CN116796617A - Rolling bearing equipment residual life prediction method and system based on data identification - Google Patents
Rolling bearing equipment residual life prediction method and system based on data identification Download PDFInfo
- Publication number
- CN116796617A CN116796617A CN202210244921.7A CN202210244921A CN116796617A CN 116796617 A CN116796617 A CN 116796617A CN 202210244921 A CN202210244921 A CN 202210244921A CN 116796617 A CN116796617 A CN 116796617A
- Authority
- CN
- China
- Prior art keywords
- data
- rolling bearing
- neural network
- prediction
- data identification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005096 rolling process Methods 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 claims abstract description 41
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 18
- 230000015654 memory Effects 0.000 claims abstract description 17
- 238000003860 storage Methods 0.000 claims abstract description 7
- 108090000623 proteins and genes Proteins 0.000 claims description 33
- 238000004422 calculation algorithm Methods 0.000 claims description 29
- 230000002068 genetic effect Effects 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 25
- 238000007781 pre-processing Methods 0.000 claims description 10
- 230000007787 long-term memory Effects 0.000 claims description 9
- 238000005215 recombination Methods 0.000 claims description 9
- 230000006798 recombination Effects 0.000 claims description 9
- 230000035772 mutation Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 210000000349 chromosome Anatomy 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 239000012634 fragment Substances 0.000 claims description 2
- 238000002372 labelling Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims 1
- 239000000284 extract Substances 0.000 abstract description 4
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000004519 manufacturing process Methods 0.000 description 8
- 238000012544 monitoring process Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000012423 maintenance Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000002921 genetic algorithm search Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 229940050561 matrix product Drugs 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
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
- G06Q10/04—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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Game Theory and Decision Science (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for predicting the residual life of rolling bearing equipment based on data identification, and belongs to the field of state detection and life prediction of rotating machinery equipment. The invention uses a method of state characteristic identification to carry out certain rule identification on state storage data, combines a convolutional neural network and a two-way long-short-term memory network to form a hybrid neural network so as to effectively extract time and space characteristics and improve the prediction precision of the residual life.
Description
Technical Field
The invention belongs to the field of state detection and life prediction of rotating machinery equipment, and particularly relates to a method and a system for predicting the residual service life of rolling bearing equipment based on data identification.
Background
With the large-scale development of petrochemical industry scale, the rotary mechanical equipment in the large unit in the petrochemical industry increasingly tends to develop in the directions of large-scale, precise, high-speed and automatic, and the composition and structure of the rotary mechanical equipment are more and more complex. These developments on the one hand increase the production efficiency and reduce the production costs, and on the other hand place higher and more stringent demands on the design, manufacture, installation, use, maintenance and reliable operation of the machine. A minor failure may cause degradation, deterioration or even failure of the performance of the device, affecting the stability and safety of the overall system operation, and severely causing catastrophic accidents. Rolling bearings are important components of rotating machinery, and if the residual service life of the rolling bearings can be accurately predicted before the rolling bearings fail, preventive measures can be timely taken so as to avoid causing serious economic loss and even casualties. And under the condition of increasingly severe requirements, in order to adapt to different requirements, an identification system has to be introduced in the manufacturing process to meet the information management requirements.
As shown in fig. 1, in the prior art, the remaining life prediction techniques of mechanical devices are mainly divided into four categories: a method based on a physical model, a method based on a statistical model, a method based on machine learning, and a fusion method. Deep learning algorithms are often used to automatically obtain deeper abstract features from large-scale data, which are widely used in the field of residual life prediction. Common methods include convolutional neural networks, recurrent neural networks, long and short term memory artificial neural networks, gated recurrent units, and the like. In most conventional approaches in existence, data points of different time steps are typically assigned the same weight. In practice, however, these data points with different time steps do not provide as much information and therefore result in less than ideal prediction accuracy. At the same time, some factories still use physical tags to manage identification object information. Thus, there is a variation in the identification management system from system to system, enterprise to enterprise, and industry to industry in existing manufacturing systems. In the collaboration process, more effort and manpower are required to deal with the problems caused by the data non-standardization.
Disclosure of Invention
Aiming at solving the technical problem of low prediction precision of the residual service life existing in the traditional method, the invention provides a rolling bearing residual service life prediction method and a rolling bearing residual service life prediction system based on data identification, and the method of state characteristic identification is used for carrying out certain rule identification on state storage data; the method adopts a cooperative training mechanism comprising a mixed neural network of a convolutional neural network and a two-way long-short-term memory network and an optimized genetic algorithm, distributes different weights to data points of different time steps to highlight the data points containing more degradation information, accurately diagnoses the state of the data points, and solves the problem that the fault diagnosis and the life prediction of the state monitoring data cannot be effectively utilized.
The technical scheme adopted by the invention for achieving the purpose is as follows:
the method for predicting the residual life of the rolling bearing equipment based on the data identification comprises the following steps:
acquiring running data of rolling bearing equipment, and identifying and preprocessing the running data;
constructing a hybrid neural network as a prediction model, and training the hybrid neural network by using a genetic algorithm;
and inputting running data of the rolling bearing equipment to be predicted into a trained prediction model to obtain a residual life prediction result.
The rolling bearing device operating data is identified as labeling its state and attribute information.
The preprocessing is to perform normalization processing on running data of the rolling bearing equipment, and segment the normalization data by using a sliding window mode.
The hybrid neural network comprises a convolutional neural network and a two-way long-short-term memory network, wherein the input of the convolutional neural network is data to be predicted, the output of the convolutional neural network is used as the input of the two-way long-short-term memory network, and the output of the two-way long-term memory network is a residual life prediction result.
The genetic algorithm comprises the following steps:
initializing a population: the attention weights are coded on the chromosome of the genetic space through binary codes, so that each attention weight corresponds to a string of binary codes;
weight transfer: converting binary codes into decimal numbers as attention weights, transferring the attention weights into a hybrid neural network, and starting training and learning by the hybrid neural network and returning loss values generated by prediction errors;
loss ordering: each attention weight is taken as a group, each group of binary codes is randomly selected and divided into a plurality of groups, the steps are iterated according to the returned loss value sequence until the loss value reaches the minimum, and the attention weight code with the minimum error value in each group is selected;
cross recombination: combining the minimum error value sets selected in the previous step in a pairwise random manner to obtain a plurality of sets I and a set II, randomly selecting part of gene points in the set I, providing the remaining unselected gene points by the set II, recombining the gene points in the two sets as a new gene set, and obtaining a recombined gene segment;
variation: performing original gene inversion operation on the gene fragments subjected to cross recombination according to mutation probability;
generating a new population: reconstructing the population according to the results of the cross recombination and mutation, and starting new training until the obtained loss value is minimum, wherein the attention weight corresponding to the loss value is used as the weight in the mixed neural network.
A rolling bearing device remaining life prediction system based on data identification, comprising:
the data preprocessing module is used for acquiring the running data of the rolling bearing equipment, and identifying and preprocessing the running data;
the prediction model construction and training module is used for constructing a hybrid neural network as a prediction model and training the hybrid neural network by using a genetic algorithm;
and the data prediction module is used for inputting running data of the rolling bearing equipment to be predicted into the trained prediction model to obtain a residual life prediction result.
A rolling bearing device remaining life prediction system based on data identification, comprising a memory and a processor; the memory is used for storing a computer program; the processor is used for realizing the rolling bearing equipment residual life prediction method based on the data identification when executing the computer program.
A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of predicting remaining life of a rolling bearing device based on data identification.
The invention has the following beneficial effects and advantages:
1. the invention uses the unified mark of the mark analysis system to mark the material code, the process code, the equipment code, the personnel code, the part code, the product code, the package code, the working procedure code and the like in the production and manufacturing process, and firstly realizes the management of one object and one code, and binds the physical entity with the virtual data. The method realizes the fine management and control of information of each link in the production process, realizes the sharing of key information data, and realizes the continuity and integrity of data to a certain extent.
2. The invention combines the convolutional neural network and the two-way long-short-term memory network to form the hybrid neural network so as to effectively extract time and space characteristics and improve the prediction accuracy of the residual life, on the basis, attention weights are introduced into the hybrid neural network, a cooperative training mechanism of the hybrid neural network and an optimized genetic algorithm is adopted, the hybrid neural network shares, transmits parameters and loss feedback, the genetic algorithm searches the optimal attention weight parameters by continuously transmitting the parameters and the feedback loss, the attention weight distribution of different time steps in the prediction of the residual life is optimized, the importance of different time steps is accurately embodied, and compared with some traditional machine learning methods and deep learning (a multi-layer perceptron, support vector regression, random forest, convolutional neural network, cyclic neural network, deep belief network and the like), the root mean square error and the prediction score are much lower, and the problem of the existing prediction accuracy of the residual life is effectively solved.
3. The invention also improves the traditional genetic algorithm, prevents the algorithm from falling into a local optimal state, and introduces the loss feedback of the hybrid neural network into the improved genetic algorithm, so that the process can obviously improve the performance by controlling the direction of random search.
Drawings
FIG. 1 is a classification diagram of a RUL prediction method of a current mechanical device;
FIG. 2 is a flow chart of a method for predicting the residual life of a bearing of a genetic optimization hybrid neural network based on data identification;
FIG. 3 is a basic structure diagram of an LSTM network with long-term memory provided by the invention;
fig. 4 is a block diagram of a system for predicting remaining service life of mechanical equipment provided by the invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
The invention provides a rolling bearing equipment residual life prediction method based on data identification, which is shown in figure 2 and comprises the following steps: marking and preprocessing the running data of the rolling bearing equipment to be used as a marking training set and a marking verification set; model building and training: firstly, transmitting initialized attention weights to a convolution-two-way long-short-term memory hybrid neural network, training a model, feeding back loss to an improved genetic algorithm, continuously learning the genetic algorithm to obtain an optimal attention weight set, transmitting the optimal attention weight set to the hybrid neural network, achieving cooperative training, finding an optimal solution, and returning to the model when the loss on a verification set is minimum; and predicting the returned model on the test set, and returning a residual life prediction result. The model built by the invention is formed by a convolutional neural network and a two-way long-short-term memory network, wherein the convolutional neural network extracts the spatial characteristics of data points, the two-way long-short-term memory network acquires the time correlation of the data, and the two-way long-term memory networks can extract the spatial characteristics and the time characteristics by combining the two-way long-term memory network and learn the complete characteristics in the data.
Normalizing the data to eliminate effects between different sensor characteristics, and segmenting the data with a sliding window for data with longer time, wherein the specific implementation mode is as follows:
normalization is performed by the following formula:
wherein x is i,j Is the raw sensor data of the sensor,is the true value of the normalized data, i represents the ith sensor, j represents the jth data point,/->Is the maximum and minimum values of the ith sensor data.
The sliding window is segmented, and in order to capture the time correlation between the data, the sliding window is used to encapsulate the data at adjacent time points, and the remaining life prediction of the last data point in the time window is used as the remaining life prediction of the time window. By analysis, the data sliding window time length was set to 30 and the step size was set to 1.
Preferably, the mixed neural network with attention weight for training consists of a convolutional neural network and a two-way long-short-term memory network, and the whole is as follows:
Y＝f H (x)
wherein x is E R m×n Is a vector comprising m time steps, n features, Y is a network output representing a residual life prediction, f H (. Cndot.) is a neural network composed of a nonlinear function that maps an input to an output.
The training is specifically as follows:
first, the original input data is selected by sliding a window, and then the attention weights for different time steps are defined as follows:
i denotes the i-th weight set, l denotes the i-th time step,representing the weight of the ith time step in the ith weight set;
by weighting the attention weights of different time steps, the following input data can be obtained:
a value representing the first data point in the t-th sample;
thereby, the data value x= [ X ] to which the attention weight is assigned 1 ,X 2 ,...,X T ]Transmitting to convolution-two-way long-short-term memory hybrid neural network X T The data point representing the T-th sample is multiplied by the value after the attention weight.
The convolutional neural network is used to acquire spatial features of the attention weighted input data, and the number of channels is set to a time step in order not to compress the time correlation of the data. Thus, convolutional neural networks do not interfere with the temporal relationship in the data, but only compress the spatial information. And then inputting the output data into a two-way long-short-term memory recurrent neural network.
As shown in fig. 3, the long-term memory network learns the spatial features extracted by the convolutional neural network and learns the time-dependent dependency relationship, and the units thereof are calculated as follows:
wherein X is t ,h t-1 ,h t ,z t ,r t ,Representing the input vector, the state storage variable at the previous time, the state storage variable at the current time, the state of the update gate, the state of the reset gate, the state of the current candidate set, and the state of the output vector at the current time, respectively. W (W) z ,W r ,/>Respectively represent update gate, reset gate, candidate set, output vector and output vector of x t And h t-1 And (5) forming the weight parameters of the connection matrix. I represents an identity matrix; []Representing vector connections; representing a matrix dot product; x represents the matrix product; sigma represents an sigmoid activation function; tanh represents the tanh activation function. b r 、b z 、b h The bias is learned in the training process;
the genetic algorithm sends the learned attention weights back to the hybrid neural network. At the same time, the hybrid network returns training loss to the genetic algorithm to guide its further learning and training. The invention can adopt the traditional genetic algorithm to cooperatively train with the mixed neural network, and preferably, the invention also improves the traditional genetic algorithm as follows:
initializing a population: the attention weight is coded on a chromosome of the genetic space through binary codes, and each weight corresponds to a string of binary gene codes;
weight transfer: converting binary codes into decimal numbers, namely initial attention weight values, transferring the attention weight values into a previous hybrid neural network, starting training and learning by the neural network, and returning corresponding loss values generated by prediction errors of the neural network;
loss ordering: randomly selecting and dividing each group of binary codes into a plurality of groups, and selecting the weight code with the smallest error value in each group according to the returned loss sequencing (prediction error);
cross recombination: combining the error minimum sets selected in the previous step in a pairwise random manner to obtain a plurality of sets I and a set II, randomly selecting part of gene points in the set I, providing the rest unselected gene points by the set II, recombining the gene points in the two sets into a new gene set, and obtaining a recombined gene segment; the selected set of gene spots is not fixed, 50% is chosen in this example; the common genetic algorithm is that random gene selection points are interchanged, but the method is not the recombination, the gene points in the first set are randomly selected, and the gene points in the first set are recombined with the gene points in the second set which are not selected to obtain a new genotype. For example: two six-bit gene segments are provided, three gene points of 1.3.5 in the first set are randomly selected (246 is not selected and 246 is provided by the second set), and then the gene points of the first set 135 and the gene points of the second set 246 are recombined into a new 6-bit gene segment; variation: performing original gene inversion operation on the genotype subjected to cross recombination according to mutation probability;
generating a new population: the population is reconstructed from the crossover and mutation and a new training is started.
The improved genetic algorithm of the invention differs from the common genetic algorithm in that:
the selection strategy in the common genetic algorithm generally calculates the occurrence probability of each individual in the filial generation to randomly select the individual to form a population, but the invention obtains a new set by randomly combining the selected minimum error set two by two to carry out the next cross mutation, so that the cross mutation process has more possibility. The crossover operator in the common genetic algorithm is mainly single-point or multi-point crossover, the crossover operator is used for carrying out gene exchange by randomly selecting crossover points, and the finally obtained new gene segment also comprises partial original genotypes.
Compared with the common genetic algorithm, the improved genetic algorithm mainly adds a random mechanism in a selection strategy and a crossover operator to reconstruct the final population space, so that the global searching capability of the algorithm is stronger.
Furthermore, the present invention introduces the loss feedback of the previous hybrid neural network into an improved genetic algorithm, so that the improvement can allow the process to significantly improve performance by controlling the direction of the random search.
In summary, the above technical solutions conceived by the present invention enable higher prediction accuracy to be achieved with relatively limited technical complexity compared to the prior art.
Experiment verification
As shown in fig. 4, the present embodiment provides a remaining service life prediction system for a mechanical device, including: a history database, a life prediction model, and a full life cycle database;
the historical database is used for storing all state monitoring data accumulated by equipment with the same model and under the same working condition, and taking the state monitoring data as training data to train a life prediction model;
the life prediction model comprises a deep neural network life prediction model which is constructed by taking a time convolution network as a characteristic extraction algorithm and a long-term and short-term memory network as a regression prediction algorithm;
the full life cycle database stores the collected real-time running data of the tested equipment, and constructs the collected real-time running data of the tested equipment into a life prediction data set with time sequence characteristics according to the model number of the tested equipment and the data collection time sequence;
and the full life cycle database inputs the life prediction data set into a life prediction module, and the trained deep neural network life prediction model predicts the life prediction data set to obtain the residual service life of the tested equipment.
The device model, the device operation configuration, the real-time data acquired during the sensor and the like are stored in the full life cycle database, the tested device is in a life cycle from the initial operation stage to the occurrence of a fault, the acquired real-time operation data of the tested device are stored in the full life cycle database, and when one life cycle of the tested device is finished, the operation data of the tested device in the life cycle is added into the historical database.
The system also comprises an operation monitoring module, wherein the operation monitoring module is used for collecting the operation data of the tested equipment in real time through a state monitoring sensor and an acceleration sensor which are arranged on the tested equipment, and storing the operation data into a full life cycle database.
In the embodiment, a life prediction model database is established by collecting historical data and enhancing the data; combining a convolutional neural network and a two-way long-short-term memory network, constructing a residual service life prediction model, and training the model by using the established historical database; and measuring vibration parameters and other parameters of the surface of the monitored equipment, preprocessing the data, and predicting the service life by using a trained prediction model. Judging the health condition of the equipment according to the prediction result, finding possible faults of the equipment in time, analyzing the position and severity of the faults, and providing different maintenance schemes according to the analysis result. The complexity of equipment maintenance is greatly reduced, the timeliness and scientificity of the equipment maintenance are improved, the equipment maintenance cost is reduced, and the safe and reliable operation of the equipment is ensured.
It will be readily appreciated by persons skilled in the art that the foregoing description is not intended to limit the invention, and any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention are intended to be included within the scope of the present invention.
Claims (8)
1. The method for predicting the residual life of the rolling bearing equipment based on the data identification is characterized by comprising the following steps of:
acquiring running data of rolling bearing equipment, and identifying and preprocessing the running data;
constructing a hybrid neural network as a prediction model, and training the hybrid neural network by using a genetic algorithm;
and inputting running data of the rolling bearing equipment to be predicted into a trained prediction model to obtain a residual life prediction result.
2. The method for predicting the remaining life of a rolling bearing apparatus based on data identification according to claim 1, wherein the rolling bearing apparatus operation data is identified as labeling its state and attribute information.
3. The method for predicting the remaining life of a rolling bearing device based on data identification according to claim 1, wherein the preprocessing is normalization processing of rolling bearing device operation data and segmentation of the normalization data is performed by using a sliding window.
4. The method for predicting the residual life of rolling bearing equipment based on data identification according to claim 1, wherein the hybrid neural network comprises a convolutional neural network and a two-way long-short-term memory network, the input of the convolutional neural network is data to be predicted, the output of the convolutional neural network is used as the input of the two-way long-term memory network, and the output of the two-way long-term memory network is a residual life prediction result.
5. The method for predicting the remaining life of a rolling bearing device based on data identification as claimed in claim 1, wherein the genetic algorithm comprises the steps of:
initializing a population: the attention weights are coded on the chromosome of the genetic space through binary codes, so that each attention weight corresponds to a string of binary codes;
weight transfer: converting binary codes into decimal numbers as attention weights, transferring the attention weights into a hybrid neural network, and starting training and learning by the hybrid neural network and returning loss values generated by prediction errors;
loss ordering: each attention weight is taken as a group, each group of binary codes is randomly selected and divided into a plurality of groups, the steps are iterated according to the returned loss value sequence until the loss value reaches the minimum, and the attention weight code with the minimum error value in each group is selected;
cross recombination: combining the minimum error value sets selected in the previous step in a pairwise random manner to obtain a plurality of sets I and a set II, randomly selecting part of gene points in the set I, providing the remaining unselected gene points by the set II, recombining the gene points in the two sets as a new gene set, and obtaining a recombined gene segment;
variation: performing original gene inversion operation on the gene fragments subjected to cross recombination according to mutation probability;
generating a new population: reconstructing the population according to the results of the cross recombination and mutation, and starting new training until the obtained loss value is minimum, wherein the attention weight corresponding to the loss value is used as the weight in the mixed neural network.
6. A rolling bearing device remaining life prediction system based on data identification, characterized by comprising:
the data preprocessing module is used for acquiring the running data of the rolling bearing equipment, and identifying and preprocessing the running data;
the prediction model construction and training module is used for constructing a hybrid neural network as a prediction model and training the hybrid neural network by using a genetic algorithm;
and the data prediction module is used for inputting running data of the rolling bearing equipment to be predicted into the trained prediction model to obtain a residual life prediction result.
7. A rolling bearing equipment residual life prediction system based on data identification is characterized by comprising a memory and a processor; the memory is used for storing a computer program; the processor for implementing the rolling bearing device remaining life prediction method based on data identification as claimed in any one of claims 1-5 when executing the computer program.
8. A computer-readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when executed by a processor, implements the rolling bearing device remaining life prediction method based on data identification as claimed in any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210244921.7A CN116796617A (en) | 2022-03-14 | 2022-03-14 | Rolling bearing equipment residual life prediction method and system based on data identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210244921.7A CN116796617A (en) | 2022-03-14 | 2022-03-14 | Rolling bearing equipment residual life prediction method and system based on data identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116796617A true CN116796617A (en) | 2023-09-22 |
Family
ID=88040364
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210244921.7A Pending CN116796617A (en) | 2022-03-14 | 2022-03-14 | Rolling bearing equipment residual life prediction method and system based on data identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116796617A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117592383A (en) * | 2024-01-19 | 2024-02-23 | 四川晟蔚智能科技有限公司 | Method, system, equipment and medium for predicting equipment health life |
-
2022
- 2022-03-14 CN CN202210244921.7A patent/CN116796617A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117592383A (en) * | 2024-01-19 | 2024-02-23 | 四川晟蔚智能科技有限公司 | Method, system, equipment and medium for predicting equipment health life |
| CN117592383B (en) * | 2024-01-19 | 2024-03-26 | 四川晟蔚智能科技有限公司 | Method, system, equipment and medium for predicting equipment health life |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106168799B (en) | A method of batteries of electric automobile predictive maintenance is carried out based on big data machine learning | |
| CN111914873A (en) | Two-stage cloud server unsupervised anomaly prediction method | |
| CN110943857A (en) | Power communication network fault analysis and positioning method based on convolutional neural network | |
| CN113094822A (en) | Method and system for predicting residual life of mechanical equipment | |
| CN111259947A (en) | Power system fault early warning method and system based on multi-mode learning | |
| CN109740859A (en) | Transformer condition evaluation and system based on Principal Component Analysis and support vector machines | |
| CN113343581B (en) | Transformer fault diagnosis method based on graph Markov neural network | |
| CN116014722A (en) | Sub-solar photovoltaic power generation prediction method and system based on seasonal decomposition and convolution network | |
| CN116796617A (en) | Rolling bearing equipment residual life prediction method and system based on data identification | |
| CN114553671A (en) | Diagnosis method for power communication network fault alarm | |
| CN114021758A (en) | Operation and maintenance personnel intelligent recommendation method and device based on fusion of gradient lifting decision tree and logistic regression | |
| CN117557127A (en) | Power grid dispatching system supporting platform reliability assessment method, system and storage medium | |
| Pattnaik et al. | A survey on machine learning techniques used for software quality prediction | |
| Zha et al. | Recognizing plans by learning embeddings from observed action distributions | |
| CN115713027A (en) | Transformer state evaluation method, device and system | |
| CN115935814A (en) | Transformer fault prediction method based on ARIMA-SVM model | |
| CN114897262A (en) | Rail transit equipment fault prediction method based on deep learning | |
| CN113076217A (en) | Disk fault prediction method based on domestic platform | |
| CN113469013A (en) | Motor fault prediction method and system based on transfer learning and time sequence | |
| CN117454762B (en) | Coal-penetrating tunnel face gas concentration prediction method of Markov-neural network | |
| CN114297795B (en) | PR-Trans-based mechanical equipment residual life prediction method | |
| Li et al. | CapsNet-Enabled Multimodal Data Fusion: A Case Study in RUL Estimation of Turbofan Engine | |
| Li | Machine Learning Based Seismic Structural Health Monitoring and Reconnaissance | |
| Huang et al. | Remaining useful life prediction of turbofan engine with GA optimized hybrid neural network | |
| CN116405369A (en) | 5G carrier network fault diagnosis method and equipment based on graph convolution neural network |
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 |