Intelligent detection method for electric vehicle charging equipment

Abstract

The invention discloses an intelligent detection method for electric vehicle charging equipment, and relates to the technical field of power equipment fault detection. The method comprises the steps of obtaining real-time operation data of target charging equipment, and intercepting a data sequence of a preset duration to obtain a real-time operation signal; filtering and aligning the various signals to obtain a first feature matrix; performing normalization processing on the first feature matrix to obtain a second feature matrix; and taking the second feature matrix as the input of a fault detection model to obtain a fault detection result. By introducing the bidirectional gating cycle cell convolutional neural network, the expression ability of the model to a complex data mode is enhanced, the dependence on artificial feature engineering is reduced, and the detection effect is optimized. Moreover, the end-to-end training mode of the model and the comprehensive analysis of multiple signals improve the adaptability and distinguishing ability to different fault modes, and effectively solve the problem of insufficient accuracy of the traditional method.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment fault detection, and in particular to an intelligent detection method for electric vehicle charging equipment. Background Art

[0002] Electric vehicles have become the development trend of the future automobile industry. The development of electric vehicles is inseparable from the charging infrastructure. At present, the charging infrastructure mainly includes DC charging piles, AC charging piles, on-board chargers, charging stations, etc. Among them, the DC charging pile has a built-in charging module, and the charging power can reach tens to hundreds of kilowatts. The charging rate is fast and can meet the requirements of high-power fast charging. However, the charging module has a high operating voltage and contains a large number of power electronic devices and electrolytic capacitors. Under high voltage and repeated temperature cycles, it is easy to cause aging failure and device failure, becoming the component with the highest failure rate in the DC charging pile. It is of great significance to monitor the possible failures of the charging module of the charging pile.

[0003] At present, the fault detection of charging modules is mainly classified through traditional machine learning models (such as support vector machines, random forests, etc.). However, these methods rely on manual feature engineering, cannot fully explore the deep patterns of data, and have poor adaptability to changes in data distribution, resulting in low accuracy of detection results. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of low accuracy mentioned in the above background technology and to propose an intelligent detection method for electric vehicle charging equipment.

[0005] In a first aspect of the present invention, there is provided a method for intelligent detection of electric vehicle charging equipment, the method comprising: Acquire real-time operation data of the target charging device, and intercept a data sequence of a preset time length to obtain a real-time operation signal; the real-time operation signal includes a plurality of voltage signals and current signals; Filtering and aligning the multiple signals to obtain a first feature matrix; Normalizing the first characteristic matrix to obtain a second characteristic matrix; The second characteristic matrix is ​​used as the input of the fault detection model to obtain a fault detection result; the fault detection model is composed of a bidirectional gated recurrent unit and a convolutional neural network.

[0006] Optionally, the charging module of the target charging device includes a front-stage circuit and a rear-stage circuit; the front-stage circuit adopts a three-phase VIENNA rectifier structure; and the rear-stage circuit adopts a full-bridge LLC resonant transformer structure.

[0007] Optionally, the post-stage circuit includes a full-bridge inverter circuit, a resonant cavity circuit, a transformer module and a rectifier filter circuit; The operation signal includes a current signal of a three-phase input in the front-stage circuit, a voltage signal of the full-bridge inverter circuit, a current signal of the resonant cavity circuit, and a voltage signal of the rectifier and filter circuit.

[0008] Optionally, the fault detection model includes a data fusion module, a feature extraction and enhancement module, a feature fusion module and a classification prediction module; wherein: The data fusion module is used to perform feature fusion on the input data through a bidirectional gated recurrent unit to obtain a hidden state vector; The feature extraction and enhancement module is used to extract multi-scale features of the hidden state vector through multiple convolution branches, and perform attention weighting to obtain multiple feature tensors; The feature fusion module is used to perform global average pooling and concatenation on multiple feature tensors to obtain a target feature vector; The classification prediction module is used to classify the target feature vector through a fully connected layer to obtain a fault detection result.

[0009] Optionally, the feature extraction and enhancement module includes a multi-scale differentiation layer and a plurality of convolution branches, namely a first convolution branch, a second convolution branch and a third convolution branch; The multi-scale differentiation layer is used to perform maximum pooling and average pooling on the hidden state vector to obtain a first vector and a second vector; the first convolution branch receives the first vector as input; the second convolution branch receives the second vector as input; the third convolution branch receives the hidden state vector as input; The target convolution branch is used to downsample the input of the branch through one-dimensional convolution and maximum pooling to obtain a first feature tensor; and to enhance the first feature tensor through an attention module to obtain a second feature tensor; the target convolution branch is any one of the multiple convolution branches.

[0010] Optionally, the downsampling structure of each convolution branch is consistent; the attention module calculates the attention weight through two fully connected layers; The calculation process of the target convolution branch is: ; Wherein, F2 is the output of the target convolution branch; X is the input of the target convolution branch; F1 is the first feature tensor; f is the operation operator, the subscripts Conv and MP represent one-dimensional convolution and maximum pooling respectively, the superscript represents the convolution kernel size or the pooling window size, and the stride is 2; W is the channel attention weight; GAP represents global average pooling; S1 and S2 are the weight parameters of the two fully connected layers respectively; ReLu and sigmoid are activation functions.

[0011] Optionally, the operation process of the feature fusion module includes: Performing global average pooling on each channel of a target second feature tensor to obtain a temporary vector; the target second feature tensor is any one of the multiple second feature tensors; Multiple temporary vectors are concatenated and spliced ​​to obtain the target feature vector.

[0012] Optionally, the classification prediction module uses a multi-layer perceptron and a Softmax function to perform classification prediction.

[0013] Beneficial effects of the present invention: The present invention proposes an intelligent detection method for electric vehicle charging equipment, which includes: acquiring real-time operation data of a target charging equipment, and intercepting a data sequence of a preset time length to obtain a real-time operation signal; the real-time operation signal includes a plurality of voltage signals and a current signal; filtering and aligning the plurality of signals to obtain a first characteristic matrix; normalizing the first characteristic matrix to obtain a second characteristic matrix; using the second characteristic matrix as an input of a fault detection model to obtain a fault detection result; the fault detection model is composed of a bidirectional gated recurrent unit and a convolutional neural network.

[0014] By introducing a bidirectional gated recurrent cell convolutional neural network, the model's ability to express complex data patterns is enhanced, while reducing reliance on artificial feature engineering and optimizing detection results. In addition, the model's end-to-end training method and comprehensive analysis of multiple signals improve its adaptability and ability to distinguish different fault modes, effectively solving the problem of insufficient accuracy of traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings.

[0016] Figure 1 A flow chart of an intelligent detection method for electric vehicle charging equipment is provided for an embodiment of the present invention; Figure 2 A circuit topology diagram of a charging module is provided for an embodiment of the present invention; Figure 3 A structural schematic diagram of a fault detection model is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] The embodiment of the present invention provides an intelligent detection method for electric vehicle charging equipment. Figure 1 , Figure 1 A flow chart of an intelligent detection method for electric vehicle charging equipment provided by an embodiment of the present invention. The method comprises the following steps: S101, acquiring real-time operation data of a target charging device, and intercepting a data sequence of a preset time length to obtain a real-time operation signal.

[0019] S102, filtering and aligning the multiple signals to obtain a first feature matrix.

[0020] S103, normalizing the first characteristic matrix to obtain a second characteristic matrix.

[0021] S104, using the second characteristic matrix as an input of a fault detection model to obtain a fault detection result.

[0022] Among them, the real-time operation signals include a variety of voltage signals and current signals; the fault detection model is composed of a bidirectional gated recurrent unit and a convolutional neural network; the fault detection results include a variety of open circuit faults.

[0023] Based on an intelligent detection method for electric vehicle charging equipment provided by an embodiment of the present invention, by introducing a bidirectional gated recurrent cell convolutional neural network, the model's ability to express complex data patterns is enhanced, while reducing dependence on artificial feature engineering and optimizing the detection effect. In addition, the model's end-to-end training method and comprehensive analysis of multiple signals improve the adaptability and ability to distinguish different fault modes, effectively solving the problem of insufficient accuracy of traditional methods.

[0024] In one implementation, by acquiring and processing the real-time operating data of the electric vehicle charging equipment, potential fault problems can be discovered in time, maintenance can be performed in advance, and safety hazards caused by the fault can be avoided.

[0025] In one implementation, the operation signal can be denoised by a low-pass filter. Different signals are aligned according to timestamps by linear interpolation or spline interpolation, so that different signals have corresponding values ​​at the same time point. By filtering and aligning multiple voltage and current signals, noise can be removed, data redundancy can be reduced, and signal quality can be ensured, making the input data of the fault detection model more accurate and efficient.

[0026] In one implementation, the maximum and minimum values ​​of the fault detection model training data set are used to normalize the operation signal data.

[0027] In one implementation, the method can be applied to different types of electric vehicle charging equipment and different usage scenarios, and has strong versatility and scalability. Figure 2 , Figure 2 A circuit topology diagram of a charging module provided in an embodiment of the present invention. The charging module of the target charging device includes a front-stage circuit and a rear-stage circuit; the front-stage circuit adopts a three-phase VIENNA rectifier structure; the rear-stage circuit adopts a full-bridge LLC resonant transformer structure. The rear-stage circuit includes a full-bridge inverter circuit, a resonant cavity circuit, a transformer module, and a rectifier filter circuit.

[0028] Among the failures of charging pile charging modules, the proportion of open circuit failures and short circuit failures of power devices is relatively high. The short circuit failure of power devices is short in time and has greater hazards. It is usually detected and protected by the design of the hardware circuit of the driver module, or the short circuit failure is converted into an open circuit failure by a fast fuse. The system can continue to operate after the open circuit failure of the power device occurs, but the voltage and current are distorted, which accelerates the aging and failure of non-faulty power devices, endangering the safe and reliable operation of the system. Therefore, the diagnosis and identification of open circuit failures of power devices in charging modules is of great significance to improving the reliability of module operation. Multiple open circuit failures can be used as classification labels, and normal operation can be used as a special fault label to annotate the training data set. The training process of the fault detection model is as follows: Step 1: Build a charging module simulation model of the target charging device.

[0029] Step 2: According to the charging module simulation model, multiple fault simulations are performed within the preset parameter range to collect operating signals when different faults occur, and combined with the injected fault type to generate labeled data.

[0030] Step three: Normalize each feature value of the labeled data according to the maximum and minimum values ​​to establish a data set.

[0031] Step 4: According to the data set, the preset neural network model is trained with the goal of maximizing accuracy to obtain a fault detection model.

[0032] In one embodiment, for Figure 2 The charging module shown in the figure has operating signals including the current signals (Ia, Ib and Ic) of the three-phase input in the front-stage circuit, the voltage signal of the full-bridge inverter circuit (voltage U1 between terminals p1 and p2), the current signal (Ir) of the resonant cavity circuit and the voltage signal of the rectifier filter circuit (voltage U2 between terminals p3 and p4). These signals reflect the dynamic changes of current and voltage in the charging equipment, which helps to deeply analyze the problems in the power conversion process and quickly identify potential faults.

[0033] In one embodiment, see Figure 3 , Figure 3 A structural schematic diagram of a fault detection model provided by an embodiment of the present invention.

[0034] The fault detection model includes a data fusion module, a feature extraction and enhancement module, a feature fusion module and a classification prediction module; among which: The data fusion module is used to perform feature fusion on the input data through a bidirectional gated recurrent unit (Bi-GRU) to obtain a hidden state vector.

[0035] The feature extraction and enhancement module is used to extract multi-scale features of the hidden state vector through multiple convolution branches and perform attention weighting to obtain multiple feature tensors.

[0036] The feature fusion module is used to perform global average pooling and concatenation on multiple feature tensors to obtain the target feature vector.

[0037] The classification prediction module is used to classify the target feature vector through the fully connected layer to obtain the fault detection result.

[0038] In one implementation, during the operation of electric vehicle charging equipment, its faults are often fed back by complex time series signals, and the previous and subsequent states of these signals may be crucial to the diagnosis of the fault. The bidirectional gated recurrent unit can capture the dependencies at different time points in the sequence by processing data from both the forward and backward directions at the same time. Specifically, forward propagation can analyze the timing regularity of the data from the beginning to the end of the sequence, while backward propagation can reversely mine potential fault information from the end to the beginning. This capture of bidirectional time dependencies improves the model's ability to understand complex signal patterns, allowing the model to comprehensively consider the changes in signals in different time dimensions, thereby making more accurate fault predictions.

[0039] In one implementation, the hidden state vector output at the last time step is taken as the final output.

[0040] In one implementation, multi-scale feature extraction is performed on the hidden state vector through multiple convolution branches, which can capture the key features of the signal from different scales and angles. This multi-dimensional feature extraction can effectively improve the model's ability to recognize complex fault modes. In addition, attention weighting can automatically adjust the weights according to the importance of different features, allowing the model to focus on key features and further improve the accuracy of fault diagnosis.

[0041] In one embodiment, the feature extraction and enhancement module includes a multi-scale differentiation layer and a plurality of convolution branches, namely a first convolution branch, a second convolution branch and a third convolution branch; The multi-scale differentiation layer is used to perform maximum pooling and average pooling on the hidden state vector to obtain a first vector and a second vector.

[0042] The target convolution branch is used to downsample the input of the branch through one-dimensional convolution and maximum pooling to obtain a first feature tensor; and to enhance the features of the first feature tensor through an attention module to obtain a second feature tensor.

[0043] Among them, the first convolution branch receives the first vector as input; the second convolution branch receives the second vector as input; the third convolution branch receives the hidden state vector as input; and the target convolution branch is any one of the multiple convolution branches.

[0044] In one implementation, the pooling window sizes of the maximum pooling and the average pooling in the multi-scale differentiation layer can both be 5×1 with a stride of 2.

[0045] In one implementation, the downsampling structure of each convolution branch is consistent; ×n in the convolution branch represents n convolution pooling structures Conv+MaxPool (Conv represents one-dimensional convolution, MaxPool represents maximum pooling) connected in sequence. Among them, n can be equal to 3, and the convolution kernel size of each convolution branch is 5×1, 3×1, and 3×1 respectively; the pooling window size is 2×1. The attention module (Attention) can use channel attention to calculate the attention weight through two fully connected layers.

[0046] The calculation process of the target convolution branch is: ; Where F2 is the output of the target convolution branch; X is the input of the target convolution branch; F1 is the first feature tensor; f is the operation operator, the subscripts Conv and MP represent one-dimensional convolution and maximum pooling respectively, the superscript represents the convolution kernel size or the pooling window size, and the stride is 2; W is the channel attention weight; GAP represents global average pooling; S1 and S2 are the weight parameters of the two fully connected layers respectively; ReLu and sigmoid are activation functions; operation symbols Indicates element-by-element multiplication; operator symbol In one embodiment, the operation process of the feature fusion module includes: Perform global average pooling on each channel of the target second feature tensor to obtain a temporary vector.

[0047] Multiple temporary vectors are concatenated and spliced ​​to obtain the target feature vector.

[0048] Among them, the target second feature tensor is any one of multiple second feature tensors obtained by multiple convolution branches.

[0049] In one implementation, global average pooling can not only extract global features, but also effectively reduce the dimension of data, reduce computational complexity, and reduce the burden of subsequent processing. Multiple temporary vectors are concatenated to integrate feature information extracted by different convolution branches. This concatenation method combines the information of multiple feature channels to form a unified target feature vector, which improves the representation ability of features and enables the model to consider signal features at multiple levels at the same time.

[0050] In one embodiment, the classification prediction module uses a multi-layer perceptron MLP and a Softmax function to perform classification prediction.

[0051] In one implementation, the hidden layer of MLP can be two layers, and the activation function uses ReLu. As a classic feedforward neural network, the multi-layer perceptron (MLP) processes information through multiple layers of neuron nodes and can effectively capture the nonlinear relationship between input features and outputs. In the fault detection of electric vehicle charging equipment, the change of fault characteristics is usually nonlinear. MLP can effectively map the relationship between complex input features and fault types through its deep hierarchical structure, thereby improving the accuracy of fault classification.

[0052] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An intelligent detection method for electric vehicle charging equipment, characterized in that: The method comprises: Acquire real-time operation data of the target charging device, and intercept a data sequence of a preset time length to obtain a real-time operation signal; the real-time operation signal includes a plurality of voltage signals and current signals; Filtering and aligning the multiple signals to obtain a first feature matrix; Normalizing the first characteristic matrix to obtain a second characteristic matrix; The second feature matrix is ​​used as an input of a pre-trained fault detection model to obtain a fault detection result; the fault detection model includes a data fusion module, a feature extraction and enhancement module, a feature fusion module and a classification prediction module; wherein: The data fusion module is used to perform feature fusion on the input data through a bidirectional gated recurrent unit to obtain a hidden state vector; The feature extraction and enhancement module is used to perform multi-scale feature extraction on the hidden state vector through multiple convolution branches, and perform attention weighting to obtain multiple feature tensors; specifically, the feature extraction and enhancement module includes a multi-scale differentiation layer and multiple convolution branches, which are respectively a first convolution branch, a second convolution branch and a third convolution branch; The multi-scale differentiation layer is used to perform maximum pooling and average pooling on the hidden state vector to obtain a first vector and a second vector; the first convolution branch receives the first vector as input; the second convolution branch receives the second vector as input; the third convolution branch receives the hidden state vector as input; A target convolution branch is used to downsample the input of the branch through one-dimensional convolution and maximum pooling to obtain a first feature tensor; and to enhance the first feature tensor through an attention module to obtain a second feature tensor; the target convolution branch is any one of the multiple convolution branches; The feature fusion module is used to perform global average pooling and concatenation on multiple feature tensors to obtain a target feature vector; The classification prediction module is used to classify the target feature vector through a fully connected layer to obtain a fault detection result.

2. The intelligent detection method for electric vehicle charging equipment according to claim 1 is characterized in that: The charging module of the target charging device includes a front-stage circuit and a rear-stage circuit; the front-stage circuit adopts a three-phase VIENNA rectifier structure; and the rear-stage circuit adopts a full-bridge LLC resonant converter structure.

3. The intelligent detection method for electric vehicle charging equipment according to claim 2 is characterized in that: The post-stage circuit includes a full-bridge inverter circuit, a resonant cavity circuit, a transformer module and a rectifier and filter circuit; The operation signal includes a current signal of a three-phase input in the front-stage circuit, a voltage signal of the full-bridge inverter circuit, a current signal of the resonant cavity circuit, and a voltage signal of the rectifier and filter circuit.

4. The intelligent detection method for electric vehicle charging equipment according to claim 1, characterized in that: The downsampling structure of each convolution branch is consistent; the attention module calculates the attention weight through two fully connected layers; The calculation process of the target convolution branch is: ; Among them, F2 is the output of the target convolution branch; X is the input of the target convolution branch; F1 is the first feature tensor; f is the operation operator, the subscripts Conv and MP represent one-dimensional convolution and maximum pooling respectively, the superscript represents the convolution kernel size or the pooling window size, and the stride is 2; W is the channel attention weight; GAP represents global average pooling; S1 and S2 are the weight parameters of the two fully connected layers respectively; ReLu and sigmoid are activation functions.

5. The intelligent detection method for electric vehicle charging equipment according to claim 1, characterized in that: The operation process of the feature fusion module includes: Performing global average pooling on each channel of a target second feature tensor to obtain a temporary vector; the target second feature tensor is any one of the multiple second feature tensors; Multiple temporary vectors are concatenated and spliced ​​to obtain the target feature vector.

6. The intelligent detection method for electric vehicle charging equipment according to claim 1, characterized in that: The classification prediction module uses a multi-layer perceptron and a Softmax function to perform classification prediction.

Images