parts of some first material conveying pipeline samples and the actual evaluation values of the wear degrees of the connecting parts of other first material conveying pipeline samples is relatively large, it indicates that the wear abnormality of the first material conveying pipeline samples occurs in the historical first set time periods, so as to ensure that a relatively accurate wear degree prediction network of the connecting parts of the material conveying pipelines can be obtained subsequently, the actual evaluation value of the wear degree of the connecting portion of the first material conveying pipeline samples should not participate in the training process of the network, and at this time, the abnormal first material conveying pipeline samples with abnormal wear need to be removed, and the specific removal process is as follows:
calculating the distance between any two first material conveying pipeline samples according to the actual evaluation value of the abrasion degree of the connecting part of each first material conveying pipeline sample in each historical first set time period, wherein the corresponding calculation formula is as follows:
wherein the content of the first and second substances,
is the distance between any A-th first material conveying pipeline sample and the B-th first material conveying pipeline sample,
the connection part abrasion degree actual evaluation value of any A-th first material conveying pipeline sample in each first set time period in history is formed into a sequence,
the connection part abrasion degree actual evaluation value of any B-th first material conveying pipeline sample in each first set time period in history is formed into a sequence,
is a sequence of
And
dynamic time warping distance in between.
According to the distance between any two first material conveying pipeline samples, calculating a Local Outlier Factor of each first material conveying pipeline sample in each first material conveying pipeline sample by using a Local Outlier Factor (LOF) detection method algorithm, and determining the Outlier first material conveying pipeline samples in each first material conveying pipeline sample based on the Local Outlier Factor, wherein the specific implementation process is as follows:
(1) and calculating the k-th reachable distance of each first material conveying pipeline sample in the k-th distance neighborhood of each first material conveying pipeline sample.
Taking the first material conveying pipeline sample M as a center, and the kth distance neighborhood of the first material conveying pipeline sample M
The kth reachable distance from any one of the first material conveying pipeline samples Q to the first material conveying pipeline sample M is defined as:
wherein the content of the first and second substances,
k distance neighborhood for first material conveying pipeline sample M
The kth reachable distance from any one of the first material conveying line samples Q to the first material conveying line sample M,
represents the kth distance of the first material transfer line sample Q,
) Represents the distance between the first material transfer line sample Q and the first material transfer line sample M,
the k-th reachable distance from the first material conveying pipeline sample Q to the first material conveying pipeline sample M is defined as the greater of the k-th distance from the first material conveying pipeline sample Q and the distance from the first material conveying pipeline sample Q to the first material conveying pipeline sample M.
(2) Calculating the kth local reachable density of each first material conveying pipeline sample according to the kth reachable distance of each first material conveying pipeline sample in the kth distance neighborhood of each first material conveying pipeline sample:
wherein the content of the first and second substances,
k distance neighborhood for first material conveying pipeline sample M
The kth reachable distance from any one of the first material conveying line samples Q to the first material conveying line sample M,
is the kth distance neighborhood of the first material conveying pipeline sample M,
is the kth local reachable density of the first material conveying pipeline sample M, i.e. all the first material conveying pipeline samples in the kth distance neighborhood of the first material conveying pipeline sample M are conveyed to the first materialThe reciprocal of the average kth reachable distance of the sample M of the conveying pipeline is that when the concentration of the sample M of the first material conveying pipeline and the samples of the surrounding first material conveying pipelines is higher, the sample M of the first material conveying pipeline is at the moment
The larger the value, the lower the concentration of the first material conveying pipeline sample M and the surrounding first material conveying pipeline samples, at this time
The smaller the value.
(3) Calculating a kth local outlier factor of each first material conveying pipeline sample based on the kth local reachable density of each first material conveying pipeline sample:
wherein the content of the first and second substances,
is the kth local achievable density of the first material conveying line sample M,
is the k-th distance neighborhood of the first material conveying pipeline sample M
The kth local achievable density of any one of the first material conveying pipeline samples Q,
is the kth local outlier factor of the first material conveying pipeline sample M, which is the kth distance neighborhood of the first material conveying pipeline sample M
The average of the ratio of the local achievable density of the other first material transfer line samples to the local achievable density of the first material transfer line sample M.
Through the steps, the kth local outlier factor of each first material conveying pipeline sample can be determined, the kth local outlier factor of each first material conveying pipeline sample is compared with a set outlier factor threshold, if the kth local reachable density of the first material conveying pipeline sample M is smaller than or equal to the set outlier factor threshold, the density of the first material conveying pipeline sample M is different from that of the first material conveying pipeline sample in the neighborhood, and the first material conveying pipeline sample M possibly belongs to the same cluster as the neighborhood; if the kth local reachable density of the first material conveying pipeline sample M is greater than the set outlier threshold, the first material conveying pipeline sample M may be an abnormal pipeline. Through the method, abnormal first material conveying pipeline samples in the first material conveying pipeline samples can be determined, the abnormal first material conveying pipeline samples are removed, and the remaining first material conveying pipeline samples after removal are the screened first material conveying pipeline samples. By adopting the LOF algorithm to remove the samples of the first material conveying pipelines, the actual evaluation values of the abrasion degrees of the connecting parts of the samples of the first material conveying pipelines which are left after screening are relatively close, the abrasion abnormality of the connecting parts is not generated, and the effectiveness of the actual evaluation values of the abrasion degrees of the connecting parts subsequently participating in the material conveying pipeline connecting part abrasion degree prediction network training is ensured.
Step S34: constructing a material conveying pipeline connecting part abrasion degree prediction network, and determining training samples of the network and a weight value corresponding to each training sample according to the actual evaluation value of the abrasion degree of the connecting part of each screened first material conveying pipeline sample in each historical first set time period.
The method comprises the steps of utilizing a mixed long and short memory network I-LSTM which is a combination of a PLSTM and an LSTM to construct a material conveying pipeline connecting part abrasion degree prediction network which is a time sequence prediction network in essence, wherein the concrete implementation process of constructing the network belongs to the prior art, and is not repeated herein.
After the material conveying pipeline connecting part wear degree prediction network is constructed, for each first material conveying pipeline sample screened out, a sequence formed by arranging connecting part wear degree actual evaluation values of the first material conveying pipeline sample in each historical first set time period in a time sequence is used as a training sample set. Considering that the variation trends of training samples under ideal conditions are similar for different training sample sets, if a certain training sample in the training sample set is data with a sudden change, the certain training sample will subsequently affect the training result of the prediction network of the wear degree of the material conveying pipeline connection. Therefore, in order to ensure the accuracy of the network training result of the subsequent prediction of the abrasion degree of the connecting part of the material conveying pipeline, the embodiment analyzes each training sample in each training sample set, so as to determine the weight value corresponding to each training sample in each training sample set, and the specific implementation process is as follows:
step S341: and determining each relevant training sample of each training sample, wherein the historical first set time period corresponding to the relevant training sample is the historical first set time period adjacent to the historical first set time period corresponding to the corresponding training sample.
And for a training sample set formed by the actual evaluation values of the abrasion degrees of the connecting parts of each first material conveying pipeline sample in the historical first set time periods, determining 2 training samples before and after each training sample in the training sample set, wherein the historical first set time periods corresponding to the four training samples are adjacent to the historical first set time periods corresponding to the corresponding training samples in terms of time, and taking the 4 training samples as the related training samples of the corresponding training samples, so as to obtain each related training sample of each training sample in the training sample set. Of course, when determining the related training samples of each training sample in the training sample set, if there are less than 2 training samples in front of or behind a certain training sample, then other related training samples are determined by interpolation according to each related training sample that can be determined. For example, a training sample is preceded by only one correlated training sample and followed by two training samples, and then a correlated training sample before the training sample is determined by interpolation.
Step S342: and respectively calculating the absolute value of the difference between the actual evaluation value of the wear degree of the connecting part corresponding to each training sample and the actual evaluation value of the wear degree of the connecting part corresponding to each related training sample corresponding to the training sample.
And for a training sample set formed by the actual evaluation values of the wear degrees of the connecting parts of each first material conveying pipeline sample in each historical first set time period, each training sample in the training sample set corresponds to 4 related training samples, and the absolute value of the difference between the actual evaluation value of the wear degree of the connecting part corresponding to the training sample and the actual evaluation values of the wear degree of the connecting part corresponding to the 4 related training samples is calculated respectively.
Step S343: and calculating the average value of the absolute values of all the difference values corresponding to each training sample so as to determine the weight value corresponding to each training sample.
For a training sample set formed by actual evaluation values of the abrasion degrees of the connecting parts of each first material conveying pipeline sample in each historical first set time period, calculating an average value of absolute values of 4 difference values corresponding to each training sample in the training sample set, and further calculating a weight value corresponding to each training sample in the training sample set, wherein a corresponding calculation formula is as follows:
wherein, the first and the second end of the pipe are connected with each other,
a training sample set consisting of actual evaluation values of the abrasion degree of the connecting part of each first material conveying pipeline sample in each first set time period in historytThe weight value corresponding to each training sample,
is as followstThe average of the absolute values of all differences for each training sample.
The reason for determining the weight value corresponding to each training sample is that the reliability of the data of the training samples participating in the network training for predicting the abrasion degree of the connecting part of the material conveying pipeline subsequently cannot be absolutely guaranteed, the abnormal data can affect the training trend and the iteration process in the network training process, and the training data can be lost on the timestamp if the abnormal data are directly removed. In order to reduce the influence of abnormal data on the training process as much as possible, the weight values corresponding to the training samples are subsequently involved in the calculation of the network loss function result. When the difference between a certain training sample and each corresponding related training sample is larger, it means that the training sample may be abnormal data, and the weight value given to the training sample is smaller, so that the subsequent influence of the data on the loss function result of the network is smaller, and the accuracy of network training can be ensured.
Step S35: and constructing a loss function of the network according to the training samples of the network and the weight values corresponding to each training sample, and training the network according to the training samples and the loss function of the network, so as to obtain a trained prediction network of the abrasion degree of the connecting part of the material conveying pipeline.
Based on the training samples of the material conveying pipeline connecting part abrasion degree prediction network and the weight value corresponding to each training sample, determining the loss function corresponding to the network, wherein the corresponding calculation formula is as follows:
wherein the content of the first and second substances,
as a function of the losses of the network,
is as followstThe predicted value of the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample,
is as followstThe label value of the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample,
is as followstThe weight value corresponding to each training sample,Nthe number of training samples that each time the network is trained.
And after determining the loss function of the prediction network of the abrasion degree of the connecting part of the material conveying pipeline, training the network by using the training sample of the network. When the material conveying pipeline connecting part wear degree prediction network is trained, 2/3 in a training sample set formed by actual evaluation values of the connecting part wear degrees of all the first material conveying pipelines screened in each first set time period of the sample history is used as a training set, and the rest 1/3 is used as a test set. Taking the training samples in each training set as a batch of training samples, and respectively inputting the training samples of each batch into the network according to each succession in the training samples of each batchNA training sample for predicting the succession of the training samplesNThe predicted value corresponding to each training sample is based on the continuityNAnd determining a loss function of the network according to the predicted value, the label value and the weight value corresponding to each training sample, and correcting the parameters of the network according to the loss function value to finally realize the training of the network. As the specific training process of the material conveying pipeline connecting part abrasion degree prediction network belongs to the prior art, the detailed description is omitted here.
After the trained wear degree prediction network of the material conveying pipeline connecting parts is obtained through the steps S31-S35, the material conveying pipeline to be detected is arranged in front of the current positionNThe actual evaluation value of the degree of wear of the joint part in the first set time period is input to the evaluation valueIn the network, the predicted value of the connection part abrasion degree evaluation value of the material conveying pipeline to be detected in the current first set time period is output by the network.
Step S4: and comparing the predicted value of the connection part abrasion degree evaluation value of the material conveying pipeline to be detected in the current first set time period with the actual evaluation value of the connection part abrasion degree, and determining whether the connection part abrasion abnormality occurs in the material conveying pipeline to be detected.
After the predicted value of the connection wear degree evaluation value of the material conveying pipeline to be detected in the current first set time period is obtained in step S3, the predicted value of the connection wear degree evaluation value is compared with the actual evaluation value of the connection wear degree of the material conveying pipeline to be detected in the current first set time period, and if the difference between the actual evaluation value of the connection wear degree and the predicted value of the connection wear degree evaluation value is greater than the set threshold value of the connection wear degree, it means that the material conveying pipeline to be detected is worn abnormally, and at this time, an early warning is started.
Through the steps S1-S4, the detection of the abnormal wear of the connecting portion of the detected material conveying pipeline can be realized, the abnormal wear of the connecting portion is detected, the detection result is obtained every first set time period, and the abnormal wear of the connecting portion is analyzed from the long time of the pipeline failure because the set time of the first set time period is relatively long. In addition, considering that the pipeline working state abnormality of pipe blockage can occur in the material conveying pipeline due to reasons of unstable air pressure, unsatisfied material humidity and the like, the pipeline working state abnormality detection of the material conveying pipeline to be detected can be realized by analyzing the air pressure value and the humidity value of the material conveying pipeline in the following steps S5-S10, and the pipeline working state abnormality detection is analyzed from the short-time property of pipeline failure because the pipeline working state abnormality detection obtains one detection result every second set time period and the set time of the second set time period is relatively short. How to realize the pipeline work state abnormity detection of the material conveying pipeline to be detected is described in detail below.
Step S5: the method comprises the steps of obtaining each position air pressure value and each position humidity value of a material conveying pipeline to be detected at each set time in a current second set time period and each position air pressure value and each position humidity value of each second material conveying pipeline sample at each set time in a historical second set time period, wherein the configuration of each second material conveying pipeline sample is the same as that of the material conveying pipeline to be detected, and the pipeline work state abnormity of each second material conveying pipeline sample does not occur in the historical second set time period.
In the process of pneumatically conveying materials, compressed air is mixed with the materials to form the flow of the materials and the air, and after a discharge valve is opened, the materials and air flow reach a storage bin through a material conveying pipeline. The conveying effect of the materials is closely related to the strength of the compressed air, and the strength of the strength can be reflected by the pressure in the material conveying pipeline. When the material conveying pipeline is blocked, leaks and the like, the air pressure in the material conveying pipeline is correspondingly changed. Therefore, the present embodiment uses the distributed air pressure sensor to collect the air pressure values at different positions of the material conveying pipeline, where the different positions can be determined according to the actual situation of the material conveying pipeline. When the air pressure value is collected, the distributed air pressure sensor records the air pressure every 1 second, so that the air pressure values at different positions every 1 second in the current second set time period can be obtained, wherein the second set time period refers to 1 minute, and every 1 second in the 1 minute is taken as a set time.
When materials are conveyed pneumatically, besides the requirement on the pressure of compressed air, the purification degree of the compressed air also has great influence on pneumatic transmission. If the compressed air is not thoroughly purified, the air contains moisture, and materials and the air with the moisture are mutually adhered after being contacted, so that the flow resistance is suddenly increased, and the pipe blockage is caused. Therefore, this embodiment uses distributed humidity sensor to gather the humidity value of each different position department of material conveying pipeline, when carrying out humidity value collection, distributed humidity sensor once records humidity every 1 second equally to can acquire the humidity value of each different position department every 1 second in the current second settlement time quantum.
Similarly, for each second material conveying pipeline sample with the same configuration as the material conveying pipeline to be detected, in the historical working process of the second material conveying pipeline samples, under the condition that the second material conveying pipeline samples are determined not to have pipeline work state abnormity, each position air pressure value and each position humidity value of each second material conveying pipeline sample at each set moment in the historical second set time period are obtained.
Step S6: and determining the pipeline abnormality index at each set time in the current second set time period and the historical second set time period according to each position air pressure value and each position humidity value at each set time in the current second set time period and the historical second set time period, so as to obtain the corresponding pipeline abnormality index vector in the current second set time period and the historical second set time period.
Since the way of determining the pipeline abnormality index at each set time in the current second set time period is exactly the same as the way of determining the pipeline abnormality index at each set time in each historical second set time period, taking the determination of the pipeline abnormality index at each set time in the current second set time period as an example, the corresponding calculation formula is:
wherein the content of the first and second substances,
is at the first within the current second set period of timeiThe abnormal index of the pipeline at each set moment,
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The standard deviation of all the elements in (A),
a sequence formed by the air pressure values of the positions at each set time in the current second set time period
The average value of all the elements in (A),
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The maximum position humidity value among all the elements in (c),
the function is hyperbolic tangent function for
And (6) carrying out normalization.
The above-mentioned abnormality index of the piping at each set time within the current second set period of time
In the calculation formula, considering that for a material conveying pipeline, a certain pressure loss is generated along the pipeline along with the extension of the pipeline, the distributed air pressure sensors are adopted to acquire air pressure values of different positions of the pipeline at the same moment, the standard deviation of the air pressure values is calculated, the fluctuation of the whole air pressure of the conveying pipeline is reflected through the standard deviation, if the air pressure in the pipeline does not fluctuate violently, the standard deviation is kept unchanged, and when the air pressure in the pipeline fluctuates violently, the standard deviation is increased relatively, which indicates that the material conveying pipeline is more likely to generate abnormity. In addition, the compressed air produced by the air compressor is normally drierThe humidity sensor is arranged to collect the air flow humidity of the pipeline at different positions at the same moment, the humidity mean value of the humidity values is used as the overall humidity evaluation, the maximum humidity value of the humidity values is considered, and when the humidity mean value is larger or the maximum humidity value is larger, the material conveying pipeline is more likely to be abnormal.
The above-mentioned pipeline abnormality index at each set time within the current second set time period
Has a value range of [0,1 ]]When the value of U is closer to 1, it means that the probability of the material conveying pipeline being abnormal is higher, otherwise, it means that the probability of the material conveying pipeline being abnormal is lower. The abnormal index of the pipeline at each set moment in the current second set time period
The abnormal index vectors are arranged according to the time sequence, so that the abnormal index vector of a pipeline corresponding to the material conveying pipeline to be detected in the current second set time period can be obtained
Subsequently based on the pipeline anomaly index vector
And whether the work state of the material conveying pipeline to be detected is abnormal or not is judged, so that the condition that the judgment is wrong possibly caused by accidental reasons when the material conveying pipeline to be detected is judged to be abnormal or not only according to the pipeline abnormality index at a set moment can be avoided.
In the same manner, according to the respective position air pressure values and the respective position humidity values of the respective second material conveying pipeline samples at each set time within the historical second set time period, the pipeline abnormality index of the respective second material conveying pipeline samples at each set time within the historical second set time period can be determined, so that the pipeline abnormality index vector corresponding to the respective second material conveying pipeline samples within the historical second set time period can be obtained.
Step S7: and determining the radius of the hyperplane of the abnormal pipeline index vector and the distance from the corresponding abnormal pipeline index vector to the radius of the hyperplane of the abnormal pipeline index vector in the current second set time period according to the corresponding abnormal pipeline index vector in the current second set time period and in the historical second set time period, and further prejudging whether the abnormal pipeline work state occurs in the material conveying pipeline to be detected in the current second set time period.
In order to pre-determine whether a pipeline condition abnormality occurs in a material conveying pipeline to be detected within a current second set time period, the embodiment provides a fault detection method based on a heterogeneous geodesic distance SVDD, and in order to implement the fault detection method, the embodiment further provides an abnormality detection system, which includes the following modules: the device comprises a parameter setting module, a heterogeneous geodesic distance calculation module, a model history module and a fault detection module.
Taking the pipeline abnormal index vector corresponding to each second material conveying pipeline sample obtained in the step S6 in a second set time period as a history sample, thereby forming a history sample set
And transmitting the historical sample set to a parameter setting module.
The parameter setting module is used for setting the parameter according to the received historical sample set
Setting the number K of the neighbors to 10, setting the zero proportion alpha to 95 percent and setting the weight expansion coefficient
And setting the parameters to be 0.1, and transmitting the set parameters to a heterogeneous geodesic distance calculation module.
The heterogeneous geodesic distance calculation module calculates the distance between the two heterogeneous geodesic lines according to the historical sample set
And calculating the distance between the heterogeneous geodesic lines of the historical samples according to the parameters
And measure the distance between the heterogeneous ground wires
And transmitting to the model history module.
Heterogeneous geodesic distance according to historical samples
The model history module obtains the optimal solution of the optimized variable by adopting a weighting support vector data-based description method
And the abnormal index vector of the pipeline exceeds the radius R of the plane, and the optimal solution of the optimized variable is obtained
And transmitting the abnormal index vector of the pipeline beyond the plane radius R to a fault detection module.
Optimal solution of fault detection module according to optimization variable
Calculating the corresponding pipeline abnormal index vector in the current second set time period
Distance to radius R of abnormal exponential vector of pipeline
。
Based on this distance
The fault detection module judges whether the working state of the material conveying pipeline to be detected is abnormal in real time, if so, the fault detection module judges whether the working state of the material conveying pipeline to be detected is abnormal
If not, the material conveying pipeline to be detected is considered to have no pipeline working state abnormity, otherwise, the material conveying pipeline to be detected is considered to have pipeline working state abnormity.
Step S8: if the pipeline is judged to be abnormal in working state in advance, the material conveying pipeline to be detected is continuously acquired in the futureGDetermining the air pressure value and the humidity value of each position at each set time in the second set time period in the futureGThe abnormal index of the pipeline at each set moment in the second set time period is obtained in the futureGAnd the corresponding pipeline abnormal index vector in the second set time period.
Since the determination of whether the pipeline work state abnormality occurs in the material conveying pipeline to be detected in step S7 is performed based on 1 second set time period, there may be a case where an accidental situation causes a misdetermination and a false alarm, which may affect the conveying efficiency of the pipeline. Therefore, in order to ensure the accuracy of the early warning of the working condition of the material conveying pipeline, in the embodiment, when the pipeline is judged to have abnormal working condition through the step S7, that is, when the pipeline is judged to have abnormal working condition for the first time, the material conveying pipeline to be detected will be detected in the futureGThe respective position air pressure values and the respective position humidity values at each set time within the second set time period are collected, and G =3 is set in the present embodiment, thereby determining that the future time is reachedGAnd the corresponding pipeline abnormal index vector in the second set time period. Due to determination in the futureGThe process of the corresponding pipeline abnormal index vector in the second set time period is completely the same as the process of determining the corresponding pipeline abnormal index vector in the current second set time period in step S6, and is not repeated herein.
Step S9: according to the futureGDetermining the corresponding pipeline abnormal index vector in the second set time period and the historical second set time period in the futureGCorresponding pipeline abnormal index vector to pipeline abnormal index vector in second set time periodAnd the distance exceeds the radius of the plane, so that the comprehensive pipeline abnormity index corresponding to the material conveying pipeline to be detected is determined.
Since the pipe abnormality index vector hyperplane radius R has been determined by the above-described step S7, it is determined that it is in the futureGThe specific implementation manner of the distance from the corresponding pipeline abnormality index vector to the pipeline abnormality index vector hyperplane radius R in the second set time period is completely the same as the specific implementation manner of determining the distance from the corresponding pipeline abnormality index vector to the pipeline abnormality index vector hyperplane radius R in the current second set time period in step S7, and is not described here again.
Hyperplane radius R and future based on pipeline anomaly index vectorGAnd calculating the distance from the corresponding pipeline abnormal index vector to the pipeline abnormal index vector hyperplane radius R in a second set time period, and calculating the comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected, wherein the corresponding calculation formula is as follows:
wherein, the first and the second end of the pipe are connected with each other,
is a comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected,
is the radius of the hyperplane of the abnormal index vector of the pipeline,
to be in the futuregThe corresponding pipeline abnormal index vector in the second set time period,Gis the total number of the second set period in the future.
Step S10: and actually judging whether the pipeline work state abnormity occurs in the material conveying pipeline to be detected according to the comprehensive pipeline abnormity index corresponding to the material conveying pipeline to be detected.
According to the comprehensive pipeline difference in the step S9Constant index
The calculation formula shows that when the value of the radius R of the hyperplane of the abnormal index vector of the pipeline is less than
When the temperature of the water is higher than the set temperature,
is between (0, 1); on the contrary, the first step is to take the reverse,
is greater than or equal to 1. The embodiment is implemented by comparing the comprehensive pipeline abnormity index
Make a judgment if
When the value of (2) is (0,1), the material conveying pipeline to be detected is considered to have pipeline working state abnormity, and early warning is started at the moment, otherwise, the material conveying pipeline to be detected is considered to have no pipeline working state abnormity, and the material conveying pipeline to be detected continues to work.
The embodiment also provides an abnormality detection device for the material conveying pipeline, which comprises a processing unit and a storage unit, wherein the storage unit is used for storing instructions, and the processing unit is used for executing the instructions stored in the storage unit, so as to realize the above abnormality detection method for the material conveying pipeline. Since the abnormality detection method for the material conveying pipeline has been described in detail in the above, it is not described herein again.
The abnormity detection method and the abnormity detection device for the material conveying pipeline are carried out in the working period of the material conveying pipeline, and are suitable for various places where the material conveying pipeline works, particularly for places where the number of pipelines is large. The invention can accurately find whether a certain material conveying pipeline is abnormal in time, thereby avoiding the pipeline from being in failure and saving the manpower for troubleshooting. The invention analyzes from two aspects of long time and short time when the pipeline has faults, namely from two aspects of abnormal abrasion of the connecting part of the pipeline and abnormal working state of the pipeline, thereby not only ensuring the running safety of the pipeline, but also not influencing the work of the pipeline. When the pipeline has an alarm signal, the current pipeline is stopped to be maintained manually, so that serious faults are avoided.
It should be noted that: the above-mentioned embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.
Claims (9)
1. An abnormality detection method for a material conveying pipeline is characterized by comprising the following steps:
acquiring the wear information of each position of each connecting part of a material conveying pipeline to be detected in a current first set time period and the current front position of each connecting partNWear information of each position within a first set time period;
according to the wear information of each connecting part at each position in the current first set time period and the current previous position of each connecting partNThe wear degree evaluation value of each connecting part in the current first set time period and the wear degree evaluation value before the current time period are calculated according to the wear information of each position in the first set time periodNEvaluating the wear degree of the material conveying pipeline to be detected in the first set time period so as to obtain the actual evaluation value of the wear degree of the connecting part of the material conveying pipeline to be detected in the current first set time period and the current wear degree of the connecting part of the material conveying pipeline to be detected in the previous set time periodNActually evaluating the abrasion degree of the connecting part in a first set time period;
the material conveying pipeline to be detected is at presentFront ofNInputting the actual evaluation value of the abrasion degree of the connecting part in the first set time period into a material conveying pipeline connecting part abrasion degree prediction network to obtain a predicted value of the evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period;
comparing the predicted value of the evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period with the actual evaluation value of the abrasion degree of the connecting part, and determining whether the material conveying pipeline to be detected is abnormal in abrasion of the connecting part;
the method further comprises the following steps:
acquiring each position air pressure value and each position humidity value of a material conveying pipeline to be detected at each set moment in a current second set time period and each position air pressure value and each position humidity value of each second material conveying pipeline sample at each set moment in a historical second set time period, wherein the configuration of each second material conveying pipeline sample is the same as that of the material conveying pipeline to be detected, and the pipeline work state abnormality of each second material conveying pipeline sample does not occur in the historical second set time period;
determining the pipeline abnormal index at each set time in the current second set time period and the historical second set time period according to each position air pressure value and each position humidity value at each set time in the current second set time period and the historical second set time period, so as to obtain the corresponding pipeline abnormal index vector in the current second set time period and the historical second set time period;
determining the radius of the hyperplane of the abnormal index vector of the pipeline and the distance from the abnormal index vector of the pipeline corresponding to the current second set time period to the radius of the hyperplane of the abnormal index vector of the pipeline according to the abnormal index vector of the pipeline corresponding to the current second set time period and the historical second set time period, and further prejudging whether the pipeline work state abnormality occurs to the material conveying pipeline to be detected in the current second set time period;
if the pipeline is judged to have abnormal pipeline state, the abnormal pipeline state is continuously obtainedTaking a material conveying pipeline to be detected in the futureGDetermining the air pressure value and the humidity value of each position at each set time in the second set time period in the futureGThe abnormal index of the pipeline at each set moment in the second set time period is obtained in the futureGCorresponding pipeline abnormal index vectors in a second set time period;
according to the futureGDetermining the corresponding pipeline abnormal index vector in the second set time period and the historical second set time period in the futureGDetermining the distance from the corresponding pipeline abnormal index vector to the pipeline abnormal index vector hyperplane radius in a second set time period, thereby determining the comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected;
and actually judging whether the pipeline work state abnormity occurs in the material conveying pipeline to be detected according to the comprehensive pipeline abnormity index corresponding to the material conveying pipeline to be detected.
2. The abnormality detection method for the material transport pipeline according to claim 1, characterized in that the wear degree evaluation value of each connection portion in the current first set period of time and before the current is calculatedNThe calculation formula corresponding to the wear degree evaluation value in the first set time period is as follows:
wherein the content of the first and second substances,
is as followsjThe wear degree evaluation value of each connecting part in the current first set time period or the wear degree evaluation value in each previous first set time period,
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Average value of all elements in (1) orjThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The average value of all the elements in (a),
is as followsjSequence formed by abrasion information of each position of each connecting part in current first set time period
Median or second of all elements injThe connection part is in a sequence formed by the abrasion information of each position in each current previous first set time period
The median value of all the elements in (a),
the function is a hyperbolic tangent function.
3. The method for detecting the abnormality of the material conveying pipeline according to claim 1, further obtaining an actual evaluation value of the degree of wear of the connecting portion of the material conveying pipeline to be detected in the current first set time period and before the current time periodNThe step of actually evaluating the degree of wear of the connecting portion for each first set period of time includes:
according to the abrasion degree evaluation value of each connecting part in the current first set time period and before the current timeNThe evaluation values of the wear degrees in the first set time period respectively calculate the average value of the evaluation values of the wear degrees of all the connecting parts in the current first set time period and the average value of the evaluation values of the wear degrees in the current first set time periodThe average value of the abrasion degree evaluation values in each current first set time period is obtained, so that the actual evaluation value of the abrasion degree of the connecting part of the material conveying pipeline to be detected in the current first set time period and the actual evaluation value of the abrasion degree of the connecting part in the current first set time period are correspondingly obtainedNAnd actually evaluating the abrasion degree of the connecting part in the first set time period.
4. The anomaly detection method for a material conveying pipeline according to claim 1, characterized in that the step of obtaining the network of prediction of the degree of wear of the material conveying pipeline connection portion comprises:
acquiring abrasion information of each position of each connecting part of each first material conveying pipeline sample in each historical first set time period, wherein each first material conveying pipeline sample is the same as the material conveying pipeline to be detected in configuration;
determining a wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period according to wear information of each connecting part of each first material conveying pipeline sample in each historical position in each historical first set time period, and further obtaining an actual wear degree evaluation value of each connecting part of each first material conveying pipeline sample in each historical first set time period;
screening each first material conveying pipeline sample according to the actual evaluation value of the abrasion degree of the connecting part of each first material conveying pipeline sample in each historical first set time period, so as to obtain each screened first material conveying pipeline sample;
constructing a material conveying pipeline connecting part abrasion degree prediction network, and determining training samples of the network and a weight value corresponding to each training sample according to an actual evaluation value of the abrasion degree of the connecting part of each screened first material conveying pipeline sample in each historical first set time period;
and constructing a loss function of the network according to the training samples of the network and the weight values corresponding to each training sample, and training the network according to the training samples and the loss function of the network, so as to obtain a trained prediction network of the abrasion degree of the connecting part of the material conveying pipeline.
5. The anomaly detection method for the material conveying pipeline according to claim 4, wherein the step of determining the weight value corresponding to each training sample comprises:
determining each relevant training sample of each training sample, wherein the historical first set time period corresponding to the relevant training sample is the historical first set time period adjacent to the historical first set time period corresponding to the corresponding training sample;
respectively calculating the absolute value of the difference between the actual evaluation value of the wear degree of the connecting part corresponding to each training sample and the actual evaluation value of the wear degree of the connecting part corresponding to each related training sample corresponding to the training sample;
and calculating the average value of the absolute values of all the difference values corresponding to each training sample, thereby determining the weight value corresponding to each training sample.
6. The anomaly detection method for the material conveying pipeline according to claim 4, wherein a calculation formula corresponding to the loss function of the network is as follows:
wherein, the first and the second end of the pipe are connected with each other,
as a function of the losses of the network,
is as followstThe predicted value of the actual evaluation value of the abrasion degree of the connecting part corresponding to each training sample,
is as followstActual evaluation value of abrasion degree of connecting part corresponding to training sampleThe value of the tag of (a) is,
is as followstThe weight value corresponding to each training sample,Nthe number of training samples that each time the network is trained.
7. The abnormality detection method for the material conveying pipeline according to claim 1, characterized in that the calculation formula for determining the pipeline abnormality index at each set time in the current second set time period corresponds to:
wherein, the first and the second end of the pipe are connected with each other,
is the first within the current second set time periodiThe abnormal index of the pipeline at each set moment,
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The standard deviation of all the elements in (A),
for the sequence of the air pressure values of the respective positions at each set time within the current second set time period
The average value of all the elements in (A),
for each setting within the current second set periodSequence of air pressure values at various positions of time
The maximum position humidity value among all the elements in (c),
the function is a hyperbolic tangent function.
8. The anomaly detection method for the material conveying pipeline according to claim 1, wherein the calculation formula corresponding to the comprehensive pipeline anomaly index for determining the material conveying pipeline to be detected is as follows:
wherein the content of the first and second substances,
is a comprehensive pipeline abnormal index corresponding to the material conveying pipeline to be detected,
is the radius of the hyperplane of the abnormal index vector of the pipeline,
to be in the futuregAnd G is the total number of the future second set time periods.
9. An abnormality detection device for a material conveying pipeline, characterized by comprising a processing unit and a storage unit, wherein the storage unit is used for storing instructions, and the processing unit is used for executing the instructions stored in the storage unit so as to implement the abnormality detection method for the material conveying pipeline according to any one of the claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210817516.XA CN114881269B (en) | 2022-07-13 | 2022-07-13 | Abnormity detection method and device for material conveying pipeline |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210817516.XA CN114881269B (en) | 2022-07-13 | 2022-07-13 | Abnormity detection method and device for material conveying pipeline |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114881269A CN114881269A (en) | 2022-08-09 |
| CN114881269B true CN114881269B (en) | 2022-09-27 |
Family
ID=82683566
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210817516.XA Active CN114881269B (en) | 2022-07-13 | 2022-07-13 | Abnormity detection method and device for material conveying pipeline |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114881269B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115121989A (en) * | 2022-08-31 | 2022-09-30 | 南通恒强轧辊有限公司 | End socket welding quality evaluation method |
| CN117851974B (en) * | 2024-03-07 | 2024-05-14 | 艾信智能环境科技(无锡)有限公司 | Medical pneumatic type garbage bedding and clothing intelligent collection system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108089962A (en) * | 2017-11-13 | 2018-05-29 | 北京奇艺世纪科技有限公司 | A kind of method for detecting abnormality, device and electronic equipment |
| CN111882338A (en) * | 2020-08-11 | 2020-11-03 | 网易（杭州）网络有限公司 | Online people number abnormality detection method and device and electronic equipment |
| CN111929091A (en) * | 2020-08-17 | 2020-11-13 | 上海明略人工智能(集团)有限公司 | Equipment abnormality detection method and device, readable storage medium and electronic equipment |
| WO2021056724A1 (en) * | 2019-09-23 | 2021-04-01 | 平安科技（深圳）有限公司 | Anomaly detection method and apparatus, electronic device and storage medium |
| CN113190976A (en) * | 2021-04-13 | 2021-07-30 | 国网江西省电力有限公司供电服务管理中心 | Charging station loss analysis method |
| CN113588179A (en) * | 2021-06-24 | 2021-11-02 | 武汉众智鸿图科技有限公司 | Water supply pipe network leakage detection method and system |
-
2022
- 2022-07-13 CN CN202210817516.XA patent/CN114881269B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108089962A (en) * | 2017-11-13 | 2018-05-29 | 北京奇艺世纪科技有限公司 | A kind of method for detecting abnormality, device and electronic equipment |
| WO2021056724A1 (en) * | 2019-09-23 | 2021-04-01 | 平安科技（深圳）有限公司 | Anomaly detection method and apparatus, electronic device and storage medium |
| CN111882338A (en) * | 2020-08-11 | 2020-11-03 | 网易（杭州）网络有限公司 | Online people number abnormality detection method and device and electronic equipment |
| CN111929091A (en) * | 2020-08-17 | 2020-11-13 | 上海明略人工智能(集团)有限公司 | Equipment abnormality detection method and device, readable storage medium and electronic equipment |
| CN113190976A (en) * | 2021-04-13 | 2021-07-30 | 国网江西省电力有限公司供电服务管理中心 | Charging station loss analysis method |
| CN113588179A (en) * | 2021-06-24 | 2021-11-02 | 武汉众智鸿图科技有限公司 | Water supply pipe network leakage detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114881269A (en) | 2022-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114881269B (en) | Abnormity detection method and device for material conveying pipeline | |
| CN110647133B (en) | Rail transit equipment state detection maintenance method and system | |
| US7536371B2 (en) | Apparatus and method for the analysis of a process having parameter-based faults | |
| CN110703214B (en) | Weather radar state evaluation and fault monitoring method | |
| CN107844067B (en) | A kind of gate of hydropower station on-line condition monitoring control method and monitoring system | |
| EP2788724A2 (en) | System and method for identifying related events in a resource network monitoring system | |
| CN111609890A (en) | Fracturing manifold working condition monitoring, service life prediction and feedback regulation and control system | |
| CN113341813B (en) | Urban gas medium-low pressure pipe network detection method and system | |
| Jin et al. | An optimal maintenance strategy for multi-state deterioration systems based on a semi-Markov decision process coupled with simulation technique | |
| CN110570013A (en) | Single-station online wave period data prediction diagnosis method | |
| CN117308275B (en) | Temperature difference-based pipeline connection abnormality detection method and system | |
| CN113065721A (en) | Method, device, equipment and medium for graded early warning of leakage events of community water supply network | |
| Berenguer et al. | Inspection and maintenance planning: An application of semi-Markov decision processes | |
| CN207992717U (en) | A kind of gate of hydropower station on-line condition monitoring system | |
| CN117927729A (en) | Industrial valve fault remote monitoring and early warning method and system | |
| EP4273520A1 (en) | Method and system for comprehensively diagnosing defect in rotating machine | |
| CN115270982A (en) | Switch cabinet fault prediction method based on multi-data neural network | |
| CN109029697A (en) | Ring cold machine based on the unilateral detection method of spectrum signature leaks out on-line fault diagnosis method | |
| CN113269266A (en) | Knowledge distillation-based wafer manufacturing process fault detection method under multi-fault mode | |
| CN110907175A (en) | Static multi-sensor rolling bearing monitoring method and device | |
| CN115953152A (en) | Intelligent operation and maintenance management system for industrial pipeline safety | |
| CN118066934A (en) | Intelligent fault positioning system for weapon equipment | |
| CN118052082B (en) | Heating armoured wire pipe orifice sealing control method and system | |
| KR102566876B1 (en) | System and method for detecting anomaly in air drying equipment for radar system using unsupervised learning method | |
| CN117589444B (en) | Wind driven generator gear box fault diagnosis method based on federal learning |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: An anomaly detection method and device for material conveying pipelines Granted publication date: 20220927 Pledgee: Shandong Liangshan Rural Commercial Bank Co.,Ltd. Pledgor: SHANDONG ENXIN SPECIAL VEHICLE MANUFACTURING Co.,Ltd. Registration number: Y2024980005930 |