CN117077327A - Bearing life prediction method and system based on digital twin - Google Patents
Bearing life prediction method and system based on digital twin Download PDFInfo
- Publication number
- CN117077327A CN117077327A CN202311344281.8A CN202311344281A CN117077327A CN 117077327 A CN117077327 A CN 117077327A CN 202311344281 A CN202311344281 A CN 202311344281A CN 117077327 A CN117077327 A CN 117077327A
- Authority
- CN
- China
- Prior art keywords
- bearing
- parameters
- digital twin
- state estimation
- model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000036541 health Effects 0.000 claims abstract description 78
- 230000015556 catabolic process Effects 0.000 claims abstract description 30
- 238000006731 degradation reaction Methods 0.000 claims abstract description 30
- 238000012544 monitoring process Methods 0.000 claims abstract description 28
- 238000003062 neural network model Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000004088 simulation Methods 0.000 claims abstract description 15
- 238000004364 calculation method Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 42
- 230000035882 stress Effects 0.000 claims description 12
- 230000032683 aging Effects 0.000 claims description 9
- 230000015654 memory Effects 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000013499 data model Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 2
- 238000013398 bayesian method Methods 0.000 claims description 2
- 230000006399 behavior Effects 0.000 claims description 2
- 239000000463 material Substances 0.000 claims description 2
- 210000004027 cell Anatomy 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 6
- 239000013598 vector Substances 0.000 description 5
- 238000012423 maintenance Methods 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003412 degenerative effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computer Hardware Design (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Medical Informatics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a bearing life prediction method and system based on digital twinning, and relates to the technical field of bearing life prediction. The method comprises the steps of establishing a bearing health state estimation function; establishing a digital twin model of the bearing to perform simulation calculation to obtain simulated bearing state parameters; respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result, updating related parameters in a digital twin model of the bearing in real time to obtain an updated digital twin model so as to obtain bearing life-span data, and extracting state monitoring data in a degradation stage; and training the LSTM neural network model by using the state monitoring data of the degradation stage. According to the invention, the full life data under multiple working conditions is obtained through the digital twin model, so that a more accurate prediction model is obtained, and a more accurate life prediction result is obtained.
Description
Technical Field
The invention belongs to the technical field of bearing life prediction, and particularly relates to a bearing life prediction method and system based on digital twinning.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Bearings are a common mechanical component, and are commonly used in devices such as transformers, and because bearings are more prone to failure than other components, research on bearings has been increasingly conducted in recent years. The maximum service life of the bearing has a large variability due to the complexity and uncertainty of the service environment and working conditions of the rolling bearing, the randomness of fatigue damage development and the diversity of failure modes, and the regular maintenance of the bearing often causes problems of 'insufficient maintenance' and 'excessive maintenance'. However, the prediction of the RUL (Remaining Useful Life, remaining service life) of the bearing can prolong the service life of the bearing to the maximum extent and reduce the maintenance cost, so the research on the prediction of the RUL of the bearing has been the focus of the research in the field.
There are a number of model-based prediction methods that describe the degradation trend of a system or device by building a physical or mathematical model, which requires a great deal of physical mechanism or empirical knowledge. The inventor finds that when the method is applied to bearing life prediction, the following technical problems often exist:
(1) Because the existing bearing cannot be replaced until the bearing is completely scrapped in use, the whole life history data of the bearing is difficult to acquire when the running data of the bearing is acquired.
(2) The model input sample size is too large, and degradation information that the same feature can express in different degradation phases, the help of subsequent predictions, is different, which is often not considered.
(3) At present, when a model is used for predicting service life, the historical data of a bearing under a certain working condition is generally considered, and the generalization of the model obtained by training is low.
Along with the development of digital twin technology, the use of a digital twin model to obtain the life-span data of a bearing, and further, the prediction of the life span of the bearing has become a trend. However, how to guarantee the accuracy of the digital twin model becomes a key factor that directly affects life prediction.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a method and a system for predicting the service life of a bearing based on digital twin, a digital twin model of the bearing is established, the full service life data of the bearing is generated through the digital twin model, meanwhile, state monitoring data in a degradation stage are accurately extracted, the service life of the bearing is predicted by utilizing an LSTM model based on the obtained state monitoring data, and finally a more accurate service life prediction result is obtained.
To achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
the first aspect of the invention provides a method for predicting bearing life based on digital twinning.
The bearing life prediction method based on digital twinning comprises the following steps:
establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
A second aspect of the invention provides a digital twinning based bearing life prediction system.
A digital twinning-based bearing life prediction system comprising:
a bearing health estimation function determination module configured to: establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
a digital twin model simulation module configured to: collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
an estimation result acquisition module configured to: respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
a contrast update module configured to: comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
a degradation phase state monitoring data acquisition module configured to: obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
a training prediction module configured to: and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
A third aspect of the invention provides a computer readable storage medium having stored thereon a program which when executed by a processor performs the steps in a method of predicting bearing life based on digital twinning according to the first aspect of the invention.
A fourth aspect of the invention provides an electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, the processor implementing the steps in the digital twin based bearing life prediction method according to the first aspect of the invention when the program is executed.
The one or more of the above technical solutions have the following beneficial effects:
the invention provides a bearing life prediction method and a system based on digital twin, which can obtain full life data under multiple working conditions through a digital twin model, train an LSTM neural network model by using state monitoring data of a degradation stage in the full life data, and further obtain a more accurate prediction model and a more accurate life prediction result.
The method comprises the steps of establishing a bearing health state estimation function, collecting bearing state parameters, determining the parameters of the bearing health state estimation function by adopting a fitting correlation method, respectively inputting the bearing state parameters collected in real time and the simulated bearing state parameters into the bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result, comparing the first health state estimation result with the second health state estimation result, and updating the related parameters in the bearing digital twin model in real time based on the comparison result, so that an updated more accurate digital twin model is obtained, and further obtaining simulation data by utilizing the updated digital twin model.
According to the invention, the digital twin model is utilized to obtain simulation data of multiple working conditions, so that the LSTM neural network model obtained through training is suitable for a cross-working condition situation, and the generalization performance of the model is improved.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
Fig. 1 is a flow chart of a method of a first embodiment.
Fig. 2 is a system configuration diagram of a second embodiment.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
The embodiment discloses a bearing life prediction method based on digital twinning.
As shown in fig. 1, the method for predicting the life of the bearing based on digital twin comprises the following steps:
establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
Further, a bearing health state estimation function is established, bearing state parameters are collected, and the parameters of the bearing health state estimation function are determined by adopting a fitting association method, specifically:
calculating the aging rate of the bearing based on the state parameters of the bearing;
estimating the actual service life of the bearing according to the aging rate and the expected service life of the bearing;
and carrying out sectional fitting on the bearing health state estimation function by adopting a least square fitting method according to the state parameters of the bearing, obtaining the proportional coefficient and the curvature coefficient of the parameters in each stage, and obtaining the bearing health state estimation function with the parameters determined.
The method comprises the following specific processes of collecting state parameters of the bearing and calculating the electrical aging rate:
obtaining real-time running state parameters of the bearing, calculating mechanical properties of the bearing according to the consultable information and the historical running data, obtaining corresponding performance indexes, and calculating the current aging rate:
wherein Q is the aging rate of the current transformer;the difference value between the performance index at the current moment and the initial performance index; />The difference value between the current time and the initial time is the difference value; i is a performance number, i=1, 2,3 … … n, n is the total number of performance categories; η is a coefficient of performance and is obtained according to the degree of influence of different performances on the aging of the bearing.
The actual service life of the bearing can be estimated according to the aging rate and the expected service life of the bearing, the health state estimation function is subjected to piecewise fitting by adopting a least square fitting method according to the acquired state parameters of the bearing, and the proportionality coefficient and the curvature coefficient of the parameters in each stage are obtained, so that a first health state estimation result and a second health state estimation result are obtained through calculation.
The health status estimation function is:
wherein S is the actual working age,is a curvature coefficient->And H is the health state estimation result and is the proportionality coefficient.
Further, the operation data of the bearing specifically includes: vibration, temperature, pressure and rotational speed; the parameters of the bearing specifically comprise: material physical parameters, structural parameters and operation condition parameters of the bearing; carrying out dynamic stress analysis on the bearing in operation and recording the stress state;
according to the obtained operation data, parameters and stress states, a physical model and a data model of the bearing are established;
and fusing the physical model and the data model, establishing a digital twin model of the bearing, and predicting and analyzing the behavior of the bearing.
A large amount of simulation data is generated through the digital twin model, so that the problem that data is short in the field of residual life prediction in neural network learning is solved, and meanwhile, the accuracy of a prediction model is improved.
In order to obtain more accurate simulation data, the first health state estimation result and the second health state estimation result are compared, and relevant parameters in the digital twin model of the bearing are updated by using a Bayesian method, so that the more accurate digital twin model is obtained.
Generating bearing life data under different working conditions through a digital twin model, and carrying out noise reduction and normalization processing on the generated bearing life data; bearing life data is divided into two phases: a healthy phase and a degenerative phase. The state monitoring data generally show relatively weak degradation characteristics in the early degradation stage, and since the state monitoring data in the early degradation stage can cause fluctuation and error of the bearing life prediction, the state monitoring data is difficult to be effectively utilized by the bearing life prediction model, and therefore, the root mean square index is adopted as a boundary line for distinguishing the healthy stage from the degradation stage, and when the root mean square index satisfies the following conditions, the obtained data is the state monitoring data for obtaining the degradation stage:
wherein,representing a root mean square index; />Representing obtaining a first parameter from vibration data of an initial health stage; />Representing obtaining a second parameter from the vibration data of the initial health phase; />Represented in [3,5 ]]Random weights in the interval.
Further, training the LSTM neural network model comprises the following specific processes:
constructing an LSTM neural network model comprising an LSTM layer, a feature awareness layer, a life prediction LSTM layer and a regression layer;
determining the number of hidden neurons of the LSTM layer, and predicting the number and the layer number of the hidden neurons of the LSTM layer; taking the characteristics of vibration signals in the state monitoring data of the degradation stage as the input of an LSTM neural network model, and outputting a residual service life percentage predicted value corresponding to each moment;
taking the percentage of the actual residual service life in the life cycle as a corresponding label, and performing iterative training on the LSTM network model by using an Adam optimizer until the loss function reaches a preset value to obtain a trained LSTM neural network model;
and inputting the sample to be predicted into the trained LSTM network model, and outputting a prediction result of the residual service life of the bearing.
Specific:
step M1: and obtaining vibration signals in the bearing degradation stage state monitoring data obtained by digital twin model simulation.
Step M2: n features are extracted from the vibration signal, including time domain features, frequency domain features and the like, and a feature set is constructed.
Step M3: carrying out normalization processing on the features in the feature set to prevent the influence of the amplitude difference among different features on the result; and converting the actual residual service life corresponding to each moment into a service life percentage, and outputting the service life percentage as the output of the LSTM network model.
The training sample collected in the step M1 is a vibration signal of the whole life cycle of the bearing, and the actual residual service life corresponding to the sample is obtained by subtracting the moment corresponding to the current sample from the bearing service life end moment.
Step M4: an LSTM network model with N input characteristic channels and characteristic awareness is constructed, wherein the LSTM network model with the characteristic awareness comprises an LSTM layer, a characteristic awareness layer, a life prediction LSTM layer and a regression layer. Determining the number of hidden neurons of the LSTM layer, and predicting the number and the layer number of the hidden neurons of the LSTM layer; presetting a time window with the length of L, taking normalized features of L periods before any moment as LSTM network model input with feature consciousness, and outputting an actual residual service life percentage predicted value corresponding to each moment.
Step M5: and taking the percentage of the actual residual service life in the life cycle as a corresponding label, and performing iterative training on the LSTM network model with the feature consciousness by using an Adam optimizer until the loss function reaches a preset value to obtain the trained LSTM network model with the feature consciousness.
Step M6: and inputting the sample to be predicted into the LSTM network model with the feature consciousness obtained after training, and outputting the prediction result of the residual service life of the bearing.
The vibration signal in the step M1 includes two channels in the X direction and the Y direction.
The step M3 includes: and carrying out normalization processing on the features in the feature set by adopting a z-score normalization method, wherein the expression is as follows:
wherein,for the j-th feature->And->The mean and standard deviation in the first L periods of the jth feature,is the j-th feature after normalization.
Specifically, the step M4 includes:
step M4.1: the feature awareness module is provided with N channels, each channel has an identical and independent structure, the normalized features of L periods before any moment are firstly input into an LSTM (long short-term memory neural network) layer for processing the feature time sequence information, and the output at the current moment is obtainedThe specific formula is as follows:
wherein,、/>、/>the values of a forget gate, an input gate and an output gate at the current moment are respectively; />And->The cell states at the current time and the previous time respectively; />And->The output of the current moment and the previous moment are respectively; />Is the input of the current moment; />、/>、/>、/>The weight matrixes are respectively a forgetting gate, an input gate, an output gate and a cell state; />、/>、/>、The input gate, the output gate and the cell state are respectively a forgetting gate, an input gate, an output gate and a cell state paranoid matrix; />And->Respectively indicate->And a hyperbolic tangent activation function; />Representing the operation of multiplication of the corresponding elements.
Step 4.2: the normalized features of the first L periods and the output ht of the LSTM layer are input into a feature consciousness layer at the same time, and a characterization vector is obtained through operation, wherein the expression is as follows:
wherein,output denoted as the final moment of the j-th channel LSTM layer,/and>for the j-th feature->And->Respectively a weight matrix and a bias matrix.
Step M4.3: the characterization vectors obtained by the channels are synthesized, and the weight of each feature is obtained through operation, and the expression is as follows:
wherein,is a parameter vector to be learned; />Representation vector->Is a transpose of (2); n represents the number of features mentioned; />A token vector representing a jth lane; />Representing the token vector for the kth channel.
Step M4.4: multiplying the jth feature by the obtained weight of the jth feature, the expression is as follows:
step M4.5: features after weightingComposing multidimensional features->In the input life prediction LSTM layer, the output of the current moment is calculated>The expression is as follows:
wherein,、/>、/>the values of a forget gate, an input gate and an output gate at the current moment are respectively; />And->The cell states at the current time and the previous time respectively; />And->The output of the current moment and the previous moment are respectively; />Is the input of the current moment; />、/>、/>、/>The weight matrixes are respectively a forgetting gate, an input gate, an output gate and a cell state; />、/>、/>、/>The input gate, the output gate and the cell state are respectively a forgetting gate, an input gate, an output gate and a cell state paranoid matrix; />And->Respectively indicate->And a hyperbolic tangent activation function; />Representing the operation of multiplication of the corresponding elements.
Step 4.6: will outputThe predicted remaining life percentage is obtained by a regression layer expressed as follows:
in the method, in the process of the invention,and->The weight matrix and the bias matrix of the regression layer are represented, respectively.
The loss function in the step M5 is a mean square error function, and the expression is as follows:
wherein,representing the number of samples->And->Respectively represent +.>The remaining life of each sample is a true value and a predicted value.
Example two
The embodiment discloses a bearing life prediction system based on digital twinning.
As shown in fig. 2, a digital twin based bearing life prediction system comprising:
a bearing health estimation function determination module configured to: establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
a digital twin model simulation module configured to: collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
an estimation result acquisition module configured to: respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
a contrast update module configured to: comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
a degradation phase state monitoring data acquisition module configured to: obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
a training prediction module configured to: and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
Example III
An object of the present embodiment is to provide a computer-readable storage medium.
A computer readable storage medium having stored thereon a computer program which when executed by a processor implements the steps in a digital twin based bearing life prediction method as described in embodiment 1 of the present disclosure.
Example IV
An object of the present embodiment is to provide an electronic apparatus.
An electronic device comprising a memory, a processor, and a program stored on the memory and executable on the processor, the processor implementing the steps in the digital twinning-based bearing life prediction method as described in embodiment 1 of the present disclosure when the program is executed.
The steps involved in the devices of the second, third and fourth embodiments correspond to those of the first embodiment of the method, and the detailed description of the embodiments can be found in the related description section of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (10)
1. The bearing life prediction method based on digital twinning is characterized by comprising the following steps of:
establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
2. The method for predicting the life of the bearing based on digital twin according to claim 1, wherein a bearing health state estimation function is established, bearing state parameters are collected, and the parameters of the bearing health state estimation function are determined by adopting a fitting association method, specifically:
calculating the aging rate of the bearing based on the state parameters of the bearing;
estimating the actual service life of the bearing according to the aging rate and the expected service life of the bearing;
and carrying out sectional fitting on the bearing health state estimation function by adopting a least square fitting method according to the state parameters of the bearing, obtaining the proportional coefficient and the curvature coefficient of the parameters in each stage, and obtaining the bearing health state estimation function with the parameters determined.
3. The method for predicting bearing life based on digital twinning of claim 2, wherein the bearing health estimation function is:
wherein S is the actual working age,is a curvature coefficient->And H is the health state estimation result and is the proportionality coefficient.
4. The digital twin based bearing life prediction method of claim 1, wherein:
the operation data of the bearing specifically comprises: vibration, temperature, pressure and rotational speed; the parameters of the bearing specifically comprise: material physical parameters, structural parameters and operation condition parameters of the bearing; carrying out dynamic stress analysis on the bearing in operation and recording the stress state;
according to the obtained operation data, parameters and stress states, a physical model and a data model of the bearing are established;
and fusing the physical model and the data model, establishing a digital twin model of the bearing, and predicting and analyzing the behavior of the bearing.
5. The digital twin based bearing life prediction method of claim 1, wherein the first health state estimation result and the second health state estimation result are compared, and related parameters in the digital twin model of the bearing are updated using a bayesian method.
6. The method for predicting bearing life based on digital twinning as set forth in claim 1, wherein a root mean square index is used as a boundary line for distinguishing a healthy phase from a degraded phase, and the obtained data is state monitoring data for obtaining the degraded phase when the root mean square index satisfies the following conditions:
wherein,representing a root mean square index; />Representing obtaining a first parameter from vibration data of an initial health stage; />Representing obtaining a second parameter from the vibration data of the initial health phase; />Represented in [3,5 ]]Random weights in the interval.
7. The method for predicting bearing life based on digital twinning according to claim 1, wherein the training of the LSTM neural network model comprises the following specific steps:
constructing an LSTM neural network model comprising an LSTM layer, a feature awareness layer, a life prediction LSTM layer and a regression layer;
determining the number of hidden neurons of the LSTM layer, and predicting the number and the layer number of the hidden neurons of the LSTM layer; taking the characteristics of vibration signals in the state monitoring data of the degradation stage as the input of an LSTM neural network model, and outputting a residual service life percentage predicted value corresponding to each moment;
taking the percentage of the actual residual service life in the life cycle as a corresponding label, and performing iterative training on the LSTM network model by using an Adam optimizer until the loss function reaches a preset value to obtain a trained LSTM neural network model;
and inputting the sample to be predicted into the trained LSTM network model, and outputting a prediction result of the residual service life of the bearing.
8. Bearing life prediction system based on digital twin, its characterized in that: comprising the following steps:
a bearing health estimation function determination module configured to: establishing a bearing health state estimation function, collecting bearing state parameters, and determining the parameters of the bearing health state estimation function by adopting a fitting association method;
a digital twin model simulation module configured to: collecting operation data, parameters and stress states of a bearing, establishing a digital twin model of the bearing, and performing simulation calculation by using the digital twin model to obtain simulated bearing state parameters;
an estimation result acquisition module configured to: respectively inputting the real-time collected bearing state parameters and the simulated bearing state parameters into a bearing health state estimation function to obtain a first health state estimation result and a second health state estimation result;
a contrast update module configured to: comparing the first health state estimation result with the second health state estimation result, and updating related parameters in the digital twin model of the bearing in real time based on the comparison result to obtain an updated digital twin model;
a degradation phase state monitoring data acquisition module configured to: obtaining bearing life data through the updated digital twin model, and extracting state monitoring data of a degradation stage;
a training prediction module configured to: and training the LSTM neural network model by taking the state monitoring data in the degradation stage as a sample set, and predicting the residual life of the bearing by using the trained LSTM neural network model.
9. A computer readable storage medium having stored thereon a program, characterized in that the program when executed by a processor realizes the steps in the digital twinning based bearing life prediction method according to any one of claims 1-7.
10. An electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, wherein the processor implements the steps of the digital twinning-based bearing life prediction method of any one of claims 1-7 when the program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311344281.8A CN117077327A (en) | 2023-10-18 | 2023-10-18 | Bearing life prediction method and system based on digital twin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311344281.8A CN117077327A (en) | 2023-10-18 | 2023-10-18 | Bearing life prediction method and system based on digital twin |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117077327A true CN117077327A (en) | 2023-11-17 |
Family
ID=88715704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311344281.8A Pending CN117077327A (en) | 2023-10-18 | 2023-10-18 | Bearing life prediction method and system based on digital twin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117077327A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117407771A (en) * | 2023-12-13 | 2024-01-16 | 湘江实验室 | Bearing health state assessment method and device based on digital twin and related equipment |
| CN117782647A (en) * | 2024-02-23 | 2024-03-29 | 杭州中港地铁装备维护有限公司 | Subway running part bearing health state monitoring system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532626A (en) * | 2019-07-31 | 2019-12-03 | 西安交通大学 | Based on the twin aeroplane engine main bearing method for predicting residual useful life of number |
| CN112129534A (en) * | 2020-09-17 | 2020-12-25 | 上海交通大学 | Method and system for predicting service life of main shaft bearing of machine tool |
| CN113092115A (en) * | 2021-04-09 | 2021-07-09 | 重庆大学 | Digital twin model construction method of digital-analog combined drive full-life rolling bearing |
| CN113221277A (en) * | 2021-05-14 | 2021-08-06 | 西安交通大学 | Bearing performance degradation evaluation method and system based on digital twin model |
| CN114626625A (en) * | 2022-03-25 | 2022-06-14 | 华南理工大学 | Bearing life prediction method based on multi-module U-BiLSTM network |
| CN115796059A (en) * | 2023-02-07 | 2023-03-14 | 中国电建集团山东电力建设第一工程有限公司 | Electrical equipment service life prediction method and system based on deep learning |
| CN115901263A (en) * | 2022-12-14 | 2023-04-04 | 昆明理工大学 | Rolling bearing residual life online prediction method and system based on digital twinning |
| CN116561927A (en) * | 2023-05-29 | 2023-08-08 | 西安交通大学 | Digital twin-driven small sample rotary machine residual life prediction method and system |
| CN116662861A (en) * | 2023-06-01 | 2023-08-29 | 燕山大学 | Multi-domain graph migration bearing life prediction method based on digital twin model driving |
| CN116842379A (en) * | 2023-06-06 | 2023-10-03 | 山东省计算中心（国家超级计算济南中心） | Mechanical bearing residual service life prediction method based on DRSN-CS and BiGRU+MLP models |
-
2023
- 2023-10-18 CN CN202311344281.8A patent/CN117077327A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532626A (en) * | 2019-07-31 | 2019-12-03 | 西安交通大学 | Based on the twin aeroplane engine main bearing method for predicting residual useful life of number |
| CN112129534A (en) * | 2020-09-17 | 2020-12-25 | 上海交通大学 | Method and system for predicting service life of main shaft bearing of machine tool |
| CN113092115A (en) * | 2021-04-09 | 2021-07-09 | 重庆大学 | Digital twin model construction method of digital-analog combined drive full-life rolling bearing |
| CN113221277A (en) * | 2021-05-14 | 2021-08-06 | 西安交通大学 | Bearing performance degradation evaluation method and system based on digital twin model |
| CN114626625A (en) * | 2022-03-25 | 2022-06-14 | 华南理工大学 | Bearing life prediction method based on multi-module U-BiLSTM network |
| CN115901263A (en) * | 2022-12-14 | 2023-04-04 | 昆明理工大学 | Rolling bearing residual life online prediction method and system based on digital twinning |
| CN115796059A (en) * | 2023-02-07 | 2023-03-14 | 中国电建集团山东电力建设第一工程有限公司 | Electrical equipment service life prediction method and system based on deep learning |
| CN116561927A (en) * | 2023-05-29 | 2023-08-08 | 西安交通大学 | Digital twin-driven small sample rotary machine residual life prediction method and system |
| CN116662861A (en) * | 2023-06-01 | 2023-08-29 | 燕山大学 | Multi-domain graph migration bearing life prediction method based on digital twin model driving |
| CN116842379A (en) * | 2023-06-06 | 2023-10-03 | 山东省计算中心（国家超级计算济南中心） | Mechanical bearing residual service life prediction method based on DRSN-CS and BiGRU+MLP models |
Non-Patent Citations (1)
| Title |
|---|
| 丁华;杨亮亮;杨兆建;王义亮;: "数字孪生与深度学习融合驱动的采煤机健康状态预测", 中国机械工程, no. 07 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117407771A (en) * | 2023-12-13 | 2024-01-16 | 湘江实验室 | Bearing health state assessment method and device based on digital twin and related equipment |
| CN117782647A (en) * | 2024-02-23 | 2024-03-29 | 杭州中港地铁装备维护有限公司 | Subway running part bearing health state monitoring system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zheng et al. | Long short-term memory network for remaining useful life estimation | |
| CN110705692B (en) | Nonlinear dynamic industrial process product prediction method of space-time attention network | |
| CN111539515B (en) | Complex equipment maintenance decision method based on fault prediction | |
| CN117077327A (en) | Bearing life prediction method and system based on digital twin | |
| Fan et al. | A novel machine learning method based approach for Li-ion battery prognostic and health management | |
| CN116757534B (en) | Intelligent refrigerator reliability analysis method based on neural training network | |
| CN112129534B (en) | Method and system for predicting service life of main shaft bearing of machine tool | |
| CN111160626B (en) | Power load time sequence control method based on decomposition fusion | |
| CN111680786B (en) | Time sequence prediction method based on improved weight gating unit | |
| CN114218872B (en) | DBN-LSTM semi-supervised joint model-based residual service life prediction method | |
| Li et al. | Domain adaptation remaining useful life prediction method based on AdaBN-DCNN | |
| CN115225516B (en) | LSSVM network flow prediction method based on improved ABC-VMD | |
| CN112784920A (en) | Cloud-side-end-coordinated dual-anti-domain self-adaptive fault diagnosis method for rotating part | |
| CN110757510A (en) | Method and system for predicting remaining life of robot | |
| Yang et al. | Remaining useful life prediction based on normalizing flow embedded sequence-to-sequence learning | |
| CN114266201B (en) | Self-attention elevator trapping prediction method based on deep learning | |
| Sadoughi et al. | A deep learning approach for failure prognostics of rolling element bearings | |
| CN114578234A (en) | Lithium ion battery degradation and capacity prediction model considering causality characteristics | |
| CN117010442A (en) | Equipment residual life prediction model training method, residual life prediction method and system | |
| CN115456287A (en) | Long-and-short-term memory network-based multi-element load prediction method for comprehensive energy system | |
| CN113469013B (en) | Motor fault prediction method and system based on transfer learning and time sequence | |
| CN113468720B (en) | Service life prediction method for digital-analog linked random degradation equipment | |
| Pei et al. | Multisensor prognostic of RUL based on EMD-ESN | |
| CN114372618A (en) | Student score prediction method and system, computer equipment and storage medium | |
| CN113435321A (en) | Method, system and equipment for evaluating state of main shaft bearing and readable storage medium |
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 |