CN116090353A - Product remaining life prediction method and device, electronic equipment and storage medium - Google Patents
Product remaining life prediction method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN116090353A CN116090353A CN202310160024.2A CN202310160024A CN116090353A CN 116090353 A CN116090353 A CN 116090353A CN 202310160024 A CN202310160024 A CN 202310160024A CN 116090353 A CN116090353 A CN 116090353A
- Authority
- CN
- China
- Prior art keywords
- degradation
- time
- product
- particle
- parameters
- 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 100
- 238000006731 degradation reaction Methods 0.000 claims abstract description 162
- 230000015556 catabolic process Effects 0.000 claims abstract description 154
- 239000002245 particle Substances 0.000 claims abstract description 130
- 238000013528 artificial neural network Methods 0.000 claims abstract description 75
- 238000012549 training Methods 0.000 claims abstract description 43
- 230000005653 Brownian motion process Effects 0.000 claims abstract description 36
- 230000007787 long-term memory Effects 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 18
- 238000012360 testing method Methods 0.000 claims abstract description 18
- 230000007704 transition Effects 0.000 claims abstract description 18
- 238000005259 measurement Methods 0.000 claims abstract description 16
- 238000005070 sampling Methods 0.000 claims abstract description 15
- 238000012546 transfer Methods 0.000 claims abstract description 15
- 238000012795 verification Methods 0.000 claims abstract description 15
- 230000007774 longterm Effects 0.000 claims abstract description 12
- 239000000047 product Substances 0.000 claims description 100
- 238000001914 filtration Methods 0.000 claims description 39
- 230000015654 memory Effects 0.000 claims description 28
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012952 Resampling Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000005537 brownian motion Methods 0.000 claims description 3
- 239000007857 degradation product Substances 0.000 claims description 3
- 238000009792 diffusion process Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 description 11
- 238000011161 development Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 238000010200 validation analysis Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 229910052744 lithium Inorganic materials 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000007723 transport mechanism Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- 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
-
- 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
- G06F2119/04—Ageing analysis or optimisation against ageing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The embodiment of the application provides a method and a device for predicting the residual life of a product, electronic equipment and a storage medium. The method comprises the following steps: acquiring degradation data of a product, and extracting a degradation curve according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment.
Description
Technical Field
The present disclosure relates to the field of fault prediction and health management technologies, and in particular, to a method and apparatus for predicting a remaining life of a product, an electronic device, and a storage medium.
Background
With the rapid development of modern industry and information age, the new generation of industrial revolution has been worldwide and the industrial modernization speed is gradually increased. The manner in which they fail, whether products or equipment, becomes increasingly complex as the demand for them increases. For some potential failure modes, it is difficult to distinguish their failure by sense, and in the overhaul of expensive precision equipment, simple detection often leads to performance degradation. Based on this, in recent years, fault Prediction and Health Management (PHM) has become increasingly widespread both academic and industrial applications. The PHM method allows for evaluation of the reliability and residual life prediction of the system under its actual life cycle conditions to predict when and where a fault will occur, thereby eliminating the risk of the system.
Although neural networks have seen more remarkable results in predicting the remaining life in recent years, neural networks still fail to give a probability density of failure for prediction, and the variable phase results in the loss of some decision methods for the users of the product. The neural network method gradually becomes the mainstream in the prediction of the residual life at present, but because the neural network method is a black box, the prior study cannot explain the real principle of the neural network, and the probability density of product failure cannot be obtained by the neural network method, so that the prediction precision is low.
Disclosure of Invention
The main purpose of the embodiments of the present application is to provide a method and apparatus for predicting the remaining lifetime of a product, an electronic device, and a storage medium, which can combine the prediction capability of a neural network with particle filtering, not only give out the probability density of failure, but also have higher prediction accuracy compared with the common particle filtering.
To achieve the above object, a first aspect of an embodiment of the present application proposes a method for predicting remaining life of a product, the method comprising:
acquiring degradation data of a product, and extracting a degradation curve according to the degradation data;
performing degradation modeling on the degradation curve according to a wiener process to obtain a degradation model, and determining offline parameters according to the degradation model;
dividing the degradation curve into a training set, a testing set and a verification set, and training a long-period and short-period neural network;
establishing a state transition equation and a measurement equation according to a wiener process, and sampling from the offline parameters;
according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time;
normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment;
and updating parameters according to the real state data acquired at each moment.
In some embodiments, the degradation model is:
the value of the fixed parameter θ and the distribution of the random parameters λ and σ are obtained by a two-stage method.
In some embodiments, the establishing state transition equations and measurement equations according to the wiener process includes:
will utilize the observation sequence
The parameter estimation problem is regarded as a recursive Bayesian filtering problem, and posterior distribution is estimated
wherein ,
for cut-off t k Time of day on-line degenerate state sequence,/->
At t k Model parameters estimated by time zones; updating the offline parameters by using the observation sequence, and constructing a state space model as follows
wherein ,
in some embodiments, the performing the state transition equation recursion to the n time according to the particle parameter at the current time and the system state at the n time after the prediction of the long-term memory neural network includes:
the predicted degradation sequence according to the neural network is defined as
Keep->
The parameters are unchanged and then based on the formula +.>
Can be recursively t k+n At the moment, the calculation of the weight is performed at each step of the recursion.
In some embodiments, the integrating the particle weights in the recursion process is the particle weight of the current time, including:
t is recorded k The weighted particle set of the parameters of the time is
M is the number of particles, and the number of particles,
wherein ,
at t k-1 Time-of-day predicted particle state and t k Weight assigned to likelihood between real states of the time system, in order +.>
As a method for converting the real state of the system into the state predicted by the long-short term neural network, N is the number of steps predicted by the neural network.
In some embodiments, the normalizing according to the particle weight to obtain the remaining lifetime probability density of the product at the current time includes:
according to the principle of particle filtering, at t k The time can be obtained
wherein ,
for m=1, 2l, m is normalized according to the weights, resulting in
Thus at t k The calculation formula of the residual life probability density of the time prediction degradation product is as follows:
in some embodiments, after said normalizing according to particle weights, comprising:
copying the particles with larger weight values, discarding the particles with smaller weight values, and introducing effective sampling size, which is defined as:
wherein ,Neff Smaller means more serious particle degradation phenomenon, and the resampling threshold is set to be N th When N eff <N th At that time, resampling is performed.
To achieve the above object, a second aspect of the embodiments of the present application proposes a product remaining life prediction apparatus, the apparatus comprising:
the acquisition module is used for acquiring degradation data of the product and extracting a degradation curve according to the degradation data;
the modeling module is used for carrying out degradation modeling on the degradation curve according to a wiener process to obtain a degradation model, and determining offline parameters according to the degradation model;
the training module is used for dividing the degradation curve into a training set, a testing set and a verification set and training a long-term and short-term neural network;
the establishing module is used for establishing a state transition equation and a measurement equation according to the wiener process and sampling from the offline parameters;
the recursion module is used for recursing a state transfer equation to the n time according to the system state of the n time after the long-term memory neural network is predicted and the particle weight in the recursion process is comprehensively the particle weight of the current time;
the prediction module is used for normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment;
and the updating module is used for updating parameters according to the real state data acquired at each moment.
To achieve the above object, a third aspect of the embodiments of the present application proposes an electronic device, which includes a memory and a processor, the memory storing a computer program, the processor implementing the method according to the first aspect when executing the computer program.
To achieve the above object, a fourth aspect of the embodiments of the present application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of the first aspect.
The method and the device for predicting the residual life of the product, the electronic equipment and the storage medium acquire degradation data of the product, and extract a degradation curve according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the application provides a product residual life prediction method integrating a long-term memory neural network and particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product. According to the embodiment of the application, through the combination of the prediction capability of the neural network and the particle filtering, the probability density of failure can be given, and the prediction accuracy is higher compared with that of the common particle filtering.
Drawings
FIG. 1 is a flow chart of a method for predicting remaining life of a product provided by an embodiment of the present application;
FIG. 2 is a signature graph of a NASA turbofan engine;
FIG. 3 is a probability density curve for the remaining life of engine number 258;
FIG. 4 is a probability density curve for the remaining life of engine number 244;
FIG. 5 is a probability density curve for the remaining life of engine number 208;
FIG. 6 is a probability density curve for the remaining life of engine number 256;
FIG. 7 is a schematic structural view of a product remaining life prediction apparatus provided in an embodiment of the present application;
fig. 8 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
It should be noted that although functional block division is performed in a device diagram and a logic sequence is shown in a flowchart, in some cases, the steps shown or described may be performed in a different order than the block division in the device, or in the flowchart. The terms first, second and the like in the description and in the claims and in the above-described figures, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Aiming at the technical problems that a residual life density curve of equipment cannot be given and the prediction precision is low in the prior art, the embodiment of the application provides a method and a device for predicting the residual life of a product, electronic equipment and a storage medium, degradation data of the product are obtained, and the degradation curve is extracted according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the application provides a product residual life prediction method integrating a long-term memory neural network and particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product. According to the embodiment of the application, through the combination of the prediction capability of the neural network and the particle filtering, the probability density of failure can be given, and the prediction accuracy is higher compared with that of the common particle filtering.
The method and device for predicting the remaining life of a product, the electronic device and the storage medium provided by the embodiment of the application are specifically described through the following embodiments, and the method for predicting the remaining life of the product in the embodiment of the application is described first.
Fig. 1 is an optional flowchart of a method for predicting remaining life of a product according to an embodiment of the present application, where the method in fig. 1 may include, but is not limited to, steps S101 to S107.
Step S101, acquiring degradation data of a product, and extracting a degradation curve according to the degradation data;
step S102, carrying out degradation modeling on a degradation curve according to a wiener process to obtain a degradation model, and determining offline parameters according to the degradation model;
step S103, dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network;
step S104, a state transition equation and a measurement equation are established according to the wiener process, and sampling is carried out from offline parameters;
step S105, according to the system state of the n time after the long-term memory neural network prediction, performing state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time;
step S106, normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment;
step S107, parameter updating is carried out according to the real state data acquired at each moment.
In some embodiments, degradation data is obtained for a product, where the product may include, but is not limited to, a bearing, a lithium battery, an engine, and the like. Extracting a degradation curve with trend according to the degradation data characteristics of the product. And carrying out degradation modeling on the degradation curve according to the wiener process. It should be noted that the present application is based on the wiener process and is therefore more versatile than the prior art.
In some embodiments, the degradation curve for the product is divided into a training set, a test set, and a validation set. Wherein samples of the training set are used to determine failure thresholds for the product and training of the neural network. The test set is used to verify the training effect of the neural network and the validation set is used to verify the validity of the proposed method.
In some embodiments, the values of the fixed parameter θ, and the distribution of the random parameters λ and σ are found by a two-stage method from samples of the training set. If for products where security monitoring is not high, reliability assessment can be made with a definition based on time of arrival. If there is a high demand for security monitoring of individual products, online updating of parameters is required.
In some embodiments, degradation data of a product is obtained, and a degradation curve is extracted according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the prediction capability of the neural network is combined with the particle filtering, so that the probability density of failure can be given, and the prediction accuracy is higher than that of the common particle filtering. The probability density curve of the residual life can be obtained by the method, and a wider product maintenance strategy can be brought to a product user.
In some embodiments, the method is a product remaining life prediction method that fuses long-term memory neural networks with particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product.
The method for predicting the residual life of the product by fusing the long-term and short-term memory neural network and the particle filtering is further described by combining a specific embodiment, and specifically comprises the following steps:
1. and acquiring degradation data of the product, and extracting a degradation curve with trend according to the degradation data characteristics of the product.
2. And carrying out degradation modeling on the degradation curve according to the wiener process to obtain a degradation model as follows:
wherein ,X(tk ) At t k The degradation state of the equipment at the moment, k is the observation times; λ is a drift parameter used to characterize the degradation rate; μ (τ; θ) represents a time-varying nonlinear function of the parameter θ for describing the nonlinearity of the device degradation state; b (t) k ) For standard brownian motion, σ is the diffusion parameter, used to describe time-varying random fluctuations and the extent of the fluctuations, respectively, in the degradation process. Different functional forms μ (τ; θ) can describe different forms of degradation processes. For example when
Time degradation is a linear degradation process, ">
And can be used to represent a power function degradation process.
3. The degradation curve of the product is divided into a training set, a testing set and a verification set. Wherein samples of the training set are used to determine failure thresholds for the product and training of the neural network. The test set is used to verify the training effect of the neural network and the validation set is used to verify the validity of the proposed method.
4. According to the samples of the training set, the values of the fixed parameters theta and the distribution of the random parameters lambda and sigma are obtained through a two-stage method. If for products where security monitoring is not high, reliability assessment can be made with a definition based on time of arrival.
Where K is the total number of samples extracted from the distribution of random parameters λ and σ, m is the index of the extracted samples, and w is the failure threshold.
5. If there is a high demand for security monitoring of individual products, online updating of parameters is required. The situation described above can only reflect the overall characteristics of the degradation process of the same equipment, and is difficult to meet the residual life prediction requirements of the engineering, which are more concerned, for the specific service equipment individuals. Taking into account the subsequent acquisition of degradation data, an observation sequence may be utilized
The problem of parameter estimation is regarded as a recursive Bayesian filtering problem, i.e. estimating posterior distribution
wherein ,
for cut-off t k Time of day on-line degenerate state sequence,/->
At t k Model parameters estimated by time zones; updating the offline parameters by using the observation sequence to better reflect individual degradation difference, and constructing a state space model as +.>
wherein ,
due to the parameter B (t k ) The existence of (2) can lead to the high weight of the current particles caused by noise, so that the following t is carried out according to the long-short-term memory neural network k+n And predicting the moment. The degenerate sequence predicted from the neural network is defined as +.>
Keep->
The parameters are unchanged and then can be recursively scaled to t based on equation (5) k+n At the moment, the calculation of the weight is performed at each step of the recursion. T is recorded k The weighted particle set of the parameters of the moment is +.>
M is the number of particles.
wherein ,
at t k-1 Time-of-day predicted particle state and t k Weight given by likelihood between real states of time system, the same thing->
It can be understood that the real state of the system is converted into the state predicted by the long-short-term neural network, and N is the number of steps predicted by the neural network. The initial condition is the procedure estimated in step 4, i.e. randomly sampling from the distribution obeyed by λ and σ. According to the principle of particle filtering, at t k The time can be obtained
wherein ,
for m=1, 2l, m is normalized according to the weights, resulting in
When the particle filtering algorithm is executed, serious particle degradation phenomenon exists, namely, the importance weight ratio of only a small number of particles is larger, and the importance weight of most particles is very small. Therefore, in the resampling step, the particles with larger weight values are copied, and the particles with smaller weight values are discarded. To measure the degree of degradation of particles, an effective sample size is introduced, defined as
wherein ,Neff Smaller means more serious particle degradation phenomenon, and the resampling threshold is set to be N th When N eff <N th At that time, resampling is performed. Thus at t k The calculation formula of the residual life probability density of the time prediction degradation product is as follows:
6. combined drawing2-6, a comparative experiment was performed using the NASA turbofan engine dataset developed using the commercial modular aero propulsion system simulation model (Commercial Modular Aero Propulsion System Simulation, C-MAPSS) developed by the united states aerospace agency army research laboratory, with 21 features selected from the system output to characterize the degradation process of the engine. The FD002 training set was used as experimental samples. Introducing a mean square error (mean square error, MSE) index that takes into account both the accuracy and uncertainty of the remaining life prediction, at t k The time is defined as follows:
in formula (13):
at t k The actual remaining life of the device is degraded at the moment. The smaller the value of MSE, the higher the accuracy of the predicted remaining life of the method, and the more accurate the result. Table 1 below shows the experimental results of MSE for the validation set:
| MSE | |
| particle filtering without neural network | 1.8686e+03 |
| Particle filtering with long and short term memory neural network | 1.8341e+03 |
As shown by the experimental results in Table 1, the MSE value corresponding to the particle filtering of the long-term memory neural network is smaller than that of the particle filtering of the neural network, which indicates that the product fused with the long-term memory neural network and the particle filtering has higher prediction accuracy and more accurate result.
Referring to fig. 7, an embodiment of the present application further provides a device for predicting a remaining life of a product, which may implement the method for predicting a remaining life of a product, where the device includes:
an acquisition module 710, configured to acquire degradation data of a product, and extract a degradation curve according to the degradation data;
the modeling module 720 is configured to perform degradation modeling on the degradation curve according to the wiener process, obtain a degradation model, and determine offline parameters according to the degradation model;
the training module 730 is configured to divide the degradation curve into a training set, a testing set and a verification set, and perform training of the long-term and short-term neural network;
the establishing module 740 is configured to establish a state transition equation and a measurement equation according to the wiener process, and sample from offline parameters;
the recursion module 750 is configured to recursively perform a state transfer equation to n time according to the system state of the n time after the long-term memory neural network predicts, and the particle weight in the recursion process is synthesized to be the particle weight of the current time;
the prediction module 760 is configured to normalize according to the particle weight, and obtain a remaining lifetime probability density of the product at the current time;
the updating module 770 is configured to update parameters according to the real state data acquired at each time.
Based on this, in the product remaining life prediction apparatus of the embodiment of the present application, the acquisition module 710 acquires degradation data of a product, extracts a degradation curve according to the degradation data, and the degradation curve is used to characterize a probability density of product failure; the modeling module 720 carries out degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determines offline parameters according to the degradation model; the training module 730 divides the degradation curve into a training set, a testing set and a verification set, and performs training of the long-term and short-term neural network; the establishing module 740 establishes a state transition equation and a measurement equation according to the wiener process, and samples from the offline parameters; the recursion module 750 recursions the state transfer equation to the n time according to the system state of the n time after the long-term memory neural network prediction and the particle parameters of the current time, and synthesizes the particle weight in the recursion process as the particle weight of the current time; the prediction module 760 normalizes according to the particle weight to obtain the remaining life probability density of the product at the current time; the update module 770 performs parameter update according to the real state data acquired at each time. According to the embodiment of the application, the degradation data of the product are obtained, and the degradation curve is extracted according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the application provides a product residual life prediction method integrating a long-term memory neural network and particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product. According to the embodiment of the application, through the combination of the prediction capability of the neural network and the particle filtering, the probability density of failure can be given, and the prediction accuracy is higher compared with that of the common particle filtering.
The specific implementation of the device for predicting the remaining life of the product is basically the same as the specific embodiment of the method for predicting the remaining life of the product, and will not be described herein.
The embodiment of the application also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the method for predicting the residual life of the product when executing the computer program. The electronic equipment can be any intelligent terminal including a tablet personal computer, a vehicle-mounted computer and the like.
Referring to fig. 8, fig. 8 illustrates a hardware structure of an electronic device according to another embodiment, the electronic device includes:
the processor 801 may be implemented by a general-purpose CPU (central processing unit), a microprocessor, an application-specific integrated circuit (ApplicationSpecificIntegratedCircuit, ASIC), or one or more integrated circuits, etc. for executing related programs to implement the technical solutions provided in the embodiments of the present application.
The memory 802 may be implemented in the form of read-only memory (ReadOnlyMemory, ROM), static storage, dynamic storage, or random access memory (RandomAccessMemory, RAM). The memory 802 may store an operating system and other application programs, and when the technical solution provided in the embodiments of the present disclosure is implemented by software or firmware, relevant program codes are stored in the memory 802, and the processor 801 invokes a method for predicting the remaining life of a product according to the embodiments of the present disclosure, that is, by acquiring degradation data of the product, extracting a degradation curve according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the application provides a product residual life prediction method integrating a long-term memory neural network and particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product. According to the embodiment of the application, through the combination of the prediction capability of the neural network and the particle filtering, the probability density of failure can be given, and the prediction accuracy is higher compared with that of the common particle filtering.
An input/output interface 803 for implementing information input and output.
The communication interface 804 is configured to implement communication interaction between the device and other devices, and may implement communication through a wired manner (such as USB, network cable, etc.), or may implement communication through a wireless manner (such as mobile network, WI F I, bluetooth, etc.).
A bus that transfers information between the various components of the device (e.g., processor 801, memory 802, input/output interface 803, and communication interface 804).
Wherein the processor 801, the memory 802, the input/output interface 803, and the communication interface 804 implement communication connection between each other inside the device through a bus.
The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program realizes the method for predicting the residual life of the product when being executed by a processor.
The memory, as a non-transitory computer readable storage medium, may be used to store non-transitory software programs as well as non-transitory computer executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory remotely located relative to the processor, the remote memory being connectable to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
According to the product remaining life prediction method, the product remaining life prediction device, the electronic equipment and the storage medium, degradation data of a product are obtained, and a degradation curve is extracted according to the degradation data; performing degradation modeling on the degradation curve according to the wiener process to obtain a degradation model, and determining offline parameters according to the degradation model; dividing a degradation curve into a training set, a testing set and a verification set, and training a long-term and short-term neural network; establishing a state transition equation and a measurement equation according to the wiener process, and sampling from offline parameters; according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time; normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment; and updating parameters according to the real state data acquired at each moment. Based on the method, the application provides a product residual life prediction method integrating a long-term memory neural network and particle filtering. According to the method, the residual service life of the product is predicted, and the prediction accuracy is good. The method not only can calculate the residual life density curve of the product, but also increases the prediction precision on the basis of common particle filtering by fusing a long-period and short-period memory neural network, reduces the particle degradation caused by noise, and can provide theoretical basis and technical support for sustainable development of the safety and economy of the product. According to the embodiment of the application, through the combination of the prediction capability of the neural network and the particle filtering, the probability density of failure can be given, and the prediction accuracy is higher compared with that of the common particle filtering.
Those of ordinary skill in the art will appreciate that all or some of the steps, systems, and methods disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable programs, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable programs, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.
The embodiments described in the embodiments of the present application are for more clearly describing the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application, and as those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.
It will be appreciated by those skilled in the art that the technical solutions shown in the figures do not constitute limitations of the embodiments of the present application, and may include more or fewer steps than shown, or may combine certain steps, or different steps.
The above described apparatus embodiments are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
Those of ordinary skill in the art will appreciate that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof.
The terms "first," "second," "third," "fourth," and the like in the description of the present application and in the above-described figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in this application, "at least one" means one or more, and "a plurality" means two or more. "and/or" for describing the association relationship of the association object, the representation may have three relationships, for example, "a and/or B" may represent: only a, only B and both a and B are present, wherein a, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one (one) of a, b or c may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the above-described division of units is merely a logical function division, and there may be another division manner in actual implementation, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including multiple instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of the various embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing a program.
Preferred embodiments of the present application are described above with reference to the accompanying drawings, and thus do not limit the scope of the claims of the embodiments of the present application. Any modifications, equivalent substitutions and improvements made by those skilled in the art without departing from the scope and spirit of the embodiments of the present application shall fall within the scope of the claims of the embodiments of the present application.
Claims (10)
1. A method of predicting remaining life of a product, the method comprising:
acquiring degradation data of a product, and extracting a degradation curve according to the degradation data;
performing degradation modeling on the degradation curve according to a wiener process to obtain a degradation model, and determining offline parameters according to the degradation model;
dividing the degradation curve into a training set, a testing set and a verification set, and training a long-period and short-period neural network;
establishing a state transition equation and a measurement equation according to a wiener process, and sampling from the offline parameters;
according to the system state of the n time after the long-term memory neural network prediction, carrying out state transfer equation recursion to the n time according to the particle parameters of the current time, and integrating the particle weight in the recursion process as the particle weight of the current time;
normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment;
and updating parameters according to the real state data acquired at each moment.
2. The method of claim 1, wherein the degradation model is:
the value of the fixed parameter θ and the distribution of the random parameters λ and σ are obtained by a two-stage method.
3. The method of claim 2, wherein the establishing state transition equations and measurement equations according to the wiener process comprises:
will utilize the observation sequence
The parameter estimation problem is regarded as a recursive Bayesian filtering problem, and posterior distribution is estimated
wherein ,
for cut-off t k Time of day on-line degenerate state sequence,/->
At t k Model parameters estimated by time zones; updating the offline parameters by using the observation sequence, and constructing a state space model as follows
wherein ,
4. the method of claim 3, wherein the recursively performing the state transition equation to the n time according to the particle parameters at the current time based on the system state at the n time after the prediction by the long-short term memory neural network comprises:
the predicted degradation sequence according to the neural network is defined as
Keep->
The parameters are unchanged and then based on the formula +.>
Can be recursively t k+n At the moment, the calculation of the weight is performed at each step of the recursion.
5. The method of claim 4, wherein the integrating the particle weights in the recursive process is the particle weight at the current time, comprising:
t is recorded k The weighted particle set of the parameters of the time is
M is the number of particles, and the number of particles,
wherein ,
at t k-1 Time-of-day predicted particle state and t k Between time of day system true statesWeight given by likelihood>
As a method for converting the real state of the system into the state predicted by the long-short term neural network, N is the number of steps predicted by the neural network.
6. The method of claim 5, wherein normalizing according to the particle weight to obtain a remaining lifetime probability density of the product at the current time comprises:
according to the principle of particle filtering, at t k The time can be obtained
Wherein, delta () is the dirac function, then
wherein ,
for m=1, 2l, m is normalized according to the weights, resulting in
Thus at t k The calculation formula of the residual life probability density of the time prediction degradation product is as follows:
7. the method of claim 6, comprising, after said normalizing according to particle weight:
copying the particles with larger weight values, discarding the particles with smaller weight values, and introducing effective sampling size, which is defined as:
wherein ,Neff Smaller means more serious particle degradation phenomenon, and the resampling threshold is set to be N th When N eff <N th At that time, resampling is performed.
8. A product remaining life prediction apparatus, the apparatus comprising:
the acquisition module is used for acquiring degradation data of the product and extracting a degradation curve according to the degradation data;
the modeling module is used for carrying out degradation modeling on the degradation curve according to a wiener process to obtain a degradation model, and determining offline parameters according to the degradation model;
the training module is used for dividing the degradation curve into a training set, a testing set and a verification set and training a long-term and short-term neural network;
the establishing module is used for establishing a state transition equation and a measurement equation according to the wiener process and sampling from the offline parameters;
the recursion module is used for recursing a state transfer equation to the n time according to the system state of the n time after the long-term memory neural network is predicted and the particle weight in the recursion process is comprehensively the particle weight of the current time;
the prediction module is used for normalizing according to the particle weight to obtain the residual life probability density of the product at the current moment;
and the updating module is used for updating parameters according to the real state data acquired at each moment.
9. An electronic device comprising a memory storing a computer program and a processor that when executing the computer program implements the method of predicting remaining life of a product as claimed in any one of claims 1 to 7.
10. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method of predicting remaining life of a product according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310160024.2A CN116090353A (en) | 2023-02-22 | 2023-02-22 | Product remaining life prediction method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310160024.2A CN116090353A (en) | 2023-02-22 | 2023-02-22 | Product remaining life prediction method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116090353A true CN116090353A (en) | 2023-05-09 |
Family
ID=86210212
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310160024.2A Pending CN116090353A (en) | 2023-02-22 | 2023-02-22 | Product remaining life prediction method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116090353A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116756881A (en) * | 2023-08-21 | 2023-09-15 | 人工智能与数字经济广东省实验室(广州) | Bearing residual service life prediction method, device and storage medium |
-
2023
- 2023-02-22 CN CN202310160024.2A patent/CN116090353A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116756881A (en) * | 2023-08-21 | 2023-09-15 | 人工智能与数字经济广东省实验室(广州) | Bearing residual service life prediction method, device and storage medium |
| CN116756881B (en) * | 2023-08-21 | 2024-01-05 | 人工智能与数字经济广东省实验室(广州) | Bearing residual service life prediction method, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109657937B (en) | Product reliability evaluation and service life prediction method based on degradation data | |
| EP3385889A1 (en) | Abnormality detection system, abnormality detection method, abnormality detection program, and method for generating learned model | |
| Xu et al. | Online identification of battery model parameters and joint state of charge and state of health estimation using dual particle filter algorithms | |
| US20180046724A1 (en) | Information recommendation method and apparatus, and server | |
| CN110647414A (en) | Improved analog function security with anomaly detection | |
| JP2019061565A (en) | Abnormality diagnostic method and abnormality diagnostic device | |
| CN112702342B (en) | Network event processing method and device, electronic equipment and readable storage medium | |
| CN109143094B (en) | Abnormal data detection method and device for power battery | |
| EP2963553A1 (en) | System analysis device and system analysis method | |
| CN112835769A (en) | Service data abnormity diagnosis method, device, equipment and storage medium | |
| D'Amico et al. | Wind speed modeled as an indexed semi‐Markov process | |
| Kachour et al. | First‐order rounded integer‐valued autoregressive (RINAR (1)) process | |
| CN116090353A (en) | Product remaining life prediction method and device, electronic equipment and storage medium | |
| Park et al. | Missing-insensitive short-term load forecasting leveraging autoencoder and LSTM | |
| CN113822366A (en) | Service index abnormality detection method and device, electronic equipment and storage medium | |
| Man et al. | Remaining useful life prediction for hard failures using joint model with extended hazard | |
| Veerakumar et al. | PMU-based real-time distribution system state estimation considering anomaly detection, discrimination and identification | |
| CN110333078B (en) | Rolling bearing degradation state stage determination method | |
| Liu et al. | Dynamic traffic demand uncertainty prediction using radio‐frequency identification data and link volume data | |
| Hua et al. | Efficient probabilistic contingency analysis through a stability measure considering wind perturbation | |
| JP2019105871A (en) | Abnormality candidate extraction program, abnormality candidate extraction method and abnormality candidate extraction apparatus | |
| CN113110961B (en) | Equipment abnormality detection method and device, computer equipment and readable storage medium | |
| Wang et al. | Evaluating weather influences on transmission line failure rate based on scarce fault records via a bi‐layer clustering technique | |
| Ouyang et al. | Optimisation of time window size for wind power ramps prediction | |
| Zhang et al. | A verification methodology for prognostic algorithms |
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 |