Degradation knowledge and data fusion driven migration decoupling method based on small samples

Abstract

The invention discloses a migration decoupling method driven by degradation knowledge and data fusion based on small samples, and relates to the technical field of predictive maintenance of rotating machinery. The method comprises the following steps: acquiring source domain data and target domain data of the rotating machine; determining a source domain invariant feature, a source domain private feature, a target domain invariant feature and a target domain private feature based on the initial life prediction model; constructing a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted life; constructing a degradation mechanism spatial state model according to the source domain invariant features and the target domain invariant features; estimating parameters of the degradation mechanism spatial state model step by step by using a particle filtering algorithm; and when the initial life prediction model continuously performs weight updating based on the loss function, the degradation mechanism spatial state model after parameter estimation is utilized to guide the weight updating process of the initial life prediction model. The generalization ability and the interpretability of the model can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of predictive maintenance of rotating machinery, and in particular to a migration decoupling method driven by degradation knowledge and data fusion based on small samples. Background Art

[0002] Industrial equipment is increasingly becoming large-scale, automated, and intelligent. Rotating machinery such as bearings and gears are widely used in aerospace, transportation, and other fields, and their operating conditions are directly related to the life of various types of mechanical equipment. Therefore, life prediction of rotating machinery such as bearings has positive practical significance. In order to solve the small sample problem faced in the life prediction of rotating machinery, transfer learning theory is introduced into the prediction model to reduce the distribution difference between the source domain and the target domain.

[0003] At present, most transfer learning research focuses on minimizing the difference in global feature distribution across domains, ignoring the inherent properties of the domain. However, if global alignment is directly adopted without proper restrictions on features, the cross-domain representation may be affected and it is difficult to ensure the generalization ability of the model. Summary of the invention

[0004] In view of this, the present application provides a migration decoupling method driven by small sample-based degradation knowledge and data fusion, which not only reveals the similarities between domains, but also takes into account the uniqueness of the domains, ensures the training effect of the model, and improves the generalization ability of the model.

[0005] According to a first aspect of the present application, a migration decoupling method driven by degradation knowledge and data fusion based on a small sample is provided, the method comprising:

[0006] Acquire source domain data and target domain data of the rotating machinery, and an initial life prediction model, wherein the initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder, and a life predictor;

[0007] Based on the source domain data and the target domain data, and the initial life prediction model, determine source domain invariant features and source domain private features corresponding to the source domain data, target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery;

[0008] Constructing a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted lifespan;

[0009] Constructing a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determining initial parameters of the degradation mechanism space state model;

[0010] Based on the initial parameters, the parameters of the degradation mechanism space state model are gradually estimated using a particle filter algorithm to obtain a degradation mechanism space state model after parameter estimation;

[0011] When the initial life prediction model continuously updates its weight based on the loss function, the degradation mechanism space state model after parameter estimation is used to guide the weight update process of the initial life prediction model until a preset number of iterations is reached, and then a preset life prediction model is output.

[0012] According to a second aspect of the present application, a migration decoupling device driven by degradation knowledge and data fusion based on a small sample is provided, the device comprising:

[0013] An acquisition unit, used to acquire source domain data and target domain data of the rotating machinery, and an initial life prediction model, wherein the initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder and a life predictor;

[0014] a determining unit, configured to determine, based on the source domain data and the target domain data, and the initial life prediction model, source domain invariant features and source domain private features corresponding to the source domain data, target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery;

[0015] A first construction unit is used to construct a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted lifespan;

[0016] A second construction unit is used to construct a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determine initial parameters of the degradation mechanism space state model;

[0017] An estimation unit, configured to gradually estimate the parameters of the degradation mechanism space state model based on the initial parameters by using a particle filter algorithm to obtain the degradation mechanism space state model after parameter estimation;

[0018] A training unit is used to guide the weight update process of the initial life prediction model by using the degradation mechanism space state model after parameter estimation when the initial life prediction model continuously updates its weight based on the loss function, until a preset number of iterations is reached, and then a preset life prediction model is output.

[0019] According to a third aspect of the present application, a storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the above-mentioned migration decoupling method driven by degradation knowledge and data fusion based on small samples is implemented.

[0020] According to the fourth aspect of the present application, an electronic device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein when the processor executes the program, the above-mentioned migration decoupling method driven by degradation knowledge and data fusion based on small samples is implemented.

[0021] Through the above technical solution, the present application provides a migration decoupling method driven by degradation knowledge and data fusion based on small samples. Compared with the prior art, by extracting the source domain invariant features, source domain private features, target domain invariant features and target domain private features respectively, the domain private features and domain invariant features can be decoupled. This method of the present application not only reveals the similarities between domains, but also takes into account the uniqueness of the domains, greatly improves the generalization ability of the model, and effectively addresses the global domain generalization problem in the case of small samples. In addition, by integrating the degradation mechanism into the life prediction model, the present application can not only enhance the accuracy of the prediction, but also improve the robustness of the prediction model to abnormal data and noise. Furthermore, in order to improve the perception ability of the prediction model for dynamic time relationships, the present application uses a particle filter algorithm for forward step-by-step prediction, captures the dependency between long-distance time features, and improves the interpretability of the prediction model.

[0022] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 A flow chart of a migration decoupling method driven by degradation knowledge and data fusion based on a small sample provided in an embodiment of the present application is shown;

[0025] Figure 2 A schematic diagram of feature alignment provided by an embodiment of the present application is shown;

[0026] Figure 3 A schematic diagram of a process flow of a feature extraction method provided in an embodiment of the present application is shown;

[0027] Figure 4 A schematic diagram of the process of data preprocessing, model training and model testing provided in an embodiment of the present application is shown;

[0028] Figure 5 A schematic diagram of the structure of a migration decoupling device driven by degradation knowledge and data fusion based on a small sample provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0029] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0030] The existing technology directly uses global alignment without proper restrictions on features, which affects the cross-domain representation and makes it difficult to ensure the generalization ability of the model.

[0031] In order to solve the above problems, the embodiment of the present invention provides a migration decoupling method based on small sample degradation knowledge and data fusion drive, such as Figure 1 As shown, the method includes:

[0032] Step 10: Obtain source domain data and target domain data of the rotating machinery, as well as an initial life prediction model.

[0033] Among them, the rotating machinery can specifically include gears, engines, bearings, etc. The initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder and a life predictor, wherein the domain invariant feature decoder and the domain mixed feature decoder are used to disentangle the domain invariant features and the domain private features, which can maximize the relative amount of information; the domain mixed feature decoder is used to restore the original feature information to ensure the minimum loss of the relative amount of information in a single domain, and the domain invariant feature decoder is used to reconstruct the original features, balance the relative amount of information in a single domain and a dual domain, and meet better cross-domain migration universality. The target domain private feature extractor, the source domain private feature extractor and the domain invariant feature extractor are based on a multi-layer perceptron framework, the life predictor includes a fully connected layer, the domain invariant feature decoder and the domain mixed feature decoder include an upsampling / deconvolution layer, a convolution layer and a decoding layer.

[0034] The embodiments of the present invention are mainly applicable to health monitoring of rotating machinery, mainly involving key equipment such as aircraft engines, railway vehicles and wind turbines.

[0035] For the embodiment of the present invention, the original vibration data is first obtained, and then the time domain features, frequency domain features and trigonometric function-based features corresponding to the original vibration data are extracted. Among them, the time domain features include entropy, energy, root mean square, kurtosis, root mean square, absolute mean, maximum absolute value, skewness, shape factor, clearance coefficient, impulse factor, crest factor, standard deviation, etc.; the frequency domain features include the energy of 16 frequency bands generated by four-level wavelet packet decomposition; the trigonometric function-based features include inverse hyperbolic cosine standard deviation, inverse hyperbolic sine standard deviation and inverse tangent standard deviation, etc. The extracted time domain features, frequency domain features and trigonometric function-based features are used as source domain data and target domain data of rotating machinery.

[0036] At the same time, the network parameters are initialized and the initial life prediction model is constructed. The initial life prediction model includes six modules, namely, target domain private feature extractor, source domain private feature extractor, domain invariant feature extractor, domain invariant feature decoder, domain mixed feature decoder and life predictor.

[0037] Step 20: Based on the source domain data and the target domain data, and the initial life prediction model, determine the source domain invariant features and source domain private features corresponding to the source domain data, the target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery.

[0038] In order to maximize the mining of cross-domain similarity features, the embodiment of the present invention separates domain-private features and domain-invariant features by feature decoupling, such as Figure 2 As shown, compared with the global feature comparison method, this division method of the embodiment of the present invention not only reveals the similarities between domains, but also takes into account the uniqueness of the domains, greatly improves the generalization ability of the model, and effectively addresses the global domain generalization problem in the case of small samples.

[0039] The extraction process of source domain invariant features, source domain private features, target domain invariant features and target domain private features is as follows: Figure 3 As shown, the method includes:

[0040] Step 21: Utilize the domain-invariant feature extractor to respectively extract source domain-invariant original features corresponding to the source domain data and target domain-invariant original features corresponding to the target domain data.

[0041] According to the embodiment of the present invention, the source domain data and the target domain data are respectively input into the domain-invariant feature extractor for feature extraction to obtain the source domain-invariant original features and the target domain-invariant original features, wherein the source domain-invariant original features and the target domain-invariant original features are features that have not been decoded.

[0042] Step 22: Utilize the target domain private feature extractor and the source domain private feature extractor to respectively extract the source domain private original features corresponding to the source domain data and the target domain private original features corresponding to the target domain data.

[0043] In the embodiment of the present invention, the source domain data and the original features unchanged in the source domain are input into the source domain private feature extractor for feature extraction to obtain the original features private in the source domain. Similarly, the target domain data and the original features unchanged in the target domain are input into the target domain private feature extractor for feature extraction to obtain the original features private in the target domain. Among them, the original features private in the source domain and the original features private in the target domain are features that have not been decoded.

[0044] Step 23: Use the domain-invariant feature decoder to decode the source domain-invariant original features and the target domain-invariant original features respectively to obtain the source domain-invariant features and the target domain-invariant features, and use the domain-mixed feature decoder to decode the source domain-private original features and the target domain-private original features respectively to obtain the source domain-private features and the target domain-private features.

[0045] For the embodiment of the present invention, the source domain invariant original features and the target domain invariant original features are respectively input into the domain invariant feature decoder for decoding to obtain the source domain invariant features and the target domain invariant features. At the same time, the source domain private original features and the target domain private original features are input into the domain mixed feature decoder for decoding to obtain the source domain private features and the target domain private features.

[0046] Step 24: input the source domain invariant features and the source domain private features into the life predictor for prediction to obtain the predicted life of the rotating machinery.

[0047] According to the embodiment of the present invention, the decoded source domain invariant features and source domain private features are input into the life predictor (fully connected layer) for prediction to obtain the predicted life of the rotating machinery.

[0048] The embodiment of the present invention can decouple domain-private features and domain-invariant features by respectively extracting source domain invariant features, source domain private features, target domain invariant features and target domain private features. This method of the embodiment of the present invention not only reveals the similarities between domains, but also takes into account the uniqueness of domains, greatly improves the generalization ability of the model, and effectively addresses the global domain generalization problem in small sample conditions.

[0049] Step 30: construct a loss function according to the source domain invariant features, the target domain invariant features, the source domain private features, the target domain private features and the predicted lifespan.

[0050] For the embodiment of the present invention, after extracting the source domain invariant features, the target domain invariant features, the source domain private features and the target domain private features, a loss function is constructed based on the source domain invariant features, the target domain invariant features, the source domain private features and the target domain private features. With respect to the construction process of the loss function, step 30 specifically includes: based on the source domain invariant features, the target domain invariant features, the source domain private features and the target domain private features, respectively calculating the first distribution difference loss between the source domain private features and the source domain invariant features, the second distribution difference loss between the target domain private features and the target domain invariant features, and the third distribution difference loss between the source domain invariant features and the target domain invariant features; calculating the life prediction loss based on the predicted life and the actual life of the rotating machinery; constructing a loss function based on the life prediction loss, the first distribution difference loss, the second distribution difference loss and the third distribution difference loss.

[0051] Among them, the first distribution difference loss, the second distribution difference loss and the third distribution difference loss can be calculated based on the domain adaptation theory, and the specific formula is as follows:

[0052]

[0053] Among them, X S represents the source domain features, including source domain private features and source domain invariant features, X T represents the target domain features, including target domain private features and target domain invariant features, n S Represents the number of samples corresponding to the source domain data, n T represents the number of samples corresponding to the target domain data, H is the regenerated Hilbert space, and φ represents X S and X T Mapping on H.

[0054] Extracting private features from the source domain Source domain invariant features Target domain private features and target domain invariant features Then, the MMD metric is used as the loss function to measure the distribution, including and The first distribution difference loss between and The difference between the second distribution loss and The third distribution difference loss L similarity .

[0055] Step 40: construct a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determine initial parameters of the degradation mechanism space state model.

[0056] For the embodiment of the present invention, after extracting the invariant features of the source domain and the invariant features of the target domain, a degradation mechanism space state model is established based on the invariant features of the source domain and the invariant features of the target domain. The degradation mechanism space state model is essentially a double exponential function, which is used to simulate the degradation process of the machinery. The main purpose of this model is to process random events in time series and predict future signal trends.

[0057] Regarding the construction process of the degradation mechanism spatial state model, the method includes: according to the source domain invariant features and the target domain invariant features, using a double exponential function to simulate the degradation process of the rotating machinery, and obtaining the degradation mechanism spatial state model.

[0058] ψ(t n ,a,b,c,d)=a*exp(b*t n )+c*exp(d*t n )

[0059] Among them, a, b, c, d are the parameters of the degradation mechanism space state model, t n For time series.

[0060] Then, the maximum likelihood algorithm is used to solve the initial parameters of the degradation mechanism space state model, that is, the maximum likelihood estimation method is used to make the likelihood function L(Φ / R) obtain the maximum value, and then the optimal initial parameters of the degradation mechanism space state model are obtained.

[0061]

[0062] Step 50: Based on the initial parameters, the parameters of the degradation mechanism space state model are gradually estimated using a particle filter algorithm to obtain a degradation mechanism space state model after parameter estimation.

[0063] According to the embodiment of the present invention, after the initial parameters of the degradation mechanism space state model are determined, a particle filter algorithm based on resampling is used to gradually estimate the parameters of the degradation mechanism space state model in chronological order to predict the future state.

[0064] Based on this, step 50 specifically includes: based on the initial parameters, using the particle filter algorithm to determine a parameter update formula; according to the parameter update formula, gradually estimating the parameters of the degradation mechanism space state model to obtain the degradation mechanism space state model after parameter estimation.

[0065] Specifically, firstly, the historical observation data R 1:n-1 ∈Rn-1 The prior probability density function at time n is obtained. The specific formula is as follows:

[0066] p(s n / R 1:n-1 )=∫p(s n / s n-1 )p(s n-1 / R 1:n-1 )ds n-1

[0067] Among them, p(s n / s n-1 ) is the probability density function of the state transition function, p(s n-1 / R 1:n-1 ) is the probability density function at time n-1. After the observation value at time n is obtained, the particle is updated and the posterior probability density function at time n is obtained. The specific formula is as follows:

[0068]

[0069] in, represents the particle weight at time n, Represents the distribution p(s n / R 1:n ) is the particle obtained by importance sampling. After the observation value at time n is obtained, the update method (parameter update formula) of the particle weight is as follows:

[0070]

[0071] After the particle filter updates the parameters at each moment, the degradation mechanism space state model can be correctly solved respectively, and the next life prediction can be carried out according to the established two-stage double exponential model.

[0072] Step 60: When the initial life prediction model continuously updates its weight based on the loss function, the degradation mechanism space state model after parameter estimation is used to guide the weight update process of the initial life prediction model until a preset number of iterations is reached, and then a preset life prediction model is output.

[0073] The preset number of iterations may be set according to actual business requirements, and the embodiment of the present invention does not specifically limit this.

[0074] For the embodiment of the present invention, after the preset life prediction model is trained, the target domain prediction data of the rotating machinery can be obtained, and the target domain prediction data can be input into the preset life prediction model to perform life prediction, thereby obtaining the life prediction result of the rotating machinery. Figure 4As shown, it specifically includes data processing, model training and model testing.

[0075] The embodiment of the present invention provides a migration decoupling method driven by degradation knowledge and data fusion based on a small sample. Compared with the prior art, by extracting the source domain invariant features, source domain private features, target domain invariant features and target domain private features respectively, the domain private features and domain invariant features can be decoupled. This method of the present application not only reveals the similarities between domains, but also takes into account the uniqueness of the domains, greatly improves the generalization ability of the model, and effectively addresses the global domain generalization problem in the case of small samples. In addition, by integrating the degradation mechanism into the life prediction model, the present application can not only enhance the accuracy of the prediction, but also improve the robustness of the prediction model to abnormal data and noise. Furthermore, in order to improve the perception ability of the prediction model for dynamic time relationships, the present application uses a particle filter algorithm for forward step-by-step prediction, captures the dependency between long-distance time features, and improves the interpretability of the prediction model.

[0076] Further, as Figure 1 and Figure 3 The specific implementation of the method shown in the embodiment provides a migration decoupling device driven by degradation knowledge and data fusion based on small samples, such as Figure 5 As shown, the device includes: an acquisition unit 101, a determination unit 102, a first construction unit 103, a second construction unit 104, an estimation unit 105 and a training unit 106.

[0077] The acquisition unit 101 can be used to acquire source domain data and target domain data of rotating machinery, as well as an initial life prediction model, wherein the initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder and a life predictor.

[0078] The determination unit 102 can be used to determine the source domain invariant features and source domain private features corresponding to the source domain data, the target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery based on the source domain data and the target domain data, and the initial life prediction model.

[0079] The first construction unit 103 may be configured to construct a loss function according to the source domain invariant features, the target domain invariant features, the source domain private features, the target domain private features and the predicted lifespan.

[0080] The second construction unit 104 may be used to construct a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determine initial parameters of the degradation mechanism space state model.

[0081] The estimation unit 105 may be used to gradually estimate the parameters of the degradation mechanism space state model based on the initial parameters using a particle filter algorithm to obtain the degradation mechanism space state model after parameter estimation.

[0082] The training unit 106 can be used to guide the weight update process of the initial life prediction model by using the degradation mechanism space state model after parameter estimation when the initial life prediction model continuously updates its weight based on the loss function, until a preset number of iterations is reached, and then a preset life prediction model is output.

[0083] In some embodiments, the determination unit 102 can be specifically used to use the domain-invariant feature extractor to respectively extract the source domain-invariant original features corresponding to the source domain data and the target domain-invariant original features corresponding to the target domain data; use the target domain-private feature extractor and the source domain-private feature extractor to respectively extract the source domain-private original features corresponding to the source domain data and the target domain-private original features corresponding to the target domain data; use the domain-invariant feature decoder to respectively decode the source domain-invariant original features and the target domain-invariant original features to obtain the source domain-invariant features and the target domain-invariant features, and use the domain mixed feature decoder to respectively decode the source domain-private original features and the target domain-private original features to obtain the source domain-private features and the target domain-private features; input the source domain-invariant features and the source domain-private features into the life predictor for prediction to obtain the predicted life of the rotating machinery.

[0084] In some embodiments, the first construction unit 103 can be specifically used to calculate, based on the source domain invariant feature, the target domain invariant feature, the source domain private feature and the target domain private feature, a first distribution difference loss between the source domain private feature and the source domain invariant feature, a second distribution difference loss between the target domain private feature and the target domain invariant feature, and a third distribution difference loss between the source domain invariant feature and the target domain invariant feature; calculate the life prediction loss based on the predicted life and the actual life of the rotating machinery; and construct a loss function according to the life prediction loss, the first distribution difference loss, the second distribution difference loss and the third distribution difference loss.

[0085] In some embodiments, the first construction unit 103 may be specifically configured to simulate the degradation process of the rotating machinery using a double exponential function according to the source domain invariant features and the target domain invariant features, so as to obtain the degradation mechanism spatial state model.

[0086] In some embodiments, the first construction unit 103 may also be specifically configured to solve the initial parameters of the degradation mechanism space state model using a maximum likelihood algorithm.

[0087] In some embodiments, the estimation unit 105 can be specifically used to determine a parameter update formula based on the initial parameters using the particle filter algorithm; according to the parameter update formula, the parameters of the degradation mechanism space state model are gradually estimated to obtain the degradation mechanism space state model after parameter estimation.

[0088] In some embodiments, the apparatus further comprises: a prediction unit.

[0089] The acquisition unit 101 may also be used to acquire target domain prediction data of the rotating machinery.

[0090] The prediction unit may be used to input the target domain prediction data into the preset life prediction model to perform life prediction, thereby obtaining a life prediction result of the rotating machinery.

[0091] It should be noted that for other corresponding descriptions of the functional units involved in the migration decoupling device driven by small sample degradation knowledge and data fusion provided by the embodiment of the present invention, reference can be made to Figure 1 and Figure 2 The corresponding description in will not be repeated here.

[0092] Based on the above Figure 1 and Figure 3 The method shown in the embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned Figure 1 and Figure 3 The migration decoupling method driven by small sample-based degradation knowledge and data fusion is shown.

[0093] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling an electronic device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0094] Based on the above Figure 1 and Figure 3 The method shown, and Figure 5In order to achieve the above-mentioned purpose, the embodiment of the present application also provides an electronic device, which can be a personal computer, a tablet computer, a server, or other network equipment, etc. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 and Figure 3 The migration decoupling method driven by small sample-based degradation knowledge and data fusion is shown.

[0095] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0096] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0097] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0098] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform, or by hardware.

[0099] The embodiments of the present invention can decouple domain private features and domain invariant features by respectively extracting source domain invariant features, source domain private features, target domain invariant features, and target domain private features. This method of the present application not only reveals the similarities between domains, but also takes into account the uniqueness of the domains, greatly improves the generalization ability of the model, and effectively addresses the global domain generalization problem in small sample conditions. In addition, by incorporating the degradation mechanism into the life prediction model, the present application can not only enhance the accuracy of the prediction, but also improve the robustness of the prediction model to abnormal data and noise. Furthermore, in order to improve the perception ability of the prediction model for dynamic time relationships, the present application uses a particle filter algorithm for forward step-by-step prediction, captures the dependencies between long-distance time features, and improves the interpretability of the prediction model.

[0100] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present application. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the description of the implementation scenario, or can be changed accordingly and located in one or more devices different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0101] The above serial numbers of this application are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of this application, but this application is not limited to them, and any changes that can be thought of by technicians in this field should fall within the scope of protection of this application.

Claims

1. A migration decoupling method driven by degradation knowledge and data fusion based on small samples, characterized in that: include: Acquire source domain data and target domain data of the rotating machinery, and an initial life prediction model, wherein the initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder, and a life predictor; Based on the source domain data and the target domain data, and the initial life prediction model, determine source domain invariant features and source domain private features corresponding to the source domain data, target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery; Constructing a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted lifespan; Constructing a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determining initial parameters of the degradation mechanism space state model; Based on the initial parameters, the parameters of the degradation mechanism space state model are gradually estimated using a particle filter algorithm to obtain a degradation mechanism space state model after parameter estimation; When the initial life prediction model continuously updates its weight based on the loss function, the degradation mechanism space state model after parameter estimation is used to guide the weight update process of the initial life prediction model until a preset number of iterations is reached, and then a preset life prediction model is output.

2. The method according to claim 1, characterized in that The determining, based on the source domain data and the target domain data, and the initial life prediction model, source domain invariant features and source domain private features corresponding to the source domain data, target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery includes: The domain-invariant feature extractor is used to respectively extract source domain-invariant original features corresponding to the source domain data and target domain-invariant original features corresponding to the target domain data; Utilizing the target domain private feature extractor and the source domain private feature extractor, respectively extracting source domain private original features corresponding to the source domain data and target domain private original features corresponding to the target domain data; The source domain invariant original features and the target domain invariant original features are decoded by the domain invariant feature decoder to obtain the source domain invariant features and the target domain invariant features, and the source domain private original features and the target domain private original features are decoded by the domain mixed feature decoder to obtain the source domain private features and the target domain private features; The source domain invariant features and the source domain private features are input into the life predictor for prediction to obtain the predicted life of the rotating machinery.

3. The method according to claim 1, characterized in that The constructing a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted lifespan includes: Based on the source domain invariant feature, the target domain invariant feature, the source domain private feature and the target domain private feature, respectively calculating a first distribution difference loss between the source domain private feature and the source domain invariant feature, a second distribution difference loss between the target domain private feature and the target domain invariant feature, and a third distribution difference loss between the source domain invariant feature and the target domain invariant feature; Calculating a predicted life loss based on the predicted life and the actual life of the rotating machine; A loss function is constructed according to the lifetime prediction loss, the first distribution difference loss, the second distribution difference loss and the third distribution difference loss.

4. The method according to any one of claims 1 to 3, characterized in that: The step of constructing a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features includes: According to the invariant features of the source domain and the invariant features of the target domain, a double exponential function is used to simulate the degradation process of the rotating machinery to obtain the degradation mechanism spatial state model.

5. The method according to any one of claims 1 to 3, characterized in that: The determining of the initial parameters of the degradation mechanism spatial state model comprises: The maximum likelihood algorithm is used to solve the initial parameters of the degradation mechanism space state model.

6. The method according to any one of claims 1 to 3, characterized in that: The step of gradually estimating the parameters of the degradation mechanism space state model based on the initial parameters by using a particle filter algorithm to obtain the degradation mechanism space state model after parameter estimation includes: Based on the initial parameters, determining a parameter update formula using the particle filter algorithm; According to the parameter updating formula, the parameters of the degradation mechanism space state model are gradually estimated to obtain the degradation mechanism space state model after parameter estimation.

7. The method according to claim 1, characterized in that The method further comprises: Acquiring target domain prediction data of the rotating machinery; The target domain prediction data is input into the preset life prediction model to perform life prediction, and obtain the life prediction result of the rotating machinery.

8. A migration decoupling device driven by degradation knowledge and data fusion based on small samples, characterized in that: include: An acquisition unit, used to acquire source domain data and target domain data of the rotating machinery, and an initial life prediction model, wherein the initial life prediction model includes a target domain private feature extractor, a source domain private feature extractor, a domain invariant feature extractor, a domain invariant feature decoder, a domain mixed feature decoder and a life predictor; a determining unit, configured to determine, based on the source domain data and the target domain data, and the initial life prediction model, source domain invariant features and source domain private features corresponding to the source domain data, target domain invariant features and target domain private features corresponding to the target domain data, and the predicted life of the rotating machinery; A first construction unit is used to construct a loss function according to the source domain invariant feature, the target domain invariant feature, the source domain private feature, the target domain private feature and the predicted lifespan; A second construction unit is used to construct a degradation mechanism space state model according to the source domain invariant features and the target domain invariant features, and determine initial parameters of the degradation mechanism space state model; An estimation unit, configured to gradually estimate the parameters of the degradation mechanism space state model based on the initial parameters by using a particle filter algorithm to obtain the degradation mechanism space state model after parameter estimation; A training unit is used to guide the weight update process of the initial life prediction model by using the degradation mechanism space state model after parameter estimation when the initial life prediction model continuously updates its weight based on the loss function, until a preset number of iterations is reached, and then a preset life prediction model is output.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images