CN115689071B - Equipment fault fusion prediction method and system based on associated parameter mining - Google Patents
Equipment fault fusion prediction method and system based on associated parameter mining Download PDFInfo
- Publication number
- CN115689071B CN115689071B CN202310001115.1A CN202310001115A CN115689071B CN 115689071 B CN115689071 B CN 115689071B CN 202310001115 A CN202310001115 A CN 202310001115A CN 115689071 B CN115689071 B CN 115689071B
- Authority
- CN
- China
- Prior art keywords
- fault
- equipment
- prediction result
- current time
- prediction
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000004927 fusion Effects 0.000 title claims abstract description 19
- 238000005065 mining Methods 0.000 title claims abstract description 16
- 230000002159 abnormal effect Effects 0.000 claims abstract description 42
- 238000010606 normalization Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 48
- 238000012360 testing method Methods 0.000 claims description 33
- 238000010801 machine learning Methods 0.000 claims description 14
- 239000013598 vector Substances 0.000 claims description 13
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 claims 1
- 230000005856 abnormality Effects 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 4
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- -1 vapors Substances 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a device fault fusion prediction method and a system based on associated parameter mining, and the method specifically comprises the following steps: collecting historical operation parameters and time sequence data of equipment; performing sequence normalization on each fault data based on the time sequence data to obtain a fault event sequence set; constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the operation parameters of the current time, and taking the predicted abnormal working condition as a first prediction result; obtaining a second prediction result of the fault through another method step; and effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result. According to the method, the relevance of the operation parameters and the relevance of the time sequence are taken into consideration, more complete equipment abnormality information and more advanced and accurate prediction results can be obtained, and the rapid and efficient target of the fault prediction method can be achieved.
Description
Technical Field
The invention relates to the technical field of reliability maintenance engineering, in particular to a device fault fusion prediction method and system based on associated parameter mining.
Background
The traditional industrial equipment fault early warning method mostly trains a fault prediction model based on a data set, then inputs relevant data into the trained prediction model to output a fault prediction result, for example, CN115186904A (country: china, publication date: 20221014) discloses a industrial equipment fault prediction method and device based on a Transformer, which is implemented by acquiring a time sequence data set corresponding to the health state of target industrial equipment; inputting the time sequence data set into a trained fault prediction model, and outputting a fault prediction value of the time sequence data set, wherein the fault prediction model is obtained by training based on a training sample carrying a fault prediction value label; and when the fault prediction value is larger than the fault threshold value, judging that the target industrial equipment is faulty, otherwise, judging that the target industrial equipment is normal in operation. Although the conventional fault prediction method is convenient and simple to apply, the running state of the equipment is not fixed in the running process of the industrial equipment, and the accuracy of the conventional prediction method is not high enough along with the increase of data updating, so that the prediction accuracy requirement in a complex industrial environment can not be met.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims at providing the equipment fault fusion prediction method and the system based on the associated parameter mining aiming at the defects of low prediction efficiency, low accuracy and low instantaneity in the traditional industrial equipment fault prediction technology. The first prediction result obtained by the bipartite graph and the second prediction result obtained by the alternative method are effectively fused to obtain a final equipment failure prediction result, so that the failure prediction result can reach the aims of high speed, high efficiency and high accuracy.
The technical scheme is as follows: in order to achieve the aim of the invention, the invention adopts the following technical scheme: an equipment fault fusion prediction method based on associated parameter mining comprises the following steps:
s1, collecting historical operation parameters and time sequence data of equipment;
s2, carrying out sequence normalization on each fault data based on the time sequence data to obtain a fault event sequence set;
s3, constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
s4, predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time, and taking the predicted abnormal working condition as a first prediction result;
s5, segment linearization vector representation is carried out on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Is the kthOperating parameter characteristics of the time sequence section;
s6, assuming that the ith time sequence segment is in a fault-free state, the corresponding operation parameter characteristic is x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j ;
S7, calculating the operation parameter x of the current time a a And x i Characteristic distance D1 between:whereinIs a correction coefficient;
s8, calculating the operation parameter x of the current time a a And x j Characteristic distance D2 between:,representing the 2 norms of the vectors;
s9, integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment, and taking the abnormal working condition as a second prediction result;
s10, effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result.
Further, the step S9 of predicting the abnormal working condition of the device by integrating the value ranges of D1 and D2, and taking the abnormal working condition as the second prediction result specifically includes:
When D1 is smaller thanAnd D2 is greater thanWhen the equipment is in fault, the equipment at the current time is indicated to be in fault;
when D1 is greater thanAnd D2 is less thanWhen the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater thanWhen the equipment at the current time is in fault, the possibility of the equipment at the current time is indicated;
when D1 is smaller thanAnd D2 is less thanWhen the device is in fault, the possibility of the device at the current time is indicated.
Further, the step S10 of effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result specifically comprises the step of fusing the first prediction result and the second prediction result through a decision-making level model obtained by a machine learning method to obtain the final equipment failure prediction result.
Further, the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises the following steps:
taking fault data related to equipment history as training data to perform data preprocessing to form a training set and a testing set;
training the decision-level model by utilizing the training set;
testing the decision-level model by using the test set;
and updating model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
Based on the same inventive concept, the equipment fault fusion prediction system based on the associated parameter mining disclosed by the invention comprises the following components:
the acquisition module is used for acquiring historical operation parameters and time sequence data of the equipment;
the system comprises a normalization module, a fault event sequence collection module and a fault event sequence collection module, wherein the normalization module is used for performing sequence normalization on each fault data based on time sequence data to obtain a fault event sequence set;
the construction module is used for constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
the prediction module 1 is used for predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time and taking the predicted abnormal working condition as a first prediction result;
the segmentation module is used for carrying out piecewise linearization vector representation on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Operating parameter characteristics of the kth time sequence segment;
assuming that the ith sequential sequence segment is in a fault-free stateThe corresponding operation parameter is characterized as x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j ;
A calculation module 1 for calculating an operation parameter x of the current time a a And x i Characteristic distance D1 between:whereinIs a correction coefficient;
a calculation module 2 for calculating an operation parameter x of the current time a a And x j Characteristic distance D2 between:,representing the 2 norms of the vectors;
the prediction module 2 is used for integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment and taking the abnormal working condition as a second prediction result;
and the fusion prediction module is used for effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result.
Further, the value ranges of D1 and D2 are synthesized, and the abnormal working condition of the device is predicted, and as a second prediction result, the method specifically includes:
When D1 is smaller thanAnd D2 is greater thanWhen the equipment is in fault, the equipment at the current time is indicated to be in fault;
when D1 is greater thanAnd D2 is less thanWhen the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater thanWhen the equipment at the current time is in fault, the possibility of the equipment at the current time is indicated;
when D1 is smaller thanAnd D2 is less thanWhen the device is in fault, the possibility of the device at the current time is indicated.
Further, the first prediction result and the second prediction result are effectively fused to obtain a final equipment failure prediction result, and the method specifically comprises the step of fusing the first prediction result and the second prediction result through a decision-making level model obtained through a machine learning method to obtain the final equipment failure prediction result.
Further, the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises the following steps:
the preprocessing module is used for preprocessing data by taking fault data related to equipment history as training data to form a training set and a testing set;
the training module is used for training the decision-level model by utilizing the training set;
the test module is used for testing the decision-level model by utilizing the test set;
and the iterative training module is used for updating the model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
The beneficial effects are that:
1. the equipment fault fusion prediction method based on the associated parameter mining can be used for equipment fault prediction under complex industrial conditions. S1, collecting historical operation parameters and time sequence data of equipment; s2, carrying out sequence normalization on each fault data based on the time sequence data to obtain a fault event sequence set; s3, constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; s4, predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time, and taking the predicted abnormal working condition as a first prediction result; s5, carrying out piecewise linearization vector representation on the time sequence data according to a fixed time distance; s6, assuming that the ith time sequence segment is in a fault-free state, the corresponding operation parameter characteristic is x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j The method comprises the steps of carrying out a first treatment on the surface of the S7, calculating the operation parameter x of the current time a a And x i A characteristic distance D1 therebetween; s8, calculating the operation parameter x of the current time a a And x j A characteristic distance D2 therebetween; s9, integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment, and taking the abnormal working condition as a second prediction result; s10, effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result. The method provided by the invention takes the relevance of the operation parameters and the relevance of the time sequence into consideration, so that more complete equipment abnormality information and more advanced prediction results can be obtained; innovative proposal of constructing a correlation bipartite graph of operation parameters and fault events, and converting the prediction problem of faults intoThe problem of bipartite graph prediction is solved, the advantages of bipartite graph are utilized to realize fault prediction, and a first prediction result and a second prediction result obtained by the bipartite graph are effectively fused to obtain a final equipment fault prediction result; the fault prediction method can achieve the aim of high speed and high efficiency.
2. According to the invention, the second prediction result is obtained by calculating the characteristic distance between the operation parameter at the current time and the operation parameter characteristics under the two pre-assumed fault states and comprehensively comparing the two characteristic distances, so that the prediction of the fault is realized, and the accuracy of the fault prediction is greatly improved.
Drawings
FIG. 1 is a flow chart of the implementation of the steps of the method of the present invention.
Detailed Description
The present application is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only the portions related to the invention are shown in the drawings.
The invention relates to an industrial equipment, in particular to an intelligent comprehensive energy-saving cabinet (intelligent voltage optimizing device), which integrates a three-phase magnetic balance technology and a compensation type voltage regulating and stabilizing technology, adopts a 'Z-shaped isolation compensation transformer' with an iron core and a plurality of windings and special wiring to be combined with a 'column type voltage regulating transformer', and is a novel environment-friendly energy-saving technology system designed for reducing equipment failure rate, prolonging equipment service life and saving electricity expenditure of users based on optimizing equipment working voltage. The normal use condition of the intelligent voltage optimizing device is as follows: (1) ambient temperature: -15-40 ℃; (2) altitude: not more than 1000 meters; and (3) injection: when the altitude exceeds 1000 meters, the load capacity of the regulated power supply will decrease with increasing altitude. (3) relative humidity: less than or equal to 90 percent; (4) The installation place should be free from gases, vapors, chemical deposits, dust, dirt and other explosive and aggressive media that seriously affect the insulating strength of the regulated power supply; (5) the installation site should be free of severe vibration or jolt. The intelligent voltage optimizing device is used for timely predicting and alarming abnormal conditions such as under-voltage, overload, overcurrent and over-temperature related to the intelligent comprehensive energy-saving cabinet (intelligent voltage optimizing device), and has important effect on use safety and reliability.
As shown in fig. 1, the present embodiment provides a device failure fusion prediction method based on associated parameter mining, which includes the following steps:
s1, collecting historical operation parameters and time sequence data of equipment;
specifically, historical operation parameters and corresponding time sequence data of the intelligent comprehensive energy-saving cabinet (intelligent voltage optimizing device) are collected; the operating parameters include: input voltage, output voltage, operating current, operating temperature, load conditions, and the like.
S2, carrying out sequence normalization on each fault data based on the time sequence data to obtain a fault event sequence set;
specifically, the fault data are integrated according to the time sequence data to form a fault event sequence set.
S3, constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
s4, predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time, and taking the predicted abnormal working condition as a first prediction result;
s5, segment linearization vector representation is carried out on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Operating parameter characteristics of the kth time sequence segment;
specifically, for example, if the time series data relates to data of 1 month, the time series division may be performed at intervals meeting the requirements such as every half minute, every 1 hour, or the like;
if the time series data relates to data of 1 whole year, the time series segmentation can be performed at intervals meeting requirements such as every 1 minute, every 1 hour, every 1 day, and the like.
S6, assuming that the ith time sequence segment is in a fault-free state, the corresponding operation parameter characteristic is x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j ;
S7, calculating the operation parameter x of the current time a a And x i Characteristic distance D1 between:whereinIs a correction coefficient;
s8, calculating the operation parameter x of the current time a a And x j Characteristic distance D2 between:,representing the 2 norms of the vectors;
s9, integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment, and taking the abnormal working condition as a second prediction result;
s10, effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result.
Specifically, the step S9 of predicting the abnormal working condition of the device by integrating the value ranges of D1 and D2, and using the abnormal working condition as the second prediction result specifically includes:
operating parameter x for current time a a And x i The characteristic distance between the two is setThreshold value;
When D1 is smaller thanAnd D2 is greater thanWhen the equipment is in fault, the equipment at the current time is indicated to be in fault;
when D1 is greater thanAnd D2 is less thanWhen the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater thanWhen the equipment at the current time is in fault, the possibility of the equipment at the current time is indicated;
when D1 is smaller thanAnd D2 is less thanWhen the device is in fault, the possibility of the device at the current time is indicated.
The method specifically comprises the step of S10, effectively fusing a first prediction result and a second prediction result to obtain a final equipment failure prediction result, wherein the first prediction result and the second prediction result are fused through a decision-making level model obtained through a machine learning method to obtain the final equipment failure prediction result.
Specifically, the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises the following steps:
taking fault data related to equipment history as training data to perform data preprocessing to form a training set and a testing set; the sample data ratio of the training set to the test set is 4:1.
Training the decision-level model by utilizing the training set;
testing the decision-level model by using the test set;
and updating model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
Specifically, inputting a training set into a decision-level model for machine learning to preliminarily obtain a fault prediction model; and inputting the test set into the preliminarily obtained prediction model, and finally, carrying out parameter adjustment on the model according to the test accuracy, thereby obtaining the optimal decision-making model.
Based on the same inventive concept, the device fault fusion prediction system based on associated parameter mining disclosed in this embodiment includes:
the acquisition module is used for acquiring historical operation parameters and time sequence data of the equipment;
specifically, historical operation parameters and corresponding time sequence data of the intelligent comprehensive energy-saving cabinet (intelligent voltage optimizing device) are collected; the operating parameters include: input voltage, output voltage, operating current, operating temperature, load conditions, and the like.
The system comprises a normalization module, a fault event sequence collection module and a fault event sequence collection module, wherein the normalization module is used for performing sequence normalization on each fault data based on time sequence data to obtain a fault event sequence set;
specifically, the fault data are integrated according to the time sequence data to form a fault event sequence set.
The construction module is used for constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
the prediction module 1 is used for predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time and taking the predicted abnormal working condition as a first prediction result;
the segmentation module is used for carrying out piecewise linearization vector representation on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Operating parameter characteristics of the kth time sequence segment;
specifically, for example, if the time series data relates to data of 1 month, the time series division may be performed at intervals meeting the requirements such as every half minute, every 1 hour, or the like;
if the time series data relates to data of 1 whole year, the time series segmentation can be performed at intervals meeting requirements such as every 1 minute, every 1 hour, every 1 day, and the like.
Specifically, assuming that the ith time sequence segment is in a fault-free state, the corresponding operating parameter characteristic is x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j ;
A calculation module 1 for calculating an operation parameter x of the current time a a And x i Characteristic distance D1 between:whereinIs a correction coefficient;
a calculation module 2 for calculating an operation parameter x of the current time a a And x j Characteristic distance D2 between:,representing the 2 norms of the vectors;
the prediction module 2 is used for integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment and taking the abnormal working condition as a second prediction result;
and the fusion prediction module is used for effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result.
Specifically, the method for predicting the abnormal working condition of the equipment by integrating the value ranges of D1 and D2 and taking the abnormal working condition as a second prediction result specifically comprises the following steps:
When D1 is smaller thanAnd D2 is greater thanWhen the equipment is in fault, the equipment at the current time is indicated to be in fault;
when D1 is greater thanAnd D2 is less thanWhen the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater thanWhen the equipment at the current time is in fault, the possibility of the equipment at the current time is indicated;
when D1 is smaller thanAnd D2 is less thanWhen the device is in fault, the possibility of the device at the current time is indicated.
The method comprises the steps of fusing a first prediction result and a second prediction result through a decision-making level model obtained through a machine learning method to obtain a final equipment failure prediction result.
Specifically, the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises the following steps:
the preprocessing module is used for preprocessing data by taking fault data related to equipment history as training data to form a training set and a testing set; the sample data ratio of the training set to the test set is 4:1.
The training module is used for training the decision-level model by utilizing the training set;
the test module is used for testing the decision-level model by utilizing the test set;
and the iterative training module is used for updating the model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
Specifically, inputting a training set into a decision-level model for machine learning to preliminarily obtain a fault prediction model; and inputting the test set into the preliminarily obtained prediction model, and finally, carrying out parameter adjustment on the model according to the test accuracy, thereby obtaining the optimal decision-making model.
Claims (6)
1. The equipment fault fusion prediction method based on the association parameter mining is characterized by comprising the following steps of:
s1, collecting historical operation parameters and time sequence data of equipment;
s2, carrying out sequence normalization on each fault data based on the time sequence data to obtain a fault event sequence set;
s3, constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
s4, predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the running parameter of the current time, and taking the predicted abnormal working condition as a first prediction result;
s5, segment linearization vector representation is carried out on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Operating parameter characteristics of the kth time sequence segment;
s6, assuming that the ith time sequence segment is in a fault-free state, the corresponding operation parameter characteristic is x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming that the jth sequential sequence segment is in fault state, the corresponding operation parameter feature is x j ;
S7, calculating the operation parameter of the current time ax a And x i Characteristic distance D1 between:whereinIs a correction coefficient;
s8, calculating the operation parameter x of the current time a a And x j Characteristic distance D2 between:,representing the 2 norms of the vectors;
s9, integrating the value ranges of the D1 and the D2, predicting the abnormal working condition of the equipment, and taking the abnormal working condition as a second prediction result;
s10, effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result;
s9, the value ranges of D1 and D2 are synthesized, the abnormal working condition of the equipment is predicted, and the abnormal working condition is used as a second prediction result, and the method specifically comprises the following steps:
When D1 is smaller thanAnd D2 is greater than->When it is, then indicate the currentThe time equipment has no faults;
when D1 is greater thanAnd D2 is less than->When the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater than->When the equipment at the current time is in fault, the possibility of the equipment at the current time is indicated;
2. The method for predicting the equipment failure fusion based on the association parameter mining according to claim 1, wherein the step S10 is characterized in that the first prediction result and the second prediction result are effectively fused to obtain a final equipment failure prediction result, and specifically comprises the step of fusing the first prediction result and the second prediction result through a decision-making level model obtained by a machine learning method to obtain the final equipment failure prediction result.
3. The method for predicting equipment failure fusion based on associated parameter mining according to claim 2, wherein the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises the following steps:
taking fault data related to equipment history as training data to perform data preprocessing to form a training set and a testing set;
training the decision-level model by utilizing the training set;
testing the decision-level model by using the test set;
and updating model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
4. An equipment fault fusion prediction system based on associated parameter mining, which is characterized by comprising:
the acquisition module is used for acquiring historical operation parameters and time sequence data of the equipment;
the system comprises a normalization module, a fault event sequence collection module and a fault event sequence collection module, wherein the normalization module is used for performing sequence normalization on each fault data based on time sequence data to obtain a fault event sequence set;
the construction module is used for constructing a correlation bipartite graph of the operation parameters and the fault events based on the fault event sequence set; the nodes in the relevance bipartite graph comprise fault object nodes and operation parameter feature nodes, and the connected edges between the fault object nodes and the operation parameter feature nodes represent the relevance relationship between the fault object nodes and the operation parameter feature nodes;
the first prediction module is used for predicting the abnormal working condition of the equipment according to the relevance bipartite graph for the operation parameters of the current time and taking the predicted abnormal working condition as a first prediction result;
the segmentation module is used for carrying out piecewise linearization vector representation on the time sequence data according to a fixed time distance: x is X K ={x 1 ,x 2 ,…, x K-1 , x K -a }; k is the total number of sequential sequence segments, k=1, 2,; x is x 1 Operating parameter characteristics for the first time sequence segment; x is x 2 Operating parameter characteristics for the second time sequence segment; x is x k-1 Operating parameter characteristics for the (k-1) th time sequence segment; x is x k Operating parameter characteristics of the kth time sequence segment;
assuming that the ith time sequence segment is in a fault-free state, the corresponding operating parameter is characterized as x i The method comprises the steps of carrying out a first treatment on the surface of the Assuming the jth sequential sequence segment is in a faulty state, it is toThe corresponding operating parameter is characterized as x j ;
A first calculation module for calculating the operation parameter x of the current time a a And x i Characteristic distance D1 between:wherein->Is a correction coefficient;
a second calculation module for calculating the operation parameter x of the current time a a And x j Characteristic distance D2 between:,/>representing the 2 norms of the vectors;
the second prediction module is used for synthesizing the value ranges of D1 and D2, predicting the abnormal working condition of the equipment and taking the abnormal working condition as a second prediction result;
the fusion prediction module is used for effectively fusing the first prediction result and the second prediction result to obtain a final equipment failure prediction result;
the method for predicting the abnormal working condition of the equipment by integrating the value ranges of D1 and D2 and taking the abnormal working condition as a second prediction result specifically comprises the following steps:
When D1 is smaller thanAnd D2 is greater than->When the equipment is in fault, the equipment at the current time is indicated to be in fault;
when D1 is greater thanAnd D2 is less than->When the equipment at the current time is in fault, the equipment at the current time is indicated;
when D1 is greater thanAnd D2 is greater than->When the device is at the present time, the possibility of failure of the device is indicated>
5. The equipment failure fusion prediction system based on the association parameter mining, as set forth in claim 4, is characterized in that the first prediction result and the second prediction result are effectively fused to obtain a final equipment failure prediction result, and specifically comprises the steps of fusing the first prediction result and the second prediction result through a decision-making level model obtained by a machine learning method to obtain the final equipment failure prediction result.
6. The system for predicting equipment failure fusion based on associated parameter mining of claim 5, wherein the decision-level model obtained by the machine learning method is obtained through pre-training, and specifically comprises:
the preprocessing module is used for preprocessing data by taking fault data related to equipment history as training data to form a training set and a testing set;
the training module is used for training the decision-level model by utilizing the training set;
the test module is used for testing the decision-level model by utilizing the test set;
and the iterative training module is used for updating the model parameters of the decision-making level model according to the test result, and obtaining the decision-making level model through iterative training.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310001115.1A CN115689071B (en) | 2023-01-03 | 2023-01-03 | Equipment fault fusion prediction method and system based on associated parameter mining |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310001115.1A CN115689071B (en) | 2023-01-03 | 2023-01-03 | Equipment fault fusion prediction method and system based on associated parameter mining |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115689071A CN115689071A (en) | 2023-02-03 |
CN115689071B true CN115689071B (en) | 2023-05-02 |
Family
ID=85057536
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310001115.1A Active CN115689071B (en) | 2023-01-03 | 2023-01-03 | Equipment fault fusion prediction method and system based on associated parameter mining |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115689071B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116755403B (en) * | 2023-06-13 | 2024-03-26 | 英利新能源(宁夏)有限公司 | Data acquisition method and system based on photovoltaic module production control system |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110008565A (en) * | 2019-03-28 | 2019-07-12 | 浙江大学 | A kind of industrial process unusual service condition prediction technique based on operating parameter association analysis |
CN112906969A (en) * | 2021-03-01 | 2021-06-04 | 三一重工股份有限公司 | Engine fault prediction method, engine fault prediction device, electronic equipment and storage medium |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110018670B (en) * | 2019-03-28 | 2020-07-10 | 浙江大学 | Industrial process abnormal working condition prediction method based on dynamic association rule mining |
CN110008253B (en) * | 2019-03-28 | 2021-02-23 | 浙江大学 | Industrial data association rule mining and abnormal working condition prediction method |
CN112801426B (en) * | 2021-04-06 | 2021-06-22 | 浙江浙能技术研究院有限公司 | Industrial process fault fusion prediction method based on correlation parameter mining |
CN114818515A (en) * | 2022-06-24 | 2022-07-29 | 中国海洋大学 | Multidimensional time sequence prediction method based on self-attention mechanism and graph convolution network |
-
2023
- 2023-01-03 CN CN202310001115.1A patent/CN115689071B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110008565A (en) * | 2019-03-28 | 2019-07-12 | 浙江大学 | A kind of industrial process unusual service condition prediction technique based on operating parameter association analysis |
CN112906969A (en) * | 2021-03-01 | 2021-06-04 | 三一重工股份有限公司 | Engine fault prediction method, engine fault prediction device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN115689071A (en) | 2023-02-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhu et al. | Imbalance learning machine-based power system short-term voltage stability assessment | |
CN111639921B (en) | Intelligent equipment fault judgment and auxiliary disposal method based on expert system | |
CN115689071B (en) | Equipment fault fusion prediction method and system based on associated parameter mining | |
CN104504607A (en) | Method for diagnosing photovoltaic power station faults on the basis of fuzzy clustering algorithm | |
CN107844628B (en) | Redundancy optimization method for large-scale offshore wind farm power collection system | |
CN106327359A (en) | Electricity consumption mode analysis-based meter centralized reading data anomaly judgment method | |
CN116031867A (en) | Power grid transient safety stability assessment method | |
CN114692787B (en) | Multi-source data-based three-winding main transformer event analysis processing method and system | |
CN115358355B (en) | Method and device for judging main transformer oil temperature gauge and top layer oil temperature abnormity | |
CN116204839A (en) | Transformer oil temperature anomaly analysis method | |
Wang et al. | State prediction of power transformer based on grey-lagrange method with weighted coefficient of variation | |
CN114355182A (en) | Online monitoring and fault diagnosis method for high-voltage circuit breaker | |
Huang et al. | Extracting substation cyber-physical architecture through intelligent electronic devices' data | |
CN112434955A (en) | Distribution network line operation risk sensing method based on multi-data fusion | |
CN112529395A (en) | Power transmission line typical ice wind disaster analysis method based on fusion of FCE and SVM | |
CN112001073A (en) | Reliability analysis research method of traction power supply system | |
Qin et al. | An Identification Method of Metering Anomaly Based on Line Loss Analysis of Low Voltage Station | |
de Oliveira et al. | Voltage dips caused by faults in a transmission system: a monitoring case study of a sensitive industrial consumer | |
Xie et al. | Large power transformer condition evaluation based on multilevel extension theory | |
Zeng et al. | Fault Diagnosis Analysis of Distribution Network Based on Compound Statistics of Residual and Score | |
CN117648588B (en) | Meteorological radar parameter anomaly identification method based on correlation network graph cluster analysis | |
CN118261577A (en) | Operation state grading method for power equipment of substation in front-end system | |
Bai et al. | Abnormal Detection Scheme of Substation Equipment based on Intelligent Fusion Terminal | |
Liu et al. | A Multilevel Diagnosis Method for Large Transformer Fault Modes under Multi-sensor Information Fusion | |
Du et al. | Fault Diagnosis Based on the Adaptive Network Fuzzy Inference System for Power System of Offshore Platform |
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 | ||
TR01 | Transfer of patent right |
Effective date of registration: 20240208 Address after: No.198 GANGLONG Road, Yangzhong Development Zone, Zhenjiang City, Jiangsu Province, 212000 Patentee after: Zhenjiang Anhua Electric Group Co.,Ltd. Country or region after: China Address before: Room 1001-3, Tower C, Zhimengyuan, 4 Gupinggang, Gulou District, Nanjing, Jiangsu, 210000 Patentee before: Nanjing University of Technology Jinhong Energy Technology Co.,Ltd. Country or region before: China |
|
TR01 | Transfer of patent right |