Electrical element magnetic core loss prediction method based on soft gating hybrid expert model

Abstract

The invention discloses an electrical element magnetic core loss prediction method based on a soft gating hybrid expert model, and relates to an electrical element magnetic core loss prediction method, which adaptively adjusts the expert weight, so that the model is more flexible and has strong adaptability when facing different types of data. Diversified expert model combinations can complement respective advantages, so that the overall robustness and generalization ability are improved, and the risk of overfitting is reduced; in addition, the modular design enables the model to be easy to expand and maintain, and the interpretability of the model can be enhanced by analyzing the output weight of the gating network; furthermore, automatic parameter adjustment and end-to-end training are carried out through an intelligent optimization algorithm, so that the performance of the model is further improved; and finally, the soft gating hybrid expert model can effectively process non-linear and complex relationships, the powerful non-linear modeling capability of the MLP is utilized to capture a complex mode in the data, and the calculation efficiency is improved through parallel calculation and dynamic resource allocation. The advantages enable the soft gating hybrid expert model to show higher accuracy and reliability when processing complex tasks such as magnetic core loss prediction and the like.

Description

Technical Field

[0001] The invention relates to a method for predicting the magnetic core loss of an electrical component, and in particular to a method for predicting the magnetic core loss of an electrical component based on a soft-gated hybrid expert model. Technical Background

[0002] With the steady progress of China's modernization, power electronics technology has developed rapidly as the core technology of power conversion. Power conversion or power conversion technology based on power electronics has become an important part of the modern power system, and has been widely used in large-scale intelligent equipment power supplies, electric vehicles, rail transportation and smart grids. With the advancement of semiconductor technology, power converters are developing towards high frequency, high power density and high performance. Magnetic components play a key role in power converters, and their losses account for a large proportion of the total power loss. Therefore, optimizing the performance of magnetic components is crucial to improving the efficiency of the entire system.

[0003] Magnetic components are usually composed of windings and cores, and the losses mainly include core loss and winding loss. For winding loss, commonly used methods include electromagnetic field finite element simulation, direct measurement under specific excitation waveform conditions, and measurement based on AC power. In contrast, core loss is the energy loss caused by magnetic materials under the action of alternating magnetic fields. It is affected by many factors such as operating frequency, flux density, excitation waveform, temperature and core material, making the calculation and modeling of core loss complex and challenging.

[0004] The existing core loss models can be roughly divided into two categories: one is the separation model, which decomposes the core loss into three parts: hysteresis loss, eddy current loss and residual loss, and estimates the total core loss by independently calculating these components; the other is the empirical calculation model, which is mainly based on a large amount of experimental data fitting, among which the most widely used is the method based on the Steinmetz Equation (SE). However, the traditional Steinmetz equation is only applicable to the case of sinusoidal wave excitation. In order to adapt to the needs of a wider range of excitation waveforms, Rudy Severns proposed a modified Steinmetz equation (MSE), which can handle the core loss problem under non-sinusoidal waveform conditions. In addition, Venkatachalam et al. also proposed an improved scheme that takes into account the flux density change rate and double peak characteristics, which further improves the prediction accuracy.

[0005] In recent years, with the development of big data and artificial intelligence technology, the trend of global industrial automation, intelligence and digitalization has become increasingly obvious. In this context, using advanced data analysis technology and machine learning algorithms to establish a more accurate data-driven core loss analysis and prediction model has become a research hotspot. This method not only helps to deeply understand the core loss mechanism, but also provides reliable prediction results under different operating conditions, thereby guiding the research and development and application of new high-efficiency magnetic materials and promoting the continuous progress of power electronics technology. By combining traditional physical models with modern data science and technology, researchers aim to develop a new generation of core loss assessment tools that can accurately reflect actual operating conditions and meet engineering design needs. . Summary of the invention

[0006] The purpose of the present invention is to propose a method for predicting the core loss of electrical components based on a soft-gated hybrid expert model. The present invention enhances the interpretability of the model by analyzing the output weights of the gating network; further improves the performance of the model by automatically adjusting parameters and end-to-end training through an intelligent optimization algorithm; and uses the powerful nonlinear modeling capabilities of MLP to capture complex patterns in data, and improves computing efficiency through parallel computing and dynamic resource allocation. These advantages enable the soft-gated hybrid expert model to show higher accuracy and reliability when dealing with complex tasks such as core loss prediction.

[0007] The objective of the present invention is achieved through the following technical solutions:

[0008] A method for predicting core loss of electrical components based on a soft-gated hybrid expert model comprises the following steps:

[0009] 1. Obtain the core loss dataset;

[0010] 2. Perform feature engineering on the data set, extract features, perform skewness correction on continuous features, and perform one-hot encoding on discrete features;

[0011] 3. Perform feature screening on the current features, use the method combining XGBoost and SHAP to evaluate the influence of each feature on the core loss, and then remove the features with low importance;

[0012] 4. Establish an expert model, select several basic models with good performance as experts, such as Extra-tree, Ridge regression, LASSO regression, CatBoost, LightGBM, etc., and establish their integrated model as one of the experts. These models should be complementary, that is, they can capture different characteristics and patterns of data in different aspects.

[0013] 5. Train the expert model. Use the dataset to independently train each selected expert model, and each model should tune the hyperparameters through an intelligent optimization algorithm to achieve the best performance.

[0014] 6. Design a gating network. A multilayer perceptron (MLP) is usually used as a gating network, which is responsible for calculating the weight of each expert model based on the input data. The output layer of the MLP usually has the same number of nodes as the number of experts, and the output values ​​of these nodes will be normalized (for example, through a softmax function) to obtain a set of effective weights. The gating network needs to be trained to learn how to assign appropriate weights to each expert for a given data sample.

[0015] 7. Combine the expert models and use the gating network to determine the weight of each expert model according to the input data. Then, the prediction results of each expert model are weighted averaged according to the weight given by the gating network to obtain the final prediction result.

[0016] According to the present invention, the magnetic core loss prediction method based on the soft-gated hybrid expert model may also have the following additional technical features:

[0017] According to an embodiment of the present invention, in S1, the variables affecting the core loss are: temperature, material type, 1024 sampling point values ​​in one cycle, excitation waveform, and frequency, wherein there are 4 types of materials, 3 types of excitation waveforms, namely sine wave, triangle wave, and trapezoidal wave, 1024 sampling points are obtained by sampling at the same interval within one cycle, and the temperature takes 4 values: 25, 50, 70, and 90, in degrees Celsius;

[0018] According to an embodiment of the present invention, in S2, feature extraction is performed on the existing data set, and statistical features, gradient features, gradient mutation points, spectrum features and partial features of PCA principal component analysis are extracted respectively.

[0019] According to one embodiment of the present invention, in S4, Extra-tree, Ridge regression, LASSO regression, CatBoost and LightGBM are used respectively;

[0020] Extra-Tree is an ensemble learning method that makes predictions by building multiple decision trees. For regression tasks, the predicted value is the average of all tree predictions, and the formula is:

[0021]

[0022] Where T is the number of trees, f t (x) is the prediction value of the tth tree for the input x.

[0023] The prediction formula of Ridge regression is the same as that of ordinary linear regression, except that an L2 regularization term is introduced during the training process to prevent overfitting. The formula is:

[0024]

[0025] Where β0 is the intercept term, β j is feature x j These coefficients are obtained by minimizing the objective function during the training process.

[0026] The prediction formula of LASSO regression is the same as that of ordinary linear regression, but its coefficient estimation is subject to L1 regularization constraints. The formula is:

[0027]

[0028] Here β j It is also obtained during the training process by minimizing the objective function containing the L1 regularization term.

[0029] CatBoost uses gradient boosted decision trees to make predictions. Each tree outputs a prediction value, and the final prediction is the accumulation of all tree predictions. The formula is:

[0030]

[0031] Where N is the number of trees, w_i is the weight of the ith tree, and f_i(x) is the prediction of the ith tree for input x.

[0032] LightGBM is also a gradient boosting decision tree model. Its prediction formula is similar to CatBoost, and the final prediction is obtained by accumulating the prediction values ​​of all trees.

[0033]

[0034] In LightGBM, f_i(x) and f_i(x,c) represent the predicted value of the i-th tree for input x and the probability or score of predicting category c, respectively.

[0035] There are also ensemble learning methods that use LightGBM as a base learner and other algorithms as sub-learners. Suppose we have M primary models h_1,h_2,…,h_M, each with a prediction value for sample x. We can combine these predictions into a vector z(x)

[0036] z(x)=[h1(x),h2(x),…,h M (x)]

[0037] This vector z(x) is the input feature of the secondary model. The secondary model can be any regression model, such as linear regression, logistic regression, random forest, etc.

[0038] The final prediction result is:

[0039]

[0040] According to one embodiment of the present invention, the single-objective algorithm used in S5 to optimize each model is Tree-structured Parzen Estimator (TPE), which is a hyperparameter optimization algorithm based on Bayesian optimization. It guides the search process by constructing two probability density functions, thereby effectively finding the optimal hyperparameter combination. The main idea of ​​the TPE algorithm is to divide the search space into "good" and "bad" parts, and try to sample new hyperparameter combinations in the "good" area.

[0041] The single target of TPE optimization is the root mean square error (RMSE), which is the square root of the average of the squares of the differences between the predicted value and the actual value. The formula is as follows:

[0042]

[0043] n is the number of samples. y_i is the actual observed value of the i-th sample. \hat{y}_i is the predicted value of the i-th sample.

[0044] According to one embodiment of the present invention, in S7, we use a gating network to dynamically assign weights to each expert model, thereby achieving optimal prediction of the input data. The gating network is usually a multilayer perceptron (MLP), whose input is the same as the expert model, the number of nodes in the output layer is equal to the number of expert models, and the output of each node is normalized to a weight value between 0 and 1 through a Softmax function, indicating the importance of the corresponding expert model under the current input. The gating network is trained together with the expert model, and the parameters are optimized by minimizing the overall prediction error to ensure that the gating network can learn how to reasonably assign weights to the expert model for different types of input data.

[0045] In the prediction phase, for each new input, the gating network first calculates the weight of each expert model, then applies these weights to the prediction results of each expert model, and obtains the final prediction output through weighted average. This process not only improves the flexibility of the model, but also enhances the robustness and generalization ability of the system, because different expert models can play their respective advantages under different input modes. In this way, the hybrid expert model can adapt to various complex data distributions more accurately and provide more reliable prediction results.

[0046] The advantages and positive effects of the present invention are:

[0047] Compared with the existing integrated stacking technology, the technology of the present invention can adaptively adjust the expert weights, making the model more flexible and adaptable when facing different types of data; secondly, the combination of diversified expert models can complement each other's advantages, improve the overall robustness and generalization ability, and reduce the risk of overfitting; in addition, the modular design makes the model easy to expand and maintain, and the interpretability of the model can be enhanced by analyzing the output weights of the gating network; furthermore, the performance of the model is further improved through automatic parameter adjustment and end-to-end training through intelligent optimization algorithms; finally, the soft-gated hybrid expert model can effectively handle nonlinear and complex relationships, use the powerful nonlinear modeling capabilities of MLP to capture complex patterns in data, and improve computing efficiency through parallel computing and dynamic resource allocation. These advantages enable the soft-gated hybrid expert model to show higher accuracy and reliability when dealing with complex tasks such as core loss prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a step diagram of the technical solution of the present invention;

[0049] Figure 2 It is a comparative diagram of the characteristic evaluation of the present invention;

[0050] Figure 3 It is a structural diagram of the soft-gated hybrid expert model of the present invention; DETAILED DESCRIPTION

[0051] The following is a further detailed description of the specific implementation of the present invention through the description of the embodiments with reference to the accompanying drawings, with the purpose of helping those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention and to facilitate its implementation. In order to make the technical solution of the present invention clearer, the present invention is explained through the following embodiments:

[0052] A method for predicting magnetic core loss based on a soft-gated hybrid expert model. Figure 1 , including the following steps:

[0053] 1. Obtain the core loss dataset;

[0054] 2. Perform feature engineering on the data set, extract features, perform skewness correction on continuous features, and perform one-hot encoding on discrete features;

[0055] 3. Perform feature screening on the current features, use the method combining XGBoost and SHAP to evaluate the influence of each feature on the core loss, and then remove the features with low importance. It can be seen that Figure 2 ;

[0056] 4. Establish an expert model, select several basic models with good performance as experts, such as Extra-tree, Ridge regression, LASSO regression, CatBoost, LightGBM, etc., and establish their integrated model as one of the experts. These models should be complementary, that is, they can capture different characteristics and patterns of data in different aspects.

[0057] 5. Train the expert model. Use the dataset to independently train each selected expert model, and each model should tune the hyperparameters through an intelligent optimization algorithm to achieve the best performance.

[0058] 6. Design a gating network. A multilayer perceptron (MLP) is usually used as a gating network, which is responsible for calculating the weight of each expert model based on the input data. The output layer of the MLP usually has the same number of nodes as the number of experts, and the output values ​​of these nodes will be normalized (for example, through a softmax function) to obtain a set of effective weights. The gating network needs to be trained to learn how to assign appropriate weights to each expert for a given data sample.

[0059] 7. Combine the expert models and use the gating network to determine the weight of each expert model according to the input data. Then, the prediction results of each expert model are weighted averaged according to the weight given by the gating network to obtain the final prediction result.

[0060] According to the present invention, the magnetic core loss prediction method based on the soft-gated hybrid expert model may also have the following additional technical features:

[0061] According to an embodiment of the present invention, in S1, the variables affecting the core loss are: temperature, material type, 1024 sampling point values ​​in one cycle, excitation waveform, and frequency, wherein there are 4 types of materials, 3 types of excitation waveforms, namely sine wave, triangle wave, and trapezoidal wave, 1024 sampling points are obtained by sampling at the same interval within one cycle, and the temperature takes 4 values: 25, 50, 70, and 90, in degrees Celsius;

[0062] According to an embodiment of the present invention, in S2, feature extraction is performed on the existing data set, and statistical features, gradient features, gradient mutation points, spectrum features and partial features of PCA principal component analysis are extracted respectively.

[0063] According to one embodiment of the present invention, in S4, Extra-tree, Ridge regression, LASSO regression, CatBoost and LightGBM are used respectively;

[0064] Extra-Tree is an ensemble learning method that makes predictions by building multiple decision trees. For regression tasks, the predicted value is the average of all tree predictions, and the formula is:

[0065]

[0066] Where T is the number of trees, f t (x) is the prediction value of the tth tree for the input x.

[0067] The prediction formula of Ridge regression is the same as that of ordinary linear regression, except that an L2 regularization term is introduced during the training process to prevent overfitting. The formula is:

[0068]

[0069] Where β0 is the intercept term, β j is feature x j These coefficients are obtained by minimizing the objective function during the training process.

[0070] The prediction formula of LASSO regression is the same as that of ordinary linear regression, but its coefficient estimation is subject to L1 regularization constraints. The formula is:

[0071]

[0072] Here β j It is also obtained during the training process by minimizing the objective function containing the L1 regularization term.

[0073] CatBoost uses gradient boosted decision trees to make predictions. Each tree outputs a prediction value, and the final prediction is the accumulation of all tree predictions. The formula is:

[0074]

[0075] Where N is the number of trees, w_i is the weight of the ith tree, and f_i(x) is the prediction of the ith tree for input x.

[0076] LightGBM is also a gradient boosting decision tree model. Its prediction formula is similar to CatBoost, and the final prediction is obtained by accumulating the prediction values ​​of all trees.

[0077]

[0078] In LightGBM, f_i(x) and f_i(x,c) represent the predicted value of the i-th tree for input x and the probability or score of predicting category c, respectively.

[0079] There are also ensemble learning methods that use LightGBM as a base learner and other algorithms as sub-learners. Suppose we have M primary models h_1,h_2,…,h_M, each with a prediction value for sample x. We can combine these predictions into a vector z(x)

[0080] z(x)=[h1(x),h2(x),…,h M (x)]

[0081] This vector z(x) is the input feature of the secondary model. The secondary model can be any regression model, such as linear regression, logistic regression, random forest, etc.

[0082] The final prediction result is:

[0083]

[0084] According to one embodiment of the present invention, the single-objective algorithm used in S5 to optimize each model is Tree-structured Parzen Estimator (TPE), which is a hyperparameter optimization algorithm based on Bayesian optimization. It guides the search process by constructing two probability density functions, thereby effectively finding the optimal hyperparameter combination. The main idea of ​​the TPE algorithm is to divide the search space into "good" and "bad" parts, and try to sample new hyperparameter combinations in the "good" area.

[0085] The single target of TPE optimization is the root mean square error (RMSE), which is the square root of the average of the squares of the differences between the predicted value and the actual value. The formula is as follows:

[0086]

[0087] n is the number of samples. y_i is the actual observed value of the ith sample. \widehat{y_i} is the predicted value of the ith sample.

[0088] According to one embodiment of the present invention, S7 utilizes a gating network to dynamically assign weights to each expert model, thereby achieving optimal prediction of the input data. The gating network is usually a multilayer perceptron (MLP), whose input is the same as the expert model, the number of nodes in the output layer is equal to the number of expert models, and the output of each node is normalized to a weight value between 0 and 1 through a Softmax function, indicating the importance of the corresponding expert model under the current input. The gating network is trained together with the expert model, and the parameters are optimized by minimizing the overall prediction error to ensure that the gating network can learn how to reasonably assign weights to the expert models for different types of input data. The specific overall logic can be seen in Figure 3 .

[0089] In the prediction phase, for each new input, the gating network first calculates the weight of each expert model, then applies these weights to the prediction results of each expert model, and obtains the final prediction output through weighted average. This process not only improves the flexibility of the model, but also enhances the robustness and generalization ability of the system, because different expert models can play their respective advantages under different input modes. In this way, the hybrid expert model can adapt to various complex data distributions more accurately and provide more reliable prediction results.

[0090] Although the present invention is described in detail by referring to the accompanying drawings and in combination with the preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, a person of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and these modifications or substitutions shall be within the scope of the present invention. Any person of ordinary skill in the art may easily think of changes or substitutions within the technical scope disclosed by the present invention, and these shall be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A method for predicting core loss of electrical components based on a soft-gated hybrid expert model, characterized in that: The method obtains the prediction results of each expert model by weighting according to the weights provided by the gating network, and includes the following steps: 1) Obtain a core loss dataset; 2) Perform feature engineering on the data set to extract features, perform skewness correction on continuous features, and perform one-hot encoding on discrete features; 3) Perform feature screening on the current features, use the method combining XGBoost and SHAP to evaluate the influence of each feature on the core loss, and then remove the features with low importance; 4) Establish an expert model and select several basic models with good performance as experts, such as Extra-tree, Ridge regression, LASSO regression, CatBoost, and LightGBM. In addition, establish their integrated model as one of the experts; these models should be complementary, that is, they can capture different characteristics and patterns of data in different aspects; 5) Train the expert model. Use the dataset to independently train each selected expert model, and each model should tune the hyperparameters through an intelligent optimization algorithm to achieve the best performance. 6) Design a gating network, usually a multilayer perceptron (MLP) is used as the gating network, which is responsible for calculating the weight of each expert model based on the input data; the output layer of the MLP usually has the same number of nodes as the number of experts, and the output values ​​of these nodes will be normalized (for example, through the softmax function) to obtain a set of effective weights; the gating network needs to be trained to learn how to assign appropriate weights to each expert for a given data sample; 7) Combine the expert models and use the gating network to determine the weight of each expert model according to the input data; then, perform weighted averaging on the prediction results of each expert model according to the weight given by the gating network to obtain the final prediction result.

2. The method for predicting core loss of electrical components based on a soft-gated hybrid expert model according to claim 1, characterized in that: In step 1, the variables affecting the core loss are: temperature, material type, 1024 sampling point values ​​within a cycle, excitation waveform, and frequency. There are 4 types of materials and 3 types of excitation waveforms, namely sine wave, triangle wave, and trapezoidal wave. The 1024 sampling points are obtained by sampling at the same interval within a cycle time. The temperature takes 4 values: 25, 50, 70, and 90, in degrees Celsius.

3. The method for predicting core loss of electrical components based on a soft-gated hybrid expert model according to claim 1, characterized in that: In the step 2, feature extraction is performed on the existing data set, and statistical features, gradient features, gradient mutation points, spectrum features and partial features of PCA principal component analysis are extracted respectively.

4. The method for predicting core loss of electrical components based on a soft-gated hybrid expert model according to claim 1, characterized in that: In step 4, Extra-tree, Ridge regression, LASSO regression, CatBoost and LightGBM are used respectively; Extra-Tree is an ensemble learning method that makes predictions by building multiple decision trees; For regression tasks, the predicted value is the average of all tree predictions, using the formula: Where T is the number of trees, f t (x) is the prediction value of the tth tree for input x; The prediction formula of Ridge regression is the same as that of ordinary linear regression, except that the L2 regularization term is introduced during the training process to prevent overfitting; The formula is: Where β0 is the intercept term, β j is feature x j The coefficients of , which are obtained by minimizing the objective function during the training process; The prediction formula of LASSO regression is also the same as that of ordinary linear regression, but its coefficient estimation is constrained by L1 regularization; The formula is: Here β j It is also obtained by minimizing the objective function containing the L1 regularization term during the training process; CatBoost uses gradient boosted decision trees to make predictions; each tree outputs a prediction value, and the final prediction is the sum of all tree predictions. The formula is: Where N is the number of trees, w_i is the weight of the i-th tree, and f_i(x) is the prediction value of the i-th tree for input x; LightGBM is also a gradient boosting decision tree model. Its prediction formula is similar to CatBoost, and the final prediction is obtained by accumulating the prediction values ​​of all trees. In LightGBM, f_i(x) and f_i(x,c) represent the predicted value of the i-th tree for input x and the probability or score of predicting category c, respectively; There is also an ensemble learning method, using LightGBM as the base learner and other algorithms as sub-learners; suppose we have M primary models h_1,h_2,...,h_M, each model has a prediction value for sample x; we can combine these predictions into a vector z(x) z(x)=[h1(x),h2(x),…,h M (x)] This vector z(x) is the input feature of the secondary model; the secondary model can be any regression model, such as linear regression, logistic regression, random forest; The final prediction result is:

5. The method for predicting core loss of electrical components based on a soft-gated hybrid expert model according to claim 1, characterized in that: In step 5, the single-objective algorithm used to optimize each model is Tree-structured Parzen Estimator (TPE), which is a hyperparameter optimization algorithm based on Bayesian optimization. Two probability density functions are constructed to guide the search process, so as to effectively find the optimal hyperparameter combination. The idea of ​​the TPE algorithm is to divide the search space into "good" and "bad" parts, and try to sample new hyperparameter combinations in the "good" area. The single target of TPE optimization is the root mean square error (RMSE), which is the square root of the average of the squares of the differences between the predicted value and the actual value. The formula is as follows: n is the number of samples; y_i is the actual observed value of the i-th sample; \hat{y}_i is the predicted value of the i-th sample.

6. The method for predicting core loss of electrical components based on a soft-gated hybrid expert model according to claim 1, characterized in that: In step 7, a gating network is used to dynamically assign weights to each expert model, thereby achieving optimal prediction of the input data; the gating network is usually a multi-layer perceptron (MLP), whose input is the same as the expert model, the number of nodes in the output layer is equal to the number of expert models, and the output of each node is normalized to a weight value between 0 and 1 through a Softmax function, indicating the importance of the corresponding expert model under the current input; the gating network is trained together with the expert model, and the parameters are optimized by minimizing the overall prediction error, so as to ensure that the gating network can learn how to reasonably assign weights to the expert model for different types of input data; In the prediction stage, for each new input, the gating network first calculates the weight of each expert model, and then applies these weights to the prediction results of each expert model, and obtains the final prediction output through weighted average; this process not only improves the flexibility of the model, but also enhances the robustness and generalization ability of the system, because different expert models play their respective advantages under different input modes; in this way, the hybrid expert model can adapt to various complex data distributions more accurately and provide more reliable prediction results.

Images