WO2024087670A1 - Silicone rubber material performance prediction method and apparatus based on lstm neural network - Google Patents

Abstract

Disclosed in the present application are a silicone rubber material performance prediction method and apparatus based on a long short-term memory (LSTM) neural network. Dielectric performance data of a silicone rubber material in a plurality of aging stages are acquired, and fusion processing is performed on the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material; according to the comprehensive performance index sequence, network parameters of the LSTM neural network are optimized by using a discrete differential evolution method until the LSTM neural network reaches a preset convergence condition to obtain a silicone rubber material performance prediction model, thereby effectively avoiding the problems of insufficient fitting degree and low precision caused by selecting key parameters according to experience; and according to a comprehensive performance index initial sequence of a silicone rubber material to be predicted, target comprehensive performance index data of the silicone rubber material to be predicted in each aging stage is predicted by using the silicone rubber material performance prediction model, thereby implementing prediction of aging performance of the silicone rubber material, and providing a detection basis for life prediction of the silicone rubber material in a dry-type transformer.

Description

Silicone rubber material performance prediction method and device based on LSTM neural network Technical Field

The present application relates to the technical field of performance prediction of insulating materials for power equipment, and in particular to a method and device for predicting the performance of silicone rubber materials based on an LSTM neural network.

Background technique

The service life of dry-type transformers depends largely on the aging characteristics of the transformer's insulation materials, while silicone rubber materials can improve partial discharge defects and have strong heat dissipation capabilities. They also have the advantages of high strength, strong overload and short-circuit resistance, and silicone rubber materials can be recycled to avoid solid waste pollution. Liquid silicone rubber is used for vacuum casting or vacuum impregnation of windings. The silicone rubber that penetrates between turns and layers solidifies the winding into an integral structure. The silicone rubber encapsulation layer formed by casting or impregnation on the inner and outer surfaces of the winding can effectively improve the insulation performance of the dry-type transformer.

However, there is relatively little research on aging modeling and life prediction of silicone rubber materials under the condition of solid insulation of dry-type transformer windings. Therefore, a performance prediction method for silicone rubber materials under the condition of solid insulation of dry-type transformer windings is urgently needed.

Summary of the invention

The present application provides a method and device for predicting the performance of silicone rubber materials based on an LSTM neural network, so as to realize the prediction of the aging performance of silicone rubber materials and provide a detection basis for the life prediction of silicone rubber materials in dry-type transformers.

In order to solve the above technical problems, in the first aspect, the present application provides a method for predicting the properties of silicone rubber materials based on an LSTM neural network, comprising:

Obtain dielectric properties data of silicone rubber materials at multiple aging stages;

The dielectric performance data is fused and processed to obtain a comprehensive performance index sequence of silicone rubber materials. The comprehensive performance index sequence can characterize the evolution of the electrical properties of silicone rubber materials during the entire aging process. The entire aging process includes multiple aging stages.

According to the comprehensive performance index sequence, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method until the long short-term memory neural network reaches the preset convergence condition, and the performance prediction model of silicone rubber material is obtained;

The silicone rubber material performance prediction model is used to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of the comprehensive performance index of the silicone rubber material to be predicted.

In some implementations, the dielectric performance data includes volume resistivity, relative dielectric constant and dielectric loss value. The dielectric performance data is fused to obtain a comprehensive performance index sequence of the silicone rubber material, including:

The sample index matrix is constructed by taking volume resistivity, relative dielectric constant at the target frequency point and dielectric loss value as performance indicators;

The principal component analysis method is used to fuse the dielectric performance data according to the sample index matrix to generate a comprehensive performance index sequence.

In some implementations, the principal component analysis method is used to fuse the dielectric performance data according to the sample index matrix to generate a comprehensive performance index sequence, including:

Using principal component analysis, according to the sample indicator matrix, the correlation coefficient matrix between each performance indicator is calculated, and the eigenvalue of the correlation coefficient matrix is calculated;

Determine the target eigenvector corresponding to the target eigenvalue when the component contribution rate condition is satisfied;

According to the target feature vector, the dielectric performance data is fused and processed to obtain a comprehensive performance index sequence.

In some implementations, the comprehensive performance indicator sequence is:

a ₁ = [a _R , a _ei , a _ti ] ^T ;

Among them, Z represents the comprehensive performance index sequence, _a1 is the target feature vector, R _Vol is the standardized sequence of volume resistivity, Eps _i is the standardized sequence of relative dielectric constant at the i-th frequency point, and Tanδ _i is the standardized sequence of dielectric loss value at the i-th frequency point.

In some implementations, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method according to the comprehensive performance index sequence until the long short-term memory neural network reaches a preset convergence condition, thereby obtaining a silicone rubber material performance prediction model, including:

Divide the comprehensive performance indicator sequence into a training set and a test set;

Using the training set, the long short-term memory neural network is trained to obtain a trained target long short-term memory neural network;

Using the discrete differential evolution method, the network parameters of the target long short-term memory neural network are optimized according to the test set until the target long short-term memory neural network reaches the preset convergence conditions, and the silicone rubber material performance prediction model is obtained. The network parameters include the time window length, the number of hidden layer neurons and the number of hidden layer layers.

In some implementations, the discrete differential evolution method is used to optimize the network parameters of the target long short-term memory neural network according to the test set until the target long short-term memory neural network reaches a preset convergence condition, thereby obtaining a silicone rubber material performance prediction model, including:

Generate the initial population based on the time window length, the number of hidden layer neurons and the number of hidden layer layers;

According to the test set, the target LSTM neural network is trained with the initial population, and the individual fitness is calculated by the root mean square error;

Perform mutation and crossover operations on the initial population to obtain the latest population;

According to the test set, the target LSTM neural network is trained with the latest population, and the latest individual fitness is calculated with the root mean square error;

If the latest individual fitness is not greater than the individual fitness, the latest population is used as the new initial population to perform the next round of mutation and crossover operations until the termination condition is met and the optimal individual is output;

The target long short-term memory neural network is updated with the optimal individual to obtain a silicone rubber material performance prediction model.

In some implementations, a silicone rubber material performance prediction model is used to predict target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to an initial sequence of comprehensive performance indexes of the silicone rubber material to be predicted, including:

Taking the initial sequence of comprehensive performance indicators as the model input of the silicone rubber material performance prediction model, predicting the target comprehensive performance indicator data of the silicone rubber material to be predicted at the next aging stage;

Inserting the target comprehensive performance index data into the initial sequence of comprehensive performance indicators to obtain a new initial sequence of comprehensive performance indicators;

Using the new initial sequence of comprehensive performance indicators, continue to predict the target comprehensive performance indicator data of the next aging stage until the target comprehensive performance indicator data is lower than the preset threshold, and predict the remaining life of the silicone rubber material to be predicted based on the number of iterations experienced when the target comprehensive performance indicator data is lower than the preset threshold.

In a second aspect, the present application provides a silicone rubber material performance prediction device based on an LSTM neural network, comprising:

An acquisition module, used for acquiring dielectric property data of silicone rubber materials at multiple aging stages;

A fusion module is used to fuse the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material. The comprehensive performance index sequence can characterize the evolution of the electrical properties of the silicone rubber material during the entire aging process. The entire aging process includes multiple aging stages.

An optimization module is used to optimize the network parameters of the long short-term memory neural network by discrete differential evolution method according to the comprehensive performance index sequence until the long short-term memory neural network reaches a preset convergence condition to obtain a silicone rubber material performance prediction model;

The prediction module is used to use the silicone rubber material performance prediction model to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of the comprehensive performance index of the silicone rubber material to be predicted.

In a third aspect, the present application provides a computer device, including a processor and a memory, the memory being used to store a computer program, and when the computer program is executed by the processor, the method for predicting the performance of silicone rubber materials based on an LSTM neural network as in the first aspect is implemented.

In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the silicone rubber material performance prediction method based on the LSTM neural network as in the first aspect.

Compared with the prior art, this application has at least the following beneficial effects:

By obtaining the dielectric performance data of silicone rubber materials at multiple aging stages, the dielectric performance data are fused and processed. A comprehensive performance index sequence of silicone rubber materials is obtained, and an aging model of the long short-term memory neural network is built by using the comprehensive performance index sequence that can characterize the evolution of the electrical properties of the silicone rubber material throughout the aging process, so that the aging model can learn the electrical performance characteristics of the silicone rubber material at each aging stage; according to the comprehensive performance index sequence, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method until the long short-term memory neural network reaches the preset convergence conditions, and a silicone rubber material performance prediction model (aging model) is obtained, which effectively avoids the problems of insufficient fitting and low prediction accuracy caused by selecting key parameters through experience, and improves the model accuracy; finally, the silicone rubber material performance prediction model is used to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of the comprehensive performance index of the silicone rubber material to be predicted, thereby realizing the prediction of the aging performance of the silicone rubber material and providing a detection basis for the life prediction of the silicone rubber material in the dry-type transformer.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic flow chart of a method for predicting properties of silicone rubber materials based on an LSTM neural network according to an embodiment of the present application;

FIG2 is a schematic diagram of the structure of an LSTM neural network shown in an embodiment of the present application;

FIG3 is a flow chart of a DDE algorithm according to an embodiment of the present application;

FIG. 4 is a schematic diagram showing the division result of the training set and the test set according to the comprehensive performance index sequence in the embodiment of the present application.

FIG5 is a schematic diagram of prediction results of a silicone rubber material to be predicted shown in an embodiment of the present application;

FIG6 is a schematic diagram of the structure of a silicone rubber material performance prediction device based on an LSTM neural network shown in an embodiment of the present application

FIG. 7 is a schematic diagram of the structure of a computer device according to an embodiment of the present application.

Detailed ways

The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

Please refer to Figure 1, which is a schematic flow chart of a method for predicting the performance of a silicone rubber material based on an LSTM neural network provided in an embodiment of the present application. The method for predicting the performance of a silicone rubber material based on an LSTM neural network in an embodiment of the present application can be applied to computer devices, including but not limited to smart phones, laptops, tablet computers, desktop computers, physical servers, and cloud servers. As shown in Figure 1, the method for predicting the performance of a silicone rubber material based on an LSTM neural network in this embodiment includes steps S101 to S105, which are described in detail as follows:

Step S101, obtaining dielectric property data of the silicone rubber material at multiple aging stages.

In this step, the entire aging process includes multiple aging stages. The different aging stages of this embodiment can be regarded as the aging process. The dielectric properties data include but are not limited to volume resistivity, relative dielectric constant and dielectric loss value.

Optionally, the silicone rubber material is processed into a thin sheet sample, the surface of the sample is cleaned, the cleaned sample is placed in a glass container, and placed in a vacuum drying oven for drying, and then a heat aging test is performed on the sample. Part of the sample is taken out at specific intervals and preserved in a vacuum, and the volume resistivity, relative dielectric constant and dielectric loss value of the taken sample at the target frequency are measured.

Exemplarily, the silicone rubber material is processed into a square thin sheet sample with a thickness of 1 mm and a length and width of 40× ⁴⁰ mm2. The surface of the sample is cleaned with anhydrous ethanol, the cleaned sample is placed in a glass container, and the glass container is placed in a vacuum drying oven at 120°C and 50 Pa for drying for 2 hours; then the dried sample is subjected to a thermal aging test: the ambient temperature is set to 250°C, the maximum aging time is set to 125 days, part of the sample is taken out every 12 hours and preserved in a vacuum, a total of n=250 groups of test data under different aging conditions, the test data include volume resistivity, relative dielectric constant and dielectric loss value at the target frequency point in the frequency band of ^10-1 Hz to ¹⁰⁶ Hz.

Step S102, fusing the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material, wherein the comprehensive performance index sequence can characterize the evolution of the electrical performance of the silicone rubber material during the entire aging process, and the entire aging process includes multiple aging stages.

In this step, the comprehensive performance index sequence is a sequence obtained by combining the dielectric performance data of multiple aged dielectrics in order of aging time. Optionally, the fusion processing includes data dimensionality reduction and data fusion, and the comprehensive performance index sequence is generated by reducing the dimensionality of the dielectric performance data of multiple aged dielectrics and then fusing the reduced dimensionality dielectric performance data.

Step S103, according to the comprehensive performance index sequence, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method until the long short-term memory neural network reaches a preset convergence condition, thereby obtaining a silicone rubber material performance prediction model.

In this step, Discrete Differential Evolution (DDE) is an algorithm constructed after rounding off the optimization process of Standard Differential Evolution (DE). It has the characteristics of DE algorithm with few controlled parameters, strong robustness and fast convergence speed.

Optionally, the comprehensive performance index sequence is decomposed into a training set and a test set, and a long short-term memory (LSTM) neural network is used for training. The prediction error on the test set is used as the target, and the network parameters of the LSTM neural network are optimized using the DDE algorithm. The network parameters include the time window length, the number of hidden layer neurons, and the number of hidden layer layers. The finally optimized LSTM neural network is used as a silicone rubber material performance prediction model, that is, a silicone rubber material aging model. It should be noted that compared with the current setting of network parameters based on user experience, this embodiment determines the network parameters through the DDE algorithm, which can improve the model fit and model prediction accuracy.

Step S104, using the silicone rubber material performance prediction model, according to the initial sequence of comprehensive performance indicators of the silicone rubber material to be predicted, predicting target comprehensive performance indicator data of the silicone rubber material to be predicted at each aging stage.

In this step, the dielectric properties of a preset number of samples of the silicone rubber material to be predicted are measured at the current stage. The electrical performance data are fused and processed into an initial sequence of comprehensive performance indicators, which are used as the model input of the silicone rubber material performance prediction model to predict the target comprehensive performance indicator data of the silicone rubber material to be predicted in the next aging stage. The target comprehensive performance indicator data are used to update the initial sequence of comprehensive performance indicators, and the target comprehensive performance indicator data of the next aging stage are continuously predicted. By analogy, the target comprehensive performance indicator data of the silicone rubber material to be predicted in each aging stage are predicted.

In some embodiments, the step S102 includes:

Taking the volume resistivity, the relative dielectric constant at the target frequency point and the dielectric loss value as performance indicators, a sample indicator matrix is constructed;

The dielectric performance data is fused and processed according to the sample indicator matrix by using the principal component analysis method to generate the comprehensive performance indicator sequence.

In this embodiment, for the silicone rubber material sampled for the i-th time, its volume resistivity R _Vol , relative dielectric constant Eps at the target frequency point and dielectric loss value Tanδ are measured, with a total of p performance indicators, and a sample indicator matrix x = {x _ij }, i = 1, 2, ..., p, j = 1, 2, ..., p is constructed. Then, the principal component analysis method is used to reduce the dimension and fuse the dielectric performance data according to the sample indicator matrix to generate a comprehensive performance indicator sequence.

Optionally, the using of principal component analysis to fuse the dielectric performance data according to the sample indicator matrix to generate the comprehensive performance indicator sequence includes:

Using principal component analysis, according to the sample indicator matrix, calculate the correlation coefficient matrix between the various performance indicators, and calculate the eigenvalues of the correlation coefficient matrix;

Determine the target eigenvector corresponding to the target eigenvalue when the component contribution rate condition is satisfied;

The dielectric performance data is fused according to the target feature vector to obtain the comprehensive performance index sequence.

In this optional embodiment, illustratively, the sample indicator matrix is standardized:

in, is the mean of the column data in the sample indicator matrix, _Sj is the standard deviation of the column data in the sample indicator matrix;

According to the standardized sample index matrix X, the correlation coefficient matrix R = { _rij } between each performance index _Xij is calculated, and the eigenvalues ( _λ1 , _λ2 , ..., _λp ) of the correlation coefficient matrix R and the corresponding eigenvectors _a1 , ..., _ap are calculated.

Calculate the mth component sequence F _m = _Xam , m=1,2,…,p, and the contribution rate _bm of the component sequence:

The component contribution rate condition is that if the maximum value b ₁ >0.85, it is determined that the first principal component sequence F ₁ basically retains the main information of the original performance index, and the sequence Z=F ₁ is used as the comprehensive performance index sequence.

Optionally, the comprehensive performance indicator sequence is:

a ₁ = [a _R , a _ei , a _ti ] ^T ;

Among them, Z represents the comprehensive performance index sequence, _a1 is the target eigenvector corresponding to the maximum eigenvalue of the correlation coefficient matrix R, R _Vol is the standardized sequence of volume resistivity, Eps _i is the standardized sequence of relative dielectric constant at the i-th frequency point, and Tanδ _i is the standardized sequence of dielectric loss value at the i-th frequency point.

In some embodiments, the step S103 includes:

Dividing the comprehensive performance indicator sequence into a training set and a test set;

Using the training set, training the long short-term memory neural network to obtain a trained target long short-term memory neural network;

The discrete differential evolution method is used to optimize the network parameters of the target long short-term memory neural network according to the test set until the target long short-term memory neural network reaches the preset convergence condition, thereby obtaining the silicone rubber material performance prediction model, wherein the network parameters include the time window length, the number of hidden layer neurons and the number of hidden layer layers.

Exemplarily, the network structure of the LSTM neural network is shown in Figure 2. The LSTM network includes four gate structures, each of which is a neural network containing num_layer hidden layers, and each hidden layer contains num_hidden neurons and is fully connected to the input vector [h(t-1), x(t)]. h(t-1) is the output of the previous moment and is a vector of dimension num_hidden. x(t) is the input of the current moment, which is composed of the previous L outputs, and L is the length of the time window.

In this embodiment, the comprehensive performance index sequence is optionally used to form (NL) group samples, 80% of which are used as training sets for network training, and 20% are used as test sets for testing. The division result is shown in Figure 4. During the test, the output of the previous moment is added to the comprehensive performance index sequence, the sliding time window forms new L inputs, and the root mean square error (RMSE) of the neural network is defined as:

Where T is the number of samples in the test set, _Zi is the i-th comprehensive performance evaluation index in the test set, and f( _xi ) is the predicted value of the LSTM neural network for the i-th sample input _xi .

Then, based on the discrete differential evolution method combined with the root mean square error, the target LSTM neural network trained based on the training set is optimized.

Optionally, the discrete differential evolution method is used to perform the target long short-term memory neural network according to the test set. The network parameters of the network are optimized until the target long short-term memory neural network reaches the preset convergence condition, and the silicone rubber material performance prediction model is obtained, including:

Generate an initial population based on the time window length, the number of neurons in the hidden layer, and the number of hidden layers;

According to the test set, the target long short-term memory neural network is trained with the initial population, and the individual fitness is calculated with the root mean square error;

Performing mutation and crossover operations on the initial population to obtain a newest population;

According to the test set, the target long short-term memory neural network is trained with the latest population, and the latest individual fitness is calculated with the root mean square error;

If the latest individual fitness is not greater than the individual fitness, the latest population is used as the new initial population to perform the next round of mutation and crossover operations until the termination condition is met and the optimal individual is output;

The target long short-term memory neural network is updated with the optimal individual to obtain the silicone rubber material performance prediction model.

In this optional embodiment, the network parameters that determine the prediction effect of the LSTM neural network include the number of hidden layers num_layer, the number of hidden layer neurons num_hidden and the time window length L. In order to ensure the prediction effect of the model, the three network parameters of the LSTM neural network are optimized using the DDE algorithm with RMSE as the optimization target.

Exemplarily, as shown in FIG3 , num_hidden, num_layer, L range are set, and the maximum number of iterations t _max , the population size N _p , the mutation operator F ₀ and the crossover operator CR are set;

Generate an initial population, use the initial population to train the LSTM neural network, and calculate the individual fitness RSME(X(i)) based on the above root mean square error formula;

The mutation operation is:
V = X(1) + floor{F ₀ (X(2) - X(3))};

Among them, X(1), X(2) and X(3) are three different individuals in population X, and floor means rounding down;

The crossover operation is: replace the individuals in population X with the individuals in the mutant population V using the crossover operator CR, and process the out-of-bounds individuals in V using the boundary absorption method to obtain the latest population V(i);

Use the latest population V(i) to train the LSTM neural network, and calculate the latest individual fitness RSME(V(i)) based on the above root mean square error formula;

If RSME(V(i))≤RSME(X(i)), the latest population V(i) is used as the new initial population X(i) for the next mutation, crossover and selection operations until the termination condition is met and the optimal individual is output.

The termination condition is RSME(V(i))≤preset value, or the number of iterations reaches t _max . The optimal individual determined after DDE optimization, that is, the parameters of the LSTM model are num_layer＝2, num_hidden＝8, and L＝3.

In some embodiments, the step S104 includes:

The initial sequence of comprehensive performance indicators is used as the model input of the silicone rubber material performance prediction model to predict the The target comprehensive performance index data of the silicone rubber material to be predicted at the next aging stage;

Inserting the target comprehensive performance indicator data into the comprehensive performance indicator initial sequence to obtain a new comprehensive performance indicator initial sequence;

Using a new initial sequence of comprehensive performance indicators, continue to predict the target comprehensive performance indicator data of the next aging stage until the target comprehensive performance indicator data is lower than the preset threshold, and predict the remaining life of the silicone rubber material to be predicted based on the number of iterations experienced when the target comprehensive performance indicator data is lower than the preset threshold.

In this embodiment, illustratively, three consecutive volume resistivities, dielectric constants, and dielectric losses of the silicone rubber material to be predicted are obtained, and these three dielectric performance data are used as the initial values of the initial sequence Z of the comprehensive performance index, and the threshold Zend of the comprehensive performance index is calculated according to the dielectric performance data when the material fails. The LSTM neural network predicts the target comprehensive performance index data at the next moment based on the three initial values of the initial sequence Z of the comprehensive performance index, and inserts the target comprehensive performance index data into the last one of the original sequence Z, deletes the first one, keeps the sample size of the original sequence unchanged, realizes the dynamic update of the comprehensive performance index sequence, and estimates the remaining life of the silicone rubber material to be predicted according to the number of iterations experienced when the comprehensive performance index is lower than the threshold Zend.

For example, the prediction result is shown in FIG5 , from which it can be seen that after 36 iterations, the index performance of the silicone rubber material to be predicted is lower than the threshold Zend, so the remaining life of the silicone rubber material to be predicted is 36×12 hours=432 hours.

In order to execute the silicone rubber material performance prediction method based on LSTM neural network corresponding to the above method embodiment, to achieve the corresponding functions and technical effects. Referring to FIG. 7, FIG. 7 shows a block diagram of a silicone rubber material performance prediction device based on LSTM neural network provided in an embodiment of the present application. For ease of explanation, only the parts related to this embodiment are shown. The silicone rubber material performance prediction device based on LSTM neural network provided in an embodiment of the present application includes:

An acquisition module 601 is used to acquire dielectric property data of the silicone rubber material at multiple aging stages;

A fusion module 602 is used to fuse the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material, wherein the comprehensive performance index sequence can characterize the evolution of the electrical performance of the silicone rubber material during the entire aging process, wherein the entire aging process includes multiple aging stages;

An optimization module 603 is used to optimize the network parameters of the long short-term memory neural network by discrete differential evolution method according to the comprehensive performance index sequence until the long short-term memory neural network reaches a preset convergence condition to obtain a silicone rubber material performance prediction model;

The prediction module 604 is used to use the silicone rubber material performance prediction model to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of comprehensive performance indexes of the silicone rubber material to be predicted.

In some embodiments, the dielectric property data includes volume resistivity, relative dielectric constant and dielectric loss value, and the fusion module 602 includes:

A construction unit, configured to construct a sample index matrix by taking the volume resistivity, the relative dielectric constant at a target frequency point and the dielectric loss value as performance indicators;

The processing unit is used to use the principal component analysis method to fuse the dielectric performance data according to the sample indicator matrix to generate the comprehensive performance indicator sequence.

In some embodiments, the processing unit is specifically configured to:

Using principal component analysis, according to the sample indicator matrix, calculate the correlation coefficient matrix between the various performance indicators, and calculate the eigenvalues of the correlation coefficient matrix;

Determine the target eigenvector corresponding to the target eigenvalue when the component contribution rate condition is satisfied;

According to the target feature vector, the dielectric performance data is fused to obtain the comprehensive performance index sequence.

In some embodiments, the comprehensive performance indicator sequence is:

a ₁ = [a _R , a _ei , a _ti ] ^T ;

Among them, Z represents the comprehensive performance index sequence, _a1 is the target feature vector, R _Vol is the standardized sequence of volume resistivity, Eps _i is the standardized sequence of relative dielectric constant at the i-th frequency point, and Tanδ _i is the standardized sequence of dielectric loss value at the i-th frequency point.

In some embodiments, the optimization module 603 includes:

A division unit, used for dividing the comprehensive performance indicator sequence into a training set and a test set;

A training unit, used to train the long short-term memory neural network using the training set to obtain a trained target long short-term memory neural network;

An optimization unit is used to optimize the network parameters of the target long short-term memory neural network according to the test set by using the discrete differential evolution method until the target long short-term memory neural network reaches the preset convergence condition, thereby obtaining the silicone rubber material performance prediction model, wherein the network parameters include the time window length, the number of hidden layer neurons and the number of hidden layer layers.

In some embodiments, the optimization unit is specifically used to:

Generate an initial population based on the time window length, the number of neurons in the hidden layer, and the number of hidden layers;

According to the test set, the target long short-term memory neural network is trained with the initial population, and the individual fitness is calculated with the root mean square error;

Performing mutation and crossover operations on the initial population to obtain a newest population;

According to the test set, the target long short-term memory neural network is trained with the latest population, and the latest individual fitness is calculated with the root mean square error;

If the latest individual fitness is not greater than the individual fitness, the latest population is used as the new initial population to perform the next round of mutation and crossover operations until the termination condition is met and the optimal individual is output;

The target long short-term memory neural network is updated with the optimal individual to obtain the silicone rubber material performance prediction model.

In some embodiments, the prediction module 604 is specifically configured to:

Using the initial sequence of comprehensive performance indicators as the model input of the silicone rubber material performance prediction model, predicting the target comprehensive performance indicator data of the silicone rubber material to be predicted at the next aging stage;

Inserting the target comprehensive performance indicator data into the comprehensive performance indicator initial sequence to obtain a new comprehensive performance indicator initial sequence;

Using a new initial sequence of comprehensive performance indicators, continue to predict the target comprehensive performance indicator data of the next aging stage until the target comprehensive performance indicator data is lower than the preset threshold, and predict the remaining life of the silicone rubber material to be predicted based on the number of iterations experienced when the target comprehensive performance indicator data is lower than the preset threshold.

The above-mentioned silicone rubber material performance prediction device based on LSTM neural network can implement the silicone rubber material performance prediction method based on LSTM neural network of the above-mentioned method embodiment. The options in the above-mentioned method embodiment are also applicable to this embodiment and will not be described in detail here. The rest of the contents of the embodiment of this application can refer to the contents of the above-mentioned method embodiment, and will not be repeated in this embodiment.

FIG7 is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. As shown in FIG7 , the computer device 7 of this embodiment includes: at least one processor 70 (only one is shown in FIG7 ), a memory 71, and a computer program 72 stored in the memory 71 and executable on the at least one processor 70, and when the processor 70 executes the computer program 72, the steps in any of the above method embodiments are implemented.

The computer device 7 may be a computing device such as a smart phone, a tablet computer, a desktop computer, and a cloud server. The computer device may include, but is not limited to, a processor 70 and a memory 71. Those skilled in the art will appreciate that FIG. 7 is merely an example of the computer device 7 and does not constitute a limitation on the computer device 7. The computer device 7 may include more or fewer components than shown in the figure, or may combine certain components, or different components, and may also include, for example, input and output devices, network access devices, etc.

The processor 70 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

In some embodiments, the memory 71 may be an internal storage unit of the computer device 7, such as a hard disk or memory of the computer device 7. In other embodiments, the memory 71 may also be an external storage device of the computer device 7, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 7. It may also include both the internal storage unit and the external storage device of the computer device 7. The memory 71 is used to store an operating system, an application program, a boot loader, data, and other programs, such as the program code of the computer program, etc. The memory 71 may also be used to temporarily store data that has been output or is to be output.

In addition, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

An embodiment of the present application provides a computer program product. When the computer program product is run on a computer device, the computer device implements the steps in the above-mentioned method embodiments when executing the computer device.

In several embodiments provided in the present application, it is understood that each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.

If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims (10)

1. A method for predicting silicone rubber material properties based on LSTM neural network, characterized by comprising:

   Obtain dielectric properties data of silicone rubber materials at multiple aging stages;

   The dielectric performance data are fused to obtain a comprehensive performance index sequence of the silicone rubber material, wherein the comprehensive performance index sequence can characterize the evolution of the electrical performance of the silicone rubber material during the entire aging process, wherein the entire aging process includes multiple aging stages;

   According to the comprehensive performance index sequence, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method until the long short-term memory neural network reaches a preset convergence condition, thereby obtaining a silicone rubber material performance prediction model;

   The silicone rubber material performance prediction model is used to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of comprehensive performance indexes of the silicone rubber material to be predicted.
2. The method for predicting silicone rubber material performance based on LSTM neural network according to claim 1, characterized in that the dielectric performance data includes volume resistivity, relative dielectric constant and dielectric loss value, and the fusion processing of the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material includes:

   Taking the volume resistivity, the relative dielectric constant at the target frequency point and the dielectric loss value as performance indicators, a sample indicator matrix is constructed;

   The dielectric performance data is fused and processed according to the sample indicator matrix by using the principal component analysis method to generate the comprehensive performance indicator sequence.
3. The method for predicting the performance of silicone rubber materials based on an LSTM neural network according to claim 2 is characterized in that the principal component analysis method is used to fuse the dielectric performance data according to the sample indicator matrix to generate the comprehensive performance indicator sequence, including:

   Using principal component analysis, according to the sample indicator matrix, calculate the correlation coefficient matrix between the various performance indicators, and calculate the eigenvalue of the correlation coefficient matrix;

   Determine the target eigenvector corresponding to the target eigenvalue when the component contribution rate condition is satisfied;

   The dielectric performance data is fused according to the target feature vector to obtain the comprehensive performance index sequence.
4. The method for predicting silicone rubber material properties based on LSTM neural network according to claim 3, characterized in that the comprehensive performance index sequence is:
   

   a ₁ = [a _R , a _ei , a _ti ] ^T ;

   Among them, Z represents the comprehensive performance index sequence, _a1 is the target feature vector, R _Vol is the standardized sequence of volume resistivity, Eps _i is the standardized sequence of relative dielectric constant at the i-th frequency point, and Tanδ _i is the standardized sequence of dielectric loss value at the i-th frequency point.
5. The method for predicting the performance of a silicone rubber material based on an LSTM neural network according to claim 1 is characterized in that, according to the comprehensive performance index sequence, the network parameters of the long short-term memory neural network are optimized by discrete differential evolution method until the long short-term memory neural network reaches a preset convergence condition, and a silicone rubber material performance prediction model is obtained, comprising:

   Dividing the comprehensive performance indicator sequence into a training set and a test set;

   Using the training set, training the long short-term memory neural network to obtain a trained target long short-term memory neural network;

   The discrete differential evolution method is used to optimize the network parameters of the target long short-term memory neural network according to the test set until the target long short-term memory neural network reaches the preset convergence condition, thereby obtaining the silicone rubber material performance prediction model, wherein the network parameters include the time window length, the number of hidden layer neurons and the number of hidden layer layers.
6. The method for predicting the performance of a silicone rubber material based on an LSTM neural network according to claim 5, characterized in that the discrete differential evolution method is used to optimize the network parameters of the target long short-term memory neural network according to the test set until the target long short-term memory neural network reaches the preset convergence condition, thereby obtaining the silicone rubber material performance prediction model, comprising:

   Generate an initial population based on the time window length, the number of neurons in the hidden layer, and the number of hidden layers;

   According to the test set, the target long short-term memory neural network is trained with the initial population, and the individual fitness is calculated with the root mean square error;

   Performing mutation and crossover operations on the initial population to obtain a newest population;

   According to the test set, the target long short-term memory neural network is trained with the latest population, and the latest individual fitness is calculated with the root mean square error;

   If the latest individual fitness is not greater than the individual fitness, the latest population is used as the new initial population to perform the next round of mutation and crossover operations until the termination condition is met and the optimal individual is output;

   The target long short-term memory neural network is updated with the optimal individual to obtain the silicone rubber material performance prediction model.
7. The method for predicting the performance of a silicone rubber material based on an LSTM neural network according to claim 1, characterized in that the method uses the silicone rubber material performance prediction model to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of the comprehensive performance index of the silicone rubber material to be predicted, including:

   Using the initial sequence of comprehensive performance indicators as the model input of the silicone rubber material performance prediction model, predicting the target comprehensive performance indicator data of the silicone rubber material to be predicted at the next aging stage;

   Inserting the target comprehensive performance indicator data into the comprehensive performance indicator initial sequence to obtain a new comprehensive performance indicator initial sequence;

   Using a new initial sequence of comprehensive performance indicators, continue to predict the target comprehensive performance indicator data of the next aging stage until the target comprehensive performance indicator data is lower than the preset threshold, and predict the remaining life of the silicone rubber material to be predicted based on the number of iterations experienced when the target comprehensive performance indicator data is lower than the preset threshold.
8. A silicone rubber material performance prediction device based on LSTM neural network, characterized by comprising:

   An acquisition module, used for acquiring dielectric property data of silicone rubber materials at multiple aging stages;

   A fusion module, used for fusing the dielectric performance data to obtain a comprehensive performance index sequence of the silicone rubber material, wherein the comprehensive performance index sequence can characterize the evolution of the electrical performance of the silicone rubber material during the entire aging process, wherein the entire aging process includes multiple aging stages;

   An optimization module, used to optimize the network parameters of the long short-term memory neural network by discrete differential evolution method according to the comprehensive performance index sequence, until the long short-term memory neural network reaches a preset convergence condition, thereby obtaining a silicone rubber material performance prediction model;

   The prediction module is used to use the silicone rubber material performance prediction model to predict the target comprehensive performance index data of the silicone rubber material to be predicted at each aging stage according to the initial sequence of comprehensive performance indicators of the silicone rubber material to be predicted.
9. A computer device, characterized in that it includes a processor and a memory, wherein the memory is used to store a computer program, and when the computer program is executed by the processor, the method for predicting the performance of silicone rubber materials based on an LSTM neural network as described in any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that it stores a computer program, and when the computer program is executed by a processor, it implements the silicone rubber material performance prediction method based on the LSTM neural network as described in any one of claims 1 to 7.