Error change prediction method based on super-capacity energy storage

By combining random forests, support vector machines and LSTM neural networks, combined with deep learning and wavelet transformation, a hybrid prediction model is constructed, which solves the problem of limited performance of a single model in complex data scenarios, and achieves high quality in supercapacity energy storage systems Prediction of accuracy error change.

CN118917442AActive Publication Date: 2024-11-08XIAN THERMAL POWER RES INST CO LTD +1

Patent Information

Application Number
CN202411419252.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-12
Publication Date
2024-11-08
Estimated Expiration
2044-10-12

AI Technical Summary

Technical Problem

A single prediction model has limited performance when processing complex and multi-dimensional data, making it difficult to cope with complex practical application scenarios, especially in supercapacity energy storage systems, which are difficult to achieve high-precision prediction of error changes.

Method used

By combining random forests, support vector machines and LSTM neural networks, weighted voting or stacked integration strategies, combined with deep learning and wavelet transformation and other technologies, a hybrid prediction model is built, multi-algorithm parallel training is carried out, and weight allocation is performed based on prediction accuracy is performed. , optimize the prediction results.

Benefits of technology

It improves the accuracy of error change prediction and the robustness of the model, reduces the limitations and instability of a single model, enhances the stability and adaptability of the prediction, and is suitable for complex data prediction scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917442A_ABST
    Figure CN118917442A_ABST
Patent Text Reader

Abstract

The invention discloses an error change prediction method based on super-capacity energy storage, and relates to the technical field of super-capacity energy storage, and the method comprises the steps: collecting data information, carrying out the intelligent fault detection through a deep learning model, monitoring the operation state of a super-capacity energy storage system, recognizing and marking abnormal data, and carrying out the correction and elimination of the marked abnormal data. Performing multi-scale analysis on the corrected data by adopting a wavelet transform method, extracting multi-level features of error change, and further extracting nonlinear features in the error change by adopting a nonlinear feature extraction technology; and constructing a hybrid prediction model, performing error change prediction, deploying a lightweight prediction model on edge equipment, performing real-time data processing and prediction, and uploading a result to a cloud. The method not only can effectively reduce the limitation and instability of a single model, but also can optimize the prediction result through an integration strategy, thereby improving the robustness and accuracy of the model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of supercapacity energy storage, and in particular to an error change prediction method based on supercapacity energy storage. Background Art

[0002] In the field of modern data analysis and prediction, as the scale and complexity of data continue to increase, a single prediction model often has difficulty coping with various complex practical application scenarios. Traditional prediction models, such as random forests, support vector machines (SVMs), and long short-term memory networks (LSTMs), each have their own unique advantages and application scenarios, but when dealing with complex, multi-dimensional data, the performance of a single model may be limited. For example, random forests perform better when dealing with high-dimensional data and nonlinear problems, SVMs have advantages when dealing with small samples and high-dimensional data, and LSTMs have a unique ability to capture long-term dependencies in time series. Therefore, how to effectively combine the advantages of multiple models to build an efficient and accurate hybrid prediction model has become an important direction of current research. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides an error change prediction method based on supercapacity energy storage, by combining random forest, support vector machine and LSTM neural network, and using weighted voting or stacking integration strategy, so as to achieve high-precision prediction of error change. In practical applications, a single model may lead to unsatisfactory prediction results due to the characteristics of the data or the limitations of the model itself. In order to solve this problem, the present invention trains multiple algorithms in parallel, and distributes weights based on the prediction accuracy of each model, integrates the advantages of multiple models, and improves the overall prediction performance. This method can not only effectively reduce the limitations and instability of a single model, but also optimize the prediction results through an integrated strategy, thereby improving the robustness and accuracy of the model.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions, a method for predicting error changes based on supercapacity energy storage, comprising: Collect data information, use deep learning models for intelligent fault detection, monitor the operating status of the ultra-capacity energy storage system, identify and mark abnormal data, and correct and eliminate the marked abnormal data; use wavelet transform method to perform multi-scale analysis on the corrected data, extract multi-level features of error changes, and further extract nonlinear features in error changes through nonlinear feature extraction technology; build a hybrid prediction model, combine random forest, support vector machine and LSTM neural network to predict error changes, deploy lightweight prediction models on edge devices, perform real-time data processing and prediction, and upload the results to the cloud.

[0006] As a preferred solution of the error change prediction method based on super-capacity energy storage described in the present invention, wherein: the collected data information includes collecting voltage data, current data, temperature data, power data, state data, environmental data, historical operation data, frequency data, capacity data, vibration and noise data; the state data includes the charging state and discharging state of the super-capacity energy storage unit; the environmental data includes the external environment data of the operating environment of the super-capacity energy storage system; the capacity data includes the remaining capacity and the number of charge and discharge cycles of the super-capacity energy storage unit.

[0007] As a preferred solution of the error change prediction method based on ultra-capacity energy storage described in the present invention, wherein: the use of a deep learning model for intelligent fault detection includes constructing a fault detection result function, representing the fault detection result at time t: ; in, Represents the original input data, representing the The convolution kernel is Layer Input signal; Represents the weight matrix, representing the The convolution kernel is The weight parameters of the layer are obtained through training data; Represents the convolution kernel parameters, representing the The convolution kernel is Layer parameters; represents the bias term, representing the The convolution kernel is The bias parameters of the layer are obtained through training data; Represents the activation function, and uses the ReLU activation function, which is defined as ; Represents the pooling result, representing the The convolution kernel is The pooling result of the layer is obtained through the maximum pooling or average pooling function; Q represents the total number of convolution kernels; L represents the total number of layers; R represents the total number of input signals. represents the convolution kernel index, is the layer index and r is the input signal index.

[0008] As a preferred solution of the error change prediction method based on supercapacitor energy storage described in the present invention, wherein: the correction and elimination of the marked abnormal data includes identifying and marking abnormal data points through intelligent fault detection of convolutional neural networks, and the abnormal data points refer to sensor failure, communication error or unexpected situation causing the data to deviate from the preset normal range; Perform statistical analysis on the collected historical data and calculate the statistical characteristics of each parameter, including mean, standard deviation, median, and quartile; Perform data correction on abnormal data points, where for a single abnormal data point, use the mean of historical data To replace: ; For multiple consecutive abnormal data points, use linear interpolation to perform linear interpolation based on adjacent normal data points: ; in, is the interpolated data, and are the normal data points before and after the abnormal data segment, respectively. is the total number of abnormal data points; is the position of the current abnormal data point in the abnormal data segment, from 1 to ; Use a sliding window to fill in the missing data, the window size is , the average value of the data points in the window is used as the filling value: ; in, represents the value of the hth data point in the sliding window, Indicates the value of the data point after correction and filling; Perform polynomial fitting on historical data and use the fitting function to fill in missing data.

[0009] , in, are the polynomial coefficients, For time; is the highest order of the polynomial, is the order of the polynomial.

[0010] As a preferred solution of the error change prediction method based on supercapacitor energy storage described in the present invention, wherein: the wavelet transform method includes decomposing the corrected data into multi-scale components through wavelet transform, each component represents a signal component of different frequencies, and the wavelet transform includes discrete wavelet transform and continuous wavelet transform; All corrected data are sorted into time series , Represents the corrected raw data time series, including voltage, current, temperature, and power; represents the wavelet basis function, is the mother wavelet, obtained by scaling and translating; ; in, Represents the scale index, which is used to control the scaling of the wavelet; Represents the translation index, which is used to control the position of the wavelet; represents the approximation coefficient of the scale component, is a scaling function, defined as: ; Use discrete wavelet transform DWT and continuous wavelet transform CWT to process data respectively; ; Among them, a is the scaling factor, b is the translation factor; the scale index and pan index determined through a wavelet transform process; Wavelet basis functions Select the mother wavelet, select the optimal mother wavelet according to the data characteristics; select the mother wavelet according to the data characteristics , and obtained by scaling and translating ; Approximation coefficient It is obtained by calculating the cumulative result of each scale component, and the formula is: ; For each scale and each translation Perform wavelet transform to obtain multi-scale components , The smaller it is, the smaller the signal component is on the corresponding scale; For time, is the scale index, is the translation index, is the maximum size, is the corrected original data time series, is the scale function, is the result of continuous wavelet transform.

[0011] As a preferred solution of the error change prediction method based on ultracapacitor energy storage described in the present invention, the multi-scale analysis includes analyzing the multi-scale components, extracting the features that affect the error change as the input of the prediction model, and the features that affect the error change include frequency components, amplitude and phase information; Select mother wavelet according to data characteristics , and obtain multi-scale components by scaling and translating , for each multi-scale component Calculate Amplitude and Phase ; Use historical data to calculate the standard deviation of each scale and translation position , standardize the frequency components; combine the standardized amplitude and phase information to calculate the comprehensive characteristic function : ; ; in, is the frequency, is the total number of translation positions, i 1 is an imaginary unit; The first Scale and The multi-scale components of the translation position are used as the frequency Function of Indicates Scale and The standard deviation of the translation position.

[0012] As a preferred solution of the error change prediction method based on ultracapacitor energy storage described in the present invention, wherein: the extraction of nonlinear features in the error change includes processing the input data by a kernel method or an autoencoder to extract nonlinear features in a high-dimensional space, and the kernel method includes a radial basis function kernel and a polynomial kernel; Enter the data Normalize, select the preset kernel width parameter, and calculate the radial basis function kernel ;Select preset constants and the degree of the polynomial , calculate the polynomial kernel , train the autoencoder through unsupervised learning methods to obtain the weight matrix and bias terms; Combine the radial basis function kernel, the polynomial kernel, and the output of the autoencoder to compute nonlinear feature maps in high-dimensional space The details are as follows: , in, is the training sample index, is the total number of training samples, is the kernel width parameter; is the degree of the polynomial, representing the input data x and the training sample The inner product of is the autoencoder neuron index, is the total number of autoencoder neurons, is the input data feature index, is the total number of input data features, is the weight of the radial basis function kernel, is the weight of the autoencoder, is the weight matrix of the autoencoder, is the bias term of the autoencoder, is the activation function.

[0013] As a preferred solution of the error change prediction method based on ultra-capacity energy storage described in the present invention, wherein: the construction of the hybrid prediction model includes combining random forest, support vector machine and LSTM neural network through an integrated learning method to form a hybrid prediction model, and the integrated learning method includes weighted voting and stacking integration strategy; Perform single model prediction for random forest, support vector machine and LSTM neural network, introduce weighted voting strategy, assign weights to the prediction results of each model, and calculate the mean absolute error MAE of each model based on the prediction of each model: ; Assign weights based on MAE, where the weights are the inverse of MAE: ; The prediction results of each model are weighted using a weighted voting method: ; Using a secondary learner Combine the prediction results of the single models and calculate the root mean square error RMSE of the secondary learner: ; Assign weights based on RMSE, where the weights are the inverse of RMSE: ; The final prediction result is obtained by combining the weighted voting results and the secondary learner results: , in, Represents the random forest model for the input data The prediction results; Represents the support vector machine model for input data The prediction results; Represents the long short-term memory network model for input data The prediction results; The final prediction result; is the mean absolute error of model f, w f is the weight of model f, is the prediction result of the secondary learner for the input data X, is the root mean square error of the secondary learner s, is the weight of the sub-learner s, F is the total number of models participating in weighted voting, D is the total number of sub-learners participating in sub-learner fusion, s represents the sub-learner index, B represents the total number of samples for calculating errors, Indicates The true value of samples is used to calculate the prediction error; Indicates the fth model for the The predicted value of a sample.

[0014] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of an error change prediction method based on supercapacitor energy storage are implemented.

[0015] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of an error change prediction method based on supercapacitor energy storage are implemented.

[0016] Beneficial effects of the present invention: The hybrid prediction model of the present invention effectively combines the advantages of random forests, support vector machines and LSTM neural networks through multi-algorithm parallel training and weighted voting or stacking integration strategies, thereby achieving high-precision error change prediction. Compared with the traditional single model, the present invention has significant improvements in prediction accuracy and model robustness. Specifically, by allocating weights based on model prediction errors, the model with smaller errors has a greater impact on the final result, thereby improving the accuracy of the prediction. At the same time, by integrating the prediction results of multiple models, the impact of outliers or abnormal performance of a single model on the final result can be effectively reduced, and the stability and robustness of the model are enhanced. In addition, the method of the present invention has good scalability and adaptability, can be applied to various complex data prediction scenarios, provides reliable prediction results, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 A schematic flow chart of an error change prediction method based on supercapacity energy storage provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0022] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0023] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0024] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0025] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and provides an error change prediction method based on supercapacitor energy storage, comprising: Collect data information, use deep learning models for intelligent fault detection, monitor the operating status of the ultra-capacity energy storage system, identify and mark abnormal data, and correct and eliminate the marked abnormal data; use wavelet transform method to perform multi-scale analysis on the corrected data, extract multi-level features of error changes, and further extract nonlinear features in error changes through nonlinear feature extraction technology; build a hybrid prediction model, combine random forest, support vector machine and LSTM neural network to predict error changes, deploy lightweight prediction models on edge devices, perform real-time data processing and prediction, and upload the results to the cloud.

[0026] The collected data information includes collecting voltage data, current data, temperature data, power data, status data, environmental data, historical operation data, frequency data, capacity data, vibration and noise data; the status data includes the charging status and discharging status of the super-capacity energy storage unit; the environmental data includes the external environment data of the operating environment of the super-capacity energy storage system; the capacity data includes the remaining capacity and the number of charge and discharge cycles of the super-capacity energy storage unit.

[0027] The use of a deep learning model for intelligent fault detection includes constructing a fault detection result function representing the fault detection result at time t: ; in, Represents the original input data, representing the The convolution kernel is Layer Input signal; Represents the weight matrix, representing the The convolution kernel is The weight parameters of the layer are obtained through training data; Represents the convolution kernel parameters, representing the The convolution kernel is Layer parameters; represents the bias term, representing the The convolution kernel is The bias parameters of the layer are obtained through training data; Represents the activation function, and uses the ReLU activation function, which is defined as ; Represents the pooling result, representing the The convolution kernel is The pooling result of the layer is obtained through the maximum pooling or average pooling function; Q represents the total number of convolution kernels; L represents the total number of layers; R represents the total number of input signals. represents the convolution kernel index, is the layer index and r is the input signal index.

[0028] The correcting and eliminating the marked abnormal data includes identifying and marking abnormal data points through intelligent fault detection of a convolutional neural network, wherein the abnormal data points refer to data deviating from a preset normal range due to sensor failure, communication error or unexpected situation; Perform statistical analysis on the collected historical data and calculate the statistical characteristics of each parameter, including mean, standard deviation, median, and quartile; Perform data correction on abnormal data points, where for a single abnormal data point, use the mean of historical data To replace: ; For multiple consecutive abnormal data points, use linear interpolation to perform linear interpolation based on adjacent normal data points: , in, is the interpolated data, and are the normal data points before and after the abnormal data segment, respectively. is the total number of abnormal data points; is the position of the current abnormal data point in the abnormal data segment, from 1 to .

[0029] Use sliding windows to fill in missing data, and the window size is , the average value of the data points in the window is used as the filling value: , in, represents the value of the hth data point in the sliding window, Indicates the value of the data point after correction and padding.

[0030] Perform polynomial fitting on historical data and use the fitting function to fill in missing data.

[0031] , in, are the polynomial coefficients, For time; is the highest order of the polynomial, is the order of the polynomial.

[0032] Specifically, in this embodiment, it is assumed that we need to fill the marked abnormal data to ensure the quality of the input data: Assume that there is a set of voltage data as follows, some of which are marked as abnormal: , Among them, 300 and 305 are data points marked as abnormal.

[0033] Select a sliding window size of 3 (that is, calculate the average of three data points each time): The first window (including the 1st, 2nd, and 3rd data points): , the specific calculation is: ; The second window (including the 2nd, 3rd, and 4th data points): , the specific calculation is: ; (Note that 300 is an outlier here, but due to the average calculation of the sliding window, the calculation result will be unreliable).

[0034] The third window (including the 3rd, 4th, and 5th data points): , the specific calculation is: ; The fourth window (including the 4th, 5th, and 6th data points): , the specific calculation is: ; The fifth window (including the 5th, 6th, and 7th data points): , the specific calculation is: .

[0035] Assume that a quadratic polynomial (i.e., second-order polynomial) is selected for fitting. The form of the quadratic polynomial is: ; Determine the fitting points and use the normal data points before and after the abnormal data points and the average value calculated by the sliding window method for fitting: ; Fit these points and assume that the corresponding time points are .

[0036] Based on the selected data points, construct the fitting equation system: ; Solve the system of equations to get the fitting coefficients : ; Use the fitted polynomial to calculate the correction value of the abnormal data point. Assume that the time point corresponding to the abnormal data point is and :

[0037] The corrected data sequence using the polynomial fitting method and sliding window method is: ; in, is the corrected data point value calculated by the sliding window method, To fit the polynomial function, t is the time point, a 0 is the constant term of the fitted polynomial, a 1 is the coefficient of the first-order term of the fitted polynomial, a 2 is the coefficient of the quadratic term of the fitting polynomial, X' is the corrected data point value calculated by the sliding window method, w is the size of the sliding window, 220, 221, 225, 223, 222 are the original data points, 220.7 and 221 are the corrected data point values ​​calculated by the fitting polynomial, and 248.67, 276.67, 276, 250 are the intermediate results calculated by the sliding window method.

[0038] The wavelet transform method includes decomposing the corrected data into multi-scale components by wavelet transform, each component represents a signal component of a different frequency, and the wavelet transform includes discrete wavelet transform and continuous wavelet transform; All corrected data are sorted into time series , Represents the corrected raw data time series, including voltage, current, temperature, and power; represents the wavelet basis function, is the mother wavelet, obtained by scaling and translating; , in, Represents the scale index, which is used to control the scaling of the wavelet; Represents the translation index, which is used to control the position of the wavelet; Represents the approximation coefficient of the scale component.

[0039] is a scaling function, defined as: ; Use discrete wavelet transform DWT and continuous wavelet transform CWT to process data respectively: , Among them, a is the scaling factor, b is the translation factor; the scale index and pan index Determined through a wavelet transform process.

[0040] Wavelet basis functions Select the mother wavelet, select the optimal mother wavelet according to the data characteristics; select the mother wavelet according to the data characteristics , and obtained by scaling and translating ; Approximation coefficient It is obtained by calculating the cumulative result of each scale component, and the formula is: For each scale and each translation Perform wavelet transform to obtain multi-scale components , The smaller it is, the smaller the signal component is on the corresponding scale; For time, is the scale index, is the translation index, is the maximum size, is the corrected original data time series, is the scale function, is the result of continuous wavelet transform.

[0041] The multi-scale analysis includes analyzing the multi-scale components to extract the features that affect the error change as the input of the prediction model, wherein the features that affect the error change include frequency components, amplitude and phase information; Select mother wavelet according to data characteristics , and obtain multi-scale components by scaling and translating , for each multi-scale component Calculate Amplitude and Phase ; Use historical data to calculate the standard deviation of each scale and translation position , standardize the frequency components; combine the standardized amplitude and phase information to calculate the comprehensive characteristic function : ; ; in, is the frequency, is the total number of translation positions, i 1 is an imaginary unit; The first Scale and The multi-scale components of the translation position are used as the frequency Function of Indicates Scale and The standard deviation of the translation position.

[0042] The extracting of nonlinear features in error variation includes processing input data by a kernel method or an autoencoder to extract nonlinear features in a high-dimensional space, wherein the kernel method includes a radial basis function kernel and a polynomial kernel; Enter the data Normalize, select the preset kernel width parameter, and calculate the radial basis function kernel ;Select preset constants and the degree of the polynomial , calculate the polynomial kernel , train the autoencoder through unsupervised learning methods to obtain the weight matrix and bias terms; Combine the radial basis function kernel, the polynomial kernel, and the output of the autoencoder to compute nonlinear feature maps in high-dimensional space The details are as follows: , in, is the training sample index, is the total number of training samples, is the kernel width parameter; is the degree of the polynomial, representing the input data x and the training sample The inner product of is the autoencoder neuron index, is the total number of autoencoder neurons, is the input data feature index, is the total number of input data features, is the weight of the radial basis function kernel, is the weight of the autoencoder, is the weight matrix of the autoencoder, is the bias term of the autoencoder, is the activation function.

[0043] The constructing of the hybrid prediction model includes combining a random forest, a support vector machine and an LSTM neural network through an integrated learning method to form a hybrid prediction model, wherein the integrated learning method includes a weighted voting and a stacking integration strategy; Perform single model prediction for random forest, support vector machine and LSTM neural network, introduce weighted voting strategy, assign weights to the prediction results of each model, and calculate the mean absolute error MAE of each model based on the prediction of each model: ; Assign weights based on MAE, where the weights are the inverse of MAE: ; The prediction results of each model are weighted using a weighted voting method: ; Using a secondary learner Combine the prediction results of the single models and calculate the root mean square error RMSE of the secondary learner: ; Assign weights based on RMSE, where the weights are the inverse of RMSE: ; The final prediction result is obtained by combining the weighted voting results and the secondary learner results: ; in, Represents the random forest model for the input data The prediction results; Represents the support vector machine model for input data The prediction results; Represents the long short-term memory network model for input data The prediction results; The final prediction result; is the mean absolute error of model f, w f is the weight of model f, is the prediction result of the secondary learner for the input data X, is the root mean square error of the secondary learner s, is the weight of the sub-learner s, F is the total number of models participating in weighted voting, D is the total number of sub-learners participating in sub-learner fusion, s represents the sub-learner index, B represents the total number of samples for calculating errors, Indicates The true value of samples is used to calculate the prediction error; Indicates the fth model for the The predicted value of a sample.

[0044] Embodiment 2, the second embodiment of the present invention, is different from the previous embodiment in that: If the functions are implemented in the form of software functional units 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 invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0045] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0048] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0049] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for predicting error changes based on supercapacity energy storage, characterized in that: include, Collect data information, use deep learning models for intelligent fault detection, monitor the operating status of the ultra-capacity energy storage system, identify and mark abnormal data, and correct and eliminate the marked abnormal data; The wavelet transform method is used to perform multi-scale analysis on the corrected data to extract the multi-level characteristics of the error change. The nonlinear characteristics in the error change are further extracted through the nonlinear feature extraction technology. Build a hybrid prediction model that combines random forest, support vector machine, and LSTM neural network to predict error changes, deploy lightweight prediction models on edge devices, perform real-time data processing and prediction, and upload the results to the cloud.

2. The error change prediction method based on supercapacity energy storage according to claim 1, characterized in that: The collected data information includes collecting voltage data, current data, temperature data, power data, status data, environmental data, historical operation data, frequency data, capacity data, vibration and noise data; the status data includes the charging status and discharging status of the super-capacity energy storage unit; the environmental data includes the external environment data of the operating environment of the super-capacity energy storage system; the capacity data includes the remaining capacity and the number of charge and discharge cycles of the super-capacity energy storage unit.

3. The error change prediction method based on supercapacity energy storage according to claim 2, characterized in that: The use of a deep learning model for intelligent fault detection includes constructing a fault detection result function representing the fault detection result at time t: , in, Represents the original input data, representing the The convolution kernel is Layer Input signal; Represents the weight matrix, representing the The convolution kernel is The weight parameters of the layer are obtained through training data; Represents the convolution kernel parameters, representing the The convolution kernel is Layer parameters; represents the bias term, representing the The convolution kernel is The bias parameters of the layer are obtained through training data; Represents the activation function, and uses the ReLU activation function, which is defined as ; Represents the pooling result, representing the The convolution kernel is The pooling result of the layer is obtained through the maximum pooling or average pooling function; Q represents the total number of convolution kernels; L represents the total number of layers; R represents the total number of input signals. represents the convolution kernel index, is the layer index and r is the input signal index.

4. The error change prediction method based on supercapacity energy storage according to claim 3, characterized in that: The correcting and eliminating the marked abnormal data includes identifying and marking abnormal data points through intelligent fault detection of a convolutional neural network, wherein the abnormal data points refer to data deviating from a preset normal range due to sensor failure, communication error or unexpected situation; Perform statistical analysis on the collected historical data and calculate the statistical characteristics of each parameter, including mean, standard deviation, median, and quartile; Perform data correction on abnormal data points, where for a single abnormal data point, use the mean of historical data To replace: ; To correct multiple consecutive abnormal data points, use linear interpolation method to perform linear interpolation based on adjacent normal data points: ; in, is the interpolated data, and are the normal data points before and after the abnormal data segment, respectively. is the total number of abnormal data points; is the position of the current abnormal data point in the abnormal data segment, from 1 to ; Use sliding windows to fill in missing data, and the window size is , the average value of the data points in the window is used as the filling value: ; in, represents the value of the hth data point in the sliding window, Indicates the value of the data point after correction and filling; Perform polynomial fitting on historical data and use the fitting function to fill in missing data; , in, are the polynomial coefficients, For time; is the highest order of the polynomial, is the order of the polynomial.

5. The error change prediction method based on supercapacity energy storage according to claim 4, characterized in that: The wavelet transform method includes decomposing the corrected data into multi-scale components by wavelet transform, each component represents a signal component of a different frequency, and the wavelet transform includes discrete wavelet transform and continuous wavelet transform; All corrected data are sorted into time series , Represents the corrected raw data time series, including voltage, current, temperature, and power; represents the wavelet basis function, is the mother wavelet, obtained by scaling and translating; ; in, Represents the scale index, which is used to control the scaling of the wavelet; Represents the translation index, which is used to control the position of the wavelet; represents the approximation coefficient of the scale component, is a scaling function, defined as: ; Use discrete wavelet transform DWT and continuous wavelet transform CWT to process data respectively; ; Among them, a is the scaling factor, b is the translation factor; the scale index and pan index determined through a wavelet transform process; Wavelet basis functions Select the mother wavelet, select the optimal mother wavelet according to the data characteristics; select the mother wavelet according to the data characteristics , and obtained by scaling and translating ; Approximation coefficient It is obtained by calculating the cumulative result of each scale component, and the formula is: For each scale and each translation Perform wavelet transform to obtain multi-scale components , The smaller it is, the smaller the signal component is on the corresponding scale; For time, is the scale index, is the translation index, is the maximum size, is the corrected original data time series, is the scale function, is the result of continuous wavelet transform.

6. The error change prediction method based on supercapacity energy storage according to claim 5, characterized in that: The multi-scale analysis includes analyzing the multi-scale components to extract the features that affect the error change as the input of the prediction model, wherein the features that affect the error change include frequency components, amplitude and phase information; Select mother wavelet according to data characteristics , and obtain multi-scale components by scaling and translating , for each multi-scale component Calculate Amplitude and Phase ; Use historical data to calculate the standard deviation of each scale and translation position , standardize the frequency components; combine the standardized amplitude and phase information to calculate the comprehensive characteristic function : ; ; in, is the frequency, is the total number of translation positions, i1 is the imaginary unit; The first Scale and The multi-scale components of the translation position are used as the frequency Function of Indicates Scale and The standard deviation of the translation position.

7. The error change prediction method based on supercapacity energy storage according to claim 6, characterized in that: The extracting of nonlinear features in error variation includes processing input data by a kernel method or an autoencoder to extract nonlinear features in a high-dimensional space, wherein the kernel method includes a radial basis function kernel and a polynomial kernel; Enter the data Normalize, select the preset kernel width parameter, and calculate the radial basis function kernel ;Select preset constants and the degree of the polynomial , calculate the polynomial kernel , train the autoencoder through unsupervised learning methods to obtain the weight matrix and bias terms; Combine the radial basis function kernel, the polynomial kernel, and the output of the autoencoder to compute nonlinear feature maps in high-dimensional space The details are as follows: , in, is the training sample index, is the total number of training samples, is the kernel width parameter; is the degree of the polynomial, representing the input data x and the training sample The inner product of is the autoencoder neuron index, is the total number of autoencoder neurons, is the input data feature index, is the total number of input data features, is the weight of the radial basis function kernel, is the weight of the autoencoder, is the weight matrix of the autoencoder, is the bias term of the autoencoder, is the activation function.

8. The error change prediction method based on supercapacity energy storage according to claim 7, characterized in that: The constructing of the hybrid prediction model includes combining a random forest, a support vector machine and an LSTM neural network through an integrated learning method to form a hybrid prediction model, wherein the integrated learning method includes a weighted voting and a stacking integration strategy; Perform single model prediction for random forest, support vector machine and LSTM neural network, introduce weighted voting strategy, assign weights to the prediction results of each model, and calculate the mean absolute error MAE of each model based on the prediction of each model: ; Assign weights based on MAE, where the weights are the inverse of MAE: ; The prediction results of each model are weighted using a weighted voting method: ; Using a secondary learner Combine the prediction results of the single models and calculate the root mean square error RMSE of the secondary learner: ; Assign weights based on RMSE, where the weights are the inverse of RMSE: ; The final prediction result is obtained by combining the weighted voting results and the secondary learner results: , in, Represents the random forest model for the input data The prediction results; Represents the support vector machine model for input data The prediction results; Represents the long short-term memory network model for input data The prediction results; The final prediction result; is the mean absolute error of model f, w f is the weight of model f, is the prediction result of the secondary learner for the input data X, is the root mean square error of the secondary learner s, is the weight of the sub-learner s, F is the total number of models participating in weighted voting, D is the total number of sub-learners participating in sub-learner fusion, s represents the sub-learner index, B represents the total number of samples for calculating errors, Indicates The true value of samples is used to calculate the prediction error; Indicates the fth model for the The predicted value of a sample.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of an error change prediction method based on ultra-capacity energy storage according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of an error change prediction method based on ultra-capacity energy storage according to any one of claims 1 to 8 are implemented.

Citation Information

Patent Citations

  • Daily water consumption prediction method based on big data

    CN111210093A

  • Industrial control sensor numerical value anomaly detection method and device

    CN112989710A

  • Joint optimization method and system based on combined prediction of peak shaving power supply equipment

    CN117522066A

  • System and method for automatically detecting vital signs for handheld PDA (Personal Digital Assistant)

    CN117850601A

  • Power grid fault power protection maintenance plan making method, device, equipment and medium

    CN118504939A

Cited By

  • Car tire building drum error analysis method and system

    CN119475062A

  • Verification method of full-performance automatic assembly line test result

    CN119537252A

  • Bidirectional correction method for abnormity of industrial time series data

    CN119646405A

  • Energy storage system fault prediction method and system based on multi-modal data fusion

    CN120493152A

  • Renewable energy consumption potential prediction method based on combined model and application

    CN120875685A