Automatic electrical equipment calibration system based on artificial intelligence
By using deep learning models for feature extraction and calibration in the automatic calibration system of electrical equipment, the problem that traditional algorithms are difficult to capture nonlinear relationships is solved, and more efficient and reliable automatic calibration of electrical equipment is achieved.
Patent Information
- Application Number
- CN202411469757.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-21
- Publication Date
- 2025-02-07
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional machine learning algorithms are difficult to effectively capture complex nonlinear relationships in input data in electrical equipment metrology calibration, resulting in insufficient calibration accuracy and affecting the performance and reliability of the equipment.
The automatic calibration system of electrical equipment based on deep learning is adopted, and the feature extraction and calibration feedback modules are used to use deep learning models to perform feature extraction and calibration to improve calibration accuracy and reliability.
Through deep learning models, high-dimensional and low-dimensional features can be extracted and feature fusion can be performed, nonlinear relationships and key information in the operating state of electrical equipment can be more accurately identified, calibration accuracy and reliability can be improved, and more efficient automatic calibration can be achieved.
Smart Images
Figure CN119397236A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence, and in particular to an automatic calibration system for electrical equipment based on artificial intelligence. Background Art
[0002] With the rapid development of artificial intelligence technology in recent years, machine learning algorithms have been increasingly used in the field of electrical equipment calibration. However, traditional machine learning algorithms usually rely on simple mathematical models and manually designed features for feature extraction, and often have difficulty effectively capturing key information when faced with complex nonlinear relationships in input data. This limitation leads to insufficient calibration accuracy, affecting the performance and reliability of electrical equipment calibration. Summary of the invention
[0003] Based on this, the purpose of the present invention is to provide an automatic calibration system for electrical equipment based on artificial intelligence, which uses a deep learning model to calibrate electrical equipment to improve calibration accuracy and reliability.
[0004] In a first aspect, the present invention provides an automatic calibration system for electrical equipment based on artificial intelligence, the system comprising:
[0005] A data acquisition module is used to obtain raw data of electrical equipment. The raw data is used to characterize the operating status of the electrical equipment. The raw data includes current, voltage, and temperature.
[0006] The feature extraction module is used to extract features from the original data based on the deep learning model to obtain high-dimensional features and low-dimensional features. The high-dimensional features are used to describe the nonlinear relationship in the original data, and the low-dimensional features are used to describe the key information of the original data.
[0007] Feature fusion module, used to fuse high-dimensional features and low-dimensional features to obtain comprehensive features;
[0008] The calibration feedback module is used to detect the comprehensive characteristics based on the calibration model and output the calibration results. The calibration results are used to indicate whether the electrical equipment is calibrated properly.
[0009] Further, the feature fusion module includes an autoencoder, and the autoencoder includes an encoder and a decoder;
[0010] The encoder is used to perform feature mapping on high-dimensional features and low-dimensional features to obtain potential features;
[0011] The decoder is used to decode and calculate the potential features to obtain comprehensive features.
[0012] Furthermore, a model training module is included, which is used to:
[0013] Get the data set, which includes a training set and a test set;
[0014] Divide the training set into multiple subsets, which are used for cross-validation and model training;
[0015] Initialize the parameters of the LightGBM algorithm, including the learning rate, the number of trees, and the depth of the tree;
[0016] Use subsets for cross-validation to obtain the optimal model parameters;
[0017] The calibration model is trained according to the training set and the optimal model parameters to obtain a trained calibration model.
[0018] Furthermore, it also includes a model evaluation module, which includes an indicator calculation submodule and a confusion matrix analysis submodule;
[0019] The indicator calculation submodule is used to calculate the indicators of the calibration model using the test set, including accuracy, recall, and mean square error;
[0020] The confusion matrix analysis submodule is used to generate a confusion matrix based on the use of the test set, and the confusion matrix is used to visualize the calibration results of the calibration model.
[0021] Furthermore, a model updating module is included, which is used to:
[0022] Adjusting the parameters of the calibration model based on the online learning algorithm to obtain an updated calibration model;
[0023] The updated calibration model is evaluated through the model evaluation module to obtain the updated model performance;
[0024] When the updated model performance meets the preset performance index, the updated calibration model is updated to the calibration feedback module.
[0025] Furthermore, it also includes a visualization module for converting the influence data of different characteristics of the original data on the performance of the calibration model and the calibration results into visualization parameters; the visualization parameters are used to display the influence data and the calibration results in a display medium.
[0026] Furthermore, based on the calibration model, the comprehensive features are detected and the calibration results are output, including:
[0027] Based on the calibration model, the comprehensive features are detected to obtain calibration values;
[0028] Compare the calibration value with the preset threshold and output the calibration result.
[0029] In a second aspect, the present invention further provides an automatic calibration method for electrical equipment based on artificial intelligence, the method comprising:
[0030] Obtaining raw data of electrical equipment, which is used to characterize the operating status of electrical equipment, including current, voltage, and temperature;
[0031] Based on the deep learning model, the original data is feature extracted to obtain high-dimensional features and low-dimensional features. The high-dimensional features are used to describe the nonlinear relationship in the original data, and the low-dimensional features are used to describe the key information of the original data.
[0032] Fuse high-dimensional features and low-dimensional features to obtain comprehensive features;
[0033] Based on the calibration model, the comprehensive features are detected and the calibration results are output. The calibration results are used to indicate whether the electrical equipment is calibrated properly.
[0034] In a third aspect, the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of an automatic calibration method for electrical equipment based on artificial intelligence are implemented.
[0035] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of an automatic calibration method for electrical equipment based on artificial intelligence.
[0036] In the above technical solution, firstly, by acquiring the original data such as current, voltage and temperature, the operating status of the electrical equipment can be comprehensively and real-time characterized. Secondly, the deep learning model is used to extract features to obtain high-dimensional features and low-dimensional features, and feature fusion is performed to obtain comprehensive features, which can more accurately identify the nonlinear relationship and key information of the original data, and improve the accuracy and efficiency of feature extraction. Finally, the calibration model is used to detect and output the calibration results based on the comprehensive features, making the calibration process more efficient and automated. Compared with traditional machine learning algorithms, the calibration accuracy and reliability are further improved.
[0037] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 A schematic diagram of the structure of an automatic calibration system for electrical equipment based on artificial intelligence is provided in one embodiment of the present invention;
[0039] Figure 2 A flow chart of an automatic calibration method for electrical equipment based on artificial intelligence is provided in one embodiment of the present invention. DETAILED DESCRIPTION
[0040] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0041] In one embodiment, Figure 1 As shown, an automatic calibration system 100 for electrical equipment based on artificial intelligence is provided. In the embodiment of the present invention, the system is applied to a controller of an electrical equipment for example. It can be understood that the controller can also be a control terminal or a control server connected to the electrical equipment, which is not limited here. The system includes:
[0042] The data acquisition module 101 is used to obtain raw data of the electrical equipment. The raw data is used to characterize the operating state of the electrical equipment. The raw data includes current, voltage, and temperature.
[0043] The above raw data can be acquired through a sensor or a data acquisition system connected to the electrical equipment, and then the operating status of the electrical equipment can be known. Usually, the electrical equipment generally includes a voltmeter, an ammeter, a voltage source, a current source, etc. The raw data may include the voltage, current, and temperature of the electrical equipment to achieve calibration of the electrical equipment. After the raw data is acquired, it can be preprocessed for subsequent research and analysis. Data preprocessing may include data cleaning such as normalization, standardization, filtering, etc. to remove outliers, noise, and erroneous data, and then convert the processed data into a format suitable for analysis for subsequent calibration analysis.
[0044] The feature extraction module 102 is used to extract features from the original data based on the deep learning model to obtain high-dimensional features and low-dimensional features. The high-dimensional features are used to describe the nonlinear relationship in the original data, and the low-dimensional features are used to describe the key information of the original data.
[0045] Specifically, the raw data acquired in the data acquisition module 101 is processed by a deep learning model in the feature extraction module 102. The deep learning model is an artificial neural network that processes data by simulating the processing method of the human brain. It is composed of multiple layers of neurons, which are organized in layers to form a complex network structure. Each layer processes the input raw data to a certain extent and passes the processing results to the next layer until the last layer generates an output. The deep learning model can be a convolutional neural network (CNN), a recurrent neural network (RNN), or a deep neural network (DNN). Compared with traditional machine learning algorithms, this model is good at automatically learning and extracting features from raw data, and can capture complex patterns and relationships in the data, including nonlinear relationships.
[0046] Exemplarily, the deep learning model can be a DNN, including: an input layer for receiving raw data input; a hidden layer for feature learning, the number of layers can be adjusted as needed, each hidden layer contains multiple neurons, each neuron performs a nonlinear transformation on the input signal through an activation function, such as ReLU, Sigmoid or Tanh; an output layer: outputs the final feature representation or decision result according to the task type (classification, regression, etc.). Each hidden layer gradually extracts high-dimensional features from the input data by adjusting weights and biases. At each layer, DNN can capture the complex patterns and nonlinear relationships of the input data, that is, high-dimensional features, by learning the extracted features, which can be used to describe the nonlinear relationships of the original data. This process can be achieved through a back-propagation algorithm. In each iteration, DNN adjusts weights and biases according to the value of the loss function to reduce prediction errors.
[0047] Furthermore, after extracting high-dimensional features, the high-dimensional features can be reduced to low-dimensional features through dimensionality reduction techniques such as principal component analysis. Taking the principal component analysis method as an example, first, the mean of the high-dimensional feature vector is adjusted to zero, and its covariance matrix is calculated through the high-dimensional feature vector to reflect the relationship between features; secondly, the covariance matrix is decomposed by eigenvalue to obtain eigenvalues and corresponding eigenvectors, and then the principal components are selected according to the eigenvalues. These principal components constitute a low-dimensional feature space; finally, the high-dimensional feature vector is projected onto the selected principal component to form a low-dimensional feature vector, and then the low-dimensional feature is obtained.
[0048] The feature fusion module 103 is used to fuse the high-dimensional features and the low-dimensional features to obtain comprehensive features.
[0049] Specifically, the high-dimensional features and low-dimensional features from the feature extraction module 102 can be fused through a feature fusion method to obtain a comprehensive feature, which contains the complex information of the high-dimensional features and the key information of the low-dimensional features, can provide richer original data information, enhance the performance and applicability of the model, and enhance the performance of the model by combining different types of features, especially in tasks such as classification and regression. Exemplarily, the feature fusion method can achieve fusion by simply connecting the high-dimensional features and the low-dimensional features in the feature dimension, or by giving the weights of the high-dimensional features and the low-dimensional features, calculating their weighted sum for fusion, etc.
[0050] The calibration feedback module 104 is used to detect the comprehensive characteristics based on the calibration model and output the calibration result, and the calibration result is used to indicate whether the electrical device is calibrated properly.
[0051] The calibration feedback module 104 receives the output from the feature fusion module 103, that is, the comprehensive features after feature fusion processing. These features contain key and nonlinear information about the operating status of the electrical equipment. The comprehensive features are analyzed and detected using a pre-trained calibration model. Exemplarily, the verification model can be based on a deep learning or machine learning algorithm, which can evaluate different dimensions of the comprehensive features and compare them with set standards such as thresholds, distribution models, etc., to determine the current state of the device, output the calibration results, and realize automatic calibration of the electrical equipment.
[0052] The system obtains the raw data of electrical equipment, extracts high-dimensional and low-dimensional features based on the deep learning model, and performs feature fusion to generate comprehensive features. These comprehensive features not only contain the nonlinear relationships and key information in the original data, but also make full use of the advantages of different features to improve the accuracy of the model in classification and regression tasks. This method not only improves the performance of the model, reduces the dimension and computational complexity, but also enhances the feature representation ability and real-time response ability. By verifying the comprehensive features of the calibration model, the calibration results can be output to realize the automatic calibration of electrical equipment. In addition, the verification of comprehensive features helps to confirm the validity of the extracted features, thereby optimizing the features, simplifying the model structure and improving its interpretability, ensuring that the performance of the model on new data will not decrease significantly, and further enhancing the generalization ability of the model.
[0053] In one embodiment, the feature fusion module 103 includes an autoencoder, and the autoencoder includes an encoder and a decoder;
[0054] The encoder is used to perform feature mapping on high-dimensional features and low-dimensional features to obtain potential features;
[0055] The decoder is used to decode and calculate the potential features to obtain comprehensive features.
[0056] Exemplarily, in the feature fusion module 103, the effective fusion of high-dimensional features and low-dimensional features can be achieved by constructing or calling an autoencoder. First, the high-dimensional features and low-dimensional features are input into the encoder part of the autoencoder, and feature mapping is performed through a multi-layer neural network. In this process, the encoder performs a nonlinear transformation on the input features through an activation function (such as ReLU, Sigmoid, etc.) to extract potential features. Its goal is to compress the input features into a more compact representation, namely the potential features, which are usually represented in a low-dimensional space. The potential features not only retain the important information in the input features, but also capture their inherent nonlinear relationships, and can be regarded as an efficient representation of the input features, thereby effectively reflecting the characteristics of the original data.
[0057] Secondly, the latent features are fed into the decoder, which gradually reconstructs the synthetic features through a series of back-propagation operations. Preferably, the decoder also uses an activation function to ensure that the output conforms to a specific format or range. The training process of the autoencoder can be optimized by the back-propagation algorithm to minimize the reconstruction error, that is, the difference between the input features and the output synthetic features. This process continuously adjusts the weights of the encoder and decoder to ensure good fusion under different data distributions.
[0058] By using autoencoders, high-dimensional features and low-dimensional features can be effectively fused to obtain comprehensive features. These comprehensive features simplify the complexity of the model while retaining key information. They not only have better interpretability, but also show better performance in subsequent classification and regression tasks.
[0059] In one embodiment, the system further includes a model training module for:
[0060] Get the data set, which includes a training set and a test set;
[0061] Divide the training set into multiple subsets, which are used for cross-validation and model training;
[0062] Initialize the parameters of the LightGBM algorithm, including the learning rate, the number of trees, and the depth of the tree;
[0063] Use subsets for cross-validation to obtain the optimal model parameters;
[0064] The calibration model is trained according to the training set and the optimal model parameters to obtain a trained calibration model.
[0065] The model training module trains the calibration model by acquiring a data set, which includes a training set and a test set. The training set contains features and corresponding labels, and the calibration model will learn rules on these data. The test set contains features and labels so that it can be used to verify the generalization ability of the model after the model training is completed. Exemplarily, the training set can be randomly divided into k subsets of equal size through k-fold cross-validation, and 1 is selected from the k subsets each time as a validation set, and the remaining k-1 are used as training sets. This process is repeated k times to ensure that each subset can be used as a validation set in a certain iteration, thereby ensuring that the model can be evaluated on different data subsets, thereby avoiding overfitting due to uneven data distribution.
[0066] The calibration model described in this article is built based on the LightGBM algorithm. The LightGBM (LightGradient Boosting Machine) algorithm is an efficient gradient boosting framework. First, its parameters, such as learning rate, number of trees, and tree depth, are initialized to build the calibration model. Secondly, the calibration model is evaluated using a subset for k-fold cross-validation. Some evaluation indicators such as accuracy and mean square error can be used for evaluation. The performance indicators of each iteration are recorded, and the results under different parameter combinations are compared. The optimal model parameters are selected by summarizing. Finally, the calibration model is trained according to the training set and the optimal model parameters to obtain a trained calibration model.
[0067] In one of the embodiments, the system further comprises a model evaluation module, the model evaluation module comprising an indicator calculation submodule and a confusion matrix analysis submodule;
[0068] The indicator calculation submodule is used to calculate the indicators of the calibration model using the test set, including accuracy, recall, and mean square error;
[0069] The confusion matrix analysis submodule is used to generate a confusion matrix based on the use of the test set, and the confusion matrix is used to visualize the calibration results of the calibration model.
[0070] The model evaluation module includes an indicator calculation submodule and a confusion matrix analysis submodule. The model performance indicators are calculated and verified to ensure the effectiveness of the calibration model in practical applications. The main function of the indicator calculation submodule is to use the test set to evaluate the performance of the calibration model. The model performance can be evaluated by calculating the accuracy of the number of correctly predicted samples to the total number of samples, the recall rate of all true positive examples correctly predicted by the model as positive examples, and the mean square error, which is the average of the squares of the differences between all predicted values and true values. The confusion matrix analysis submodule is mainly used to generate and analyze the confusion matrix. The confusion matrix is a table that is usually used to describe the performance of a classification model. Its rows represent actual categories and its columns represent predicted categories. It can intuitively show the prediction effect of the model on each category, help analyze which categories the model performs well or poorly, and visualize the calibration results.
[0071] In one embodiment, the system further includes a model updating module, configured to:
[0072] Adjusting the parameters of the calibration model based on the online learning algorithm to obtain an updated calibration model;
[0073] The updated calibration model is evaluated through the model evaluation module to obtain the updated model performance;
[0074] When the updated model performance meets the preset performance index, the updated calibration model is updated to the calibration feedback module.
[0075] The model update module can learn and adjust instantly when receiving new data based on the online learning algorithm without retraining the entire model to obtain an updated calibration model, where the new data is the original data received. And the performance of the updated calibration model is calculated by the model evaluation module, which may include accuracy, precision, recall, etc. If the performance of the updated calibration model meets the preset performance indicators, such as the accuracy is higher than a certain threshold, the updated calibration model will be sent to the calibration feedback module. This process may involve saving the new model parameters and using them for subsequent predictions and decisions. This module integrates the system components of online learning, model evaluation and dynamic feedback. Through the synergy of these functions, it ensures that the calibration model is continuously optimized in a changing environment to achieve the preset performance indicators. This not only improves the adaptive ability of the system, but also enhances the accuracy and reliability of the model.
[0076] In one embodiment, the system further includes a visualization module for converting the influence data of different features of the original data on the performance of the calibration model and the calibration results into visualization parameters; the visualization parameters are used to display the influence data and the calibration results in a display medium.
[0077] Specifically, the main purpose of the visualization module is to visualize the influence of different features extracted from the original data on the performance of the calibration model and the calibration results, so that users can intuitively understand the operating status of the electrical equipment and the effectiveness of the calibration. This visualization process may include data format conversion, data normalization, etc. The converted visualization parameters are displayed in a display medium, which can be a variety of devices such as computer screens, mobile devices, interactive exhibits, etc.
[0078] In one embodiment, based on the calibration model, the comprehensive features are detected and the calibration results are output, including:
[0079] Based on the calibration model, the comprehensive features are detected to obtain calibration values;
[0080] Compare the calibration value with the preset threshold and output the calibration result.
[0081] The calibration model is used to detect and calibrate the comprehensive characteristics of the remote water data, and the calibration value is output. The calibration value is a quantitative evaluation of the comprehensive characteristics, indicating the deviation or degree of match between the characteristics and the expected standard. The calibration value is compared with a pre-set threshold, and the system finally outputs the calibration result. The calibration result can be qualified or unqualified. This threshold is usually determined based on statistical analysis to determine whether the characteristics meet the expected standard.
[0082] In one embodiment, if Figure 2 As shown, the present invention also provides an automatic calibration method for electrical equipment based on artificial intelligence, the method comprising:
[0083] S201: Acquire raw data of the electrical device, where the raw data is used to characterize the operating state of the electrical device, and the raw data includes current, voltage, and temperature;
[0084] S202: extracting features from the original data based on the deep learning model to obtain high-dimensional features and low-dimensional features, where the high-dimensional features are used to describe nonlinear relationships in the original data, and the low-dimensional features are used to describe key information of the original data;
[0085] S203: Fusing the high-dimensional features and the low-dimensional features to obtain comprehensive features;
[0086] S204: Based on the calibration model, the comprehensive features are detected and the calibration results are output. The calibration results are used to indicate whether the electrical equipment is calibrated properly.
[0087] Specifically, the method uses a deep learning model to obtain the raw data of electrical equipment, extract its high-dimensional and low-dimensional features, perform feature fusion, and obtain comprehensive features that can describe its nonlinear relationships and key information. The comprehensive features are calibrated through the calibration model to effectively identify and describe the complex relationships in the raw data and eliminate the errors caused by noise, drift or nonlinearity in the raw data, obtain calibration results, and realize automatic calibration of electrical equipment. As the model continues to learn and optimize, the system can better adapt to environmental changes and equipment aging, improve the calibration capability of electrical equipment under different working conditions, and improve the efficiency and accuracy of calibration.
[0088] Exemplarily, in this method, high-dimensional features and low-dimensional features are fused to obtain comprehensive features, including:
[0089] Build an autoencoder, which includes an encoder and a decoder;
[0090] The encoder is used to perform feature mapping on high-dimensional features and low-dimensional features to obtain potential features;
[0091] The decoder is used to decode and calculate the potential features to obtain comprehensive features.
[0092] Exemplarily, the method further includes:
[0093] Get the data set, which includes a training set and a test set;
[0094] Divide the training set into multiple subsets, which are used for cross-validation and model training;
[0095] Initialize the parameters of the LightGBM algorithm, including the learning rate, the number of trees, and the depth of the tree;
[0096] Use subsets for cross-validation to obtain the optimal model parameters;
[0097] The calibration model is trained according to the training set and the optimal model parameters to obtain a trained calibration model.
[0098] Exemplarily, the method further includes:
[0099] Use the test set to calculate the metrics of the calibration model, including accuracy, recall, and mean square error;
[0100] A confusion matrix is generated based on the use of the test set and is used to visualize the calibration results of the calibration model.
[0101] Exemplarily, the method further includes:
[0102] Adjusting the parameters of the calibration model based on the online learning algorithm to obtain an updated calibration model;
[0103] The updated calibration model is evaluated through the model evaluation module to obtain the updated model performance;
[0104] When the updated model performance meets the preset performance index, the updated calibration model is updated to the calibration feedback module.
[0105] Exemplarily, the method further includes:
[0106] The influence data of different characteristics of the original data on the performance of the calibration model and the calibration results are converted into visualization parameters; the visualization parameters are used to display the influence data and the calibration results in a display medium.
[0107] Exemplarily, the method detects the comprehensive features based on the calibration model and outputs the calibration results, including:
[0108] Based on the calibration model, the comprehensive features are detected to obtain calibration values;
[0109] Compare the calibration value with the preset threshold and output the calibration result.
[0110] In one embodiment, the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of an automatic calibration method for electrical equipment based on artificial intelligence are implemented. Specifically, the steps may include data acquisition, feature extraction, calibration detection, etc., so as to utilize a deep learning model to realize automatic calibration of electrical equipment and improve calibration accuracy and reliability.
[0111] In one embodiment, the present invention also provides a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of an automatic calibration method for an electrical device based on artificial intelligence. The computer storage medium may include a random access memory, a read-only memory, a flash memory or other solid-state storage technology, a digital versatile disk or other optical storage, a cassette, a magnetic tape, a disk storage or other magnetic storage device.
[0112] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.
Claims
1. An automatic calibration system for electrical equipment based on artificial intelligence, characterized in that: include: A data acquisition module is used to obtain raw data of the electrical equipment, wherein the raw data is used to characterize the operating state of the electrical equipment, and the raw data includes current, voltage, and temperature; A feature extraction module, used to extract features from the original data based on a deep learning model to obtain high-dimensional features and low-dimensional features, wherein the high-dimensional features are used to describe nonlinear relationships in the original data, and the low-dimensional features are used to describe key information of the original data; A feature fusion module, used for fusing the high-dimensional features with the low-dimensional features to obtain comprehensive features; The calibration feedback module is used to detect the comprehensive characteristics based on the calibration model and output a calibration result, wherein the calibration result is used to indicate whether the electrical device is calibrated properly.
2. The automatic calibration system for electrical equipment according to claim 1, characterized in that: The feature fusion module includes an autoencoder, and the autoencoder includes an encoder and a decoder; The encoder is used to perform feature mapping on the high-dimensional features and the low-dimensional features to obtain potential features; The decoder is used to decode and calculate the potential features to obtain the comprehensive features.
3. The automatic calibration system for electrical equipment according to claim 1, characterized in that: Also includes model training modules for: Obtaining a data set, wherein the data set includes a training set and a test set; Dividing the training set into multiple subsets, wherein the subsets are used for cross-validation and model training; Initialize the parameters of the LightGBM algorithm, including the learning rate, the number of trees, and the depth of the tree; Using the subsets to perform cross validation to obtain optimal model parameters; The calibration model is trained according to the training set and the optimal model parameters to obtain the trained calibration model.
4. The automatic calibration system for electrical equipment according to claim 3, characterized in that: Also included is a model evaluation module, the model evaluation module including an indicator calculation submodule and a confusion matrix analysis submodule; The index calculation submodule is used to calculate the index of the calibration model using the test set, and the index includes accuracy, recall rate, and mean square error; The confusion matrix analysis submodule is used to generate the confusion matrix based on the test set, and the confusion matrix is used to visualize the calibration result of the calibration model.
5. The automatic calibration system for electrical equipment according to claim 4, characterized in that: Also included is a model updating module for: Adjusting the parameters of the calibration model based on an online learning algorithm to obtain an updated calibration model; Evaluating the updated calibration model through the model evaluation module to obtain updated model performance; When the updated model performance meets the preset performance index, the updated calibration model is updated to the calibration feedback module.
6. The automatic calibration system for electrical equipment according to claim 1, characterized in that: It also includes a visualization module for converting the influence data of different characteristics of the original data on the performance of the calibration model and the calibration results into visualization parameters; the visualization parameters are used to display the influence data and the calibration results in a display medium.
7. The automatic calibration system for electrical equipment according to claim 1, characterized in that: The detecting the comprehensive features based on the calibration model and outputting the calibration results include: Based on the calibration model, the comprehensive feature is detected to obtain a calibration value; The calibration value is compared with a preset threshold value, and the calibration result is output.
8. An automatic calibration method for electrical equipment based on artificial intelligence, characterized in that: The method comprises: Acquire raw data of the electrical device, where the raw data is used to characterize the operating state of the electrical device, and the raw data includes current, voltage, and temperature; Extracting features from the original data based on a deep learning model to obtain high-dimensional features and low-dimensional features, wherein the high-dimensional features are used to describe nonlinear relationships in the original data, and the low-dimensional features are used to describe key information of the original data; Fusing the high-dimensional features with the low-dimensional features to obtain comprehensive features; Based on the calibration model, the comprehensive characteristics are detected and a calibration result is output, wherein the calibration result is used to indicate whether the electrical device is calibrated properly.
9. An electronic 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 the method according to claim 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 the method described in claim 8 are implemented.