Charging method and system capable of adjusting charging strategy based on accident rate

Abstract

The invention discloses a charging method capable of adjusting a charging strategy based on an accident rate, and the method comprises the steps: obtaining accident data of a plurality of different vehicle types, carrying out the preprocessing of the accident data, and carrying out the training based on the preprocessed accident data, so as to obtain an accident rate prediction model, the accident rate prediction model is used for predicting the accident rate of an access vehicle accessed to the charging pile; optimizing the accident rate prediction model, and deploying the optimized accident rate prediction model to the charging pile; acquiring vehicle type parameters and traveled mileage of the accessed vehicle; on the basis of the vehicle model parameters and an accident rate prediction model, calculating the probability that each type of charging accident occurs to the accessed vehicle, so as to obtain a plurality of predicted accident rates corresponding to the charging accidents; and adjusting a charging strategy of the accessed vehicle based on the predicted accident rate. Based on the charging method, the charging decision can be more scientific and accurate, and the charging safety is improved.

Description

Technical Field

[0001] The present invention relates to the field of deep learning prediction and electric vehicle charging technology, and in particular to a charging method and system capable of adjusting a charging strategy based on an accident rate. Background Art

[0002] In order to avoid various accidents that may occur during the charging process of the car, the mainstream practice in the industry has adopted a variety of technical measures to enhance the safety performance of the charging pile. Specifically, it includes adding sophisticated protection circuits and smoke detection devices inside the charging pile, and using flame-retardant materials to make the charging pile. Among them, the protection circuit can monitor key parameters such as current and voltage in real time, and can quickly cut off the power supply once an abnormality is detected, thereby effectively preventing accidents caused by electrical failures; the smoke detection device can sensitively sense tiny smoke particles, and once smoke is detected, it immediately triggers the alarm mechanism so that the staff can quickly intervene to prevent the spread of the fire; the flame-retardant material is not easy to burn at high temperatures, and even if it burns, it can effectively slow down the spread of the fire, buying precious time for rescue work. These measures have greatly improved the safety of the charging pile, but it is still difficult to completely eliminate the risks that the vehicle itself may encounter during the charging process, such as battery thermal runaway, line aging and other problems, which have potential safety hazards.

[0003] In order to further improve the ability to respond to accidents, some charging stations have also adopted AI recognition solutions such as adding cameras. These cameras can monitor the charging area around the clock and use advanced artificial intelligence technology to identify potential safety hazards such as illegal operations by personnel and abnormal vehicle conditions. Once an accident occurs, these intelligent recognition systems can quickly identify and trigger the emergency response mechanism, notifying relevant personnel to arrive at the scene immediately for processing. However, this solution focuses more on rapid response after the event. Although it can reduce the losses caused by the accident to a certain extent, it cannot fundamentally reduce the probability of accidents in advance. Summary of the invention

[0004] The object of the present invention is to provide a charging method and system that can predict the accident rate according to the vehicle model and adjust the charging strategy of the connected vehicle based on the predicted accident rate.

[0005] To achieve the above object, the present invention discloses a charging method capable of adjusting a charging strategy based on an accident rate, comprising: Acquire accident data of several different vehicle models, preprocess the accident data, and perform training based on the preprocessed accident data to obtain an accident rate prediction model, wherein the accident rate prediction model is used to predict the accident rate of vehicles connected to the charging pile; Deploying the accident rate prediction model to the charging pile; Obtaining the vehicle model parameters and mileage of the connected vehicle; Based on the vehicle type parameters and the accident rate prediction model, the probability of each type of charging accident occurring to the connected vehicle is calculated to obtain a number of predicted accident rates corresponding to the charging accidents; Based on the predicted accident rate, a charging strategy for the connected vehicle is adjusted.

[0006] Specifically, adjusting the charging strategy of the connected vehicle includes: When the predicted accident rates are all lower than a first preset threshold, controlling the charging pile to charge the connected vehicle with a preset input power set by the connected vehicle; When a certain predicted accident rate is higher than or equal to the first preset threshold, the charging pile is controlled to reduce the output power to the connected vehicle.

[0007] Specifically, the training method of the accident rate prediction model includes: In the input layer, a plurality of input nodes corresponding to the data in the accident data are created to obtain a plurality of input features; In the feature interaction layer, feature crossover is performed on the input features; In the fully connected layer, high-level features are extracted from the input features after feature crossing, and dimensionality reduction is performed on the input features; In the output layer, the predicted accident rate corresponding to each type of charging accident is output through an activation function.

[0008] Specifically, before deploying the accident rate prediction model to the charging pile, the accident rate prediction model needs to be optimized.

[0009] Specifically, the charging method also evaluates the performance of the accident rate prediction model through K-fold cross validation. When the accuracy of the accident rate prediction model is less than or equal to the second preset threshold or the recall rate of the accident rate prediction model is greater than or equal to the third preset threshold, the accident rate prediction model is retrained.

[0010] Specifically, when the connected vehicle finishes charging, the charging data generated by the connected vehicle during the charging process is saved, and the accident rate prediction model is regularly updated and optimized based on the charging data and the accident data.

[0011] The present invention also discloses a charging system capable of adjusting a charging strategy based on an accident rate. The charging system operates based on the charging method described above.

[0012] The present invention also discloses a charging system capable of adjusting a charging strategy based on an accident rate, comprising: one or more processors; Memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for executing the charging method as described above.

[0013] The present invention also discloses a computer-readable storage medium, characterized in that it includes a computer program, and the computer program can be executed by a processor to complete the charging method as described above.

[0014] Compared with the prior art, the charging method provided by the above technical solution of the present invention can adjust the charging strategy based on the accident rate, and obtains an accident rate prediction model by acquiring accident data of different models for training. The accident rate prediction model can predict the accident rate of various charging accidents of the connected vehicles connected to the charging pile, and adjust the charging strategy of the connected vehicles in real time according to the predicted accident rate. The charging method proposed by the present invention can ensure the safety of the charging process of the connected vehicles, and the model trained according to the historical accident data can make the decision more scientific and accurate, improve the operational efficiency and safety of the charging pile, and save the cost of safety prevention from the source. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The figure is a flow chart of a charging method capable of adjusting a charging strategy based on an accident rate in an embodiment of the present invention.

[0016] Figure 2 The figure is a flow chart of a charging method capable of adjusting a charging strategy based on an accident rate in another embodiment of the present invention.

[0017] Figure 3 Flow chart of the training method of the accident rate prediction model in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to explain the technical content, structural features, achieved objectives and effects of the present invention in detail, the following is a detailed description in conjunction with the implementation methods and the accompanying drawings.

[0019] The embodiment of the present invention discloses a charging method that can adjust the charging strategy based on the accident rate to prevent charging accidents and improve the safety of the vehicle charging process. Figure 1 As shown, the charging method specifically includes the following steps: S11: Obtain accident data of several different vehicle models and pre-process the accident data.

[0020] S12: Perform training based on the pre-processed accident data to obtain an accident rate prediction model. The accident rate prediction model is used to predict the accident rate of the connected vehicle connected to the charging pile.

[0021] S13: Deploy the accident rate prediction model to the charging pile.

[0022] S14: Obtain the vehicle model parameters and mileage of the connected vehicle.

[0023] S15: Based on the vehicle model parameters and the accident rate prediction model, the probability of various types of charging accidents occurring in the connected vehicle is calculated to obtain a number of predicted accident rates corresponding to the charging accidents.

[0024] S16: Based on the predicted accident rate, the charging strategy of the connected vehicle is adjusted.

[0025] Compared with the prior art, the charging method proposed in the present invention, which can adjust the charging strategy based on the accident rate, shows advantages and innovations. This method collects and analyzes accident data that occurred in the actual charging process of various different models, and constructs a highly accurate accident rate prediction model. The accident rate prediction model can be based on the model parameters and mileage of each vehicle connected to the charging pile, and can accurately predict the probability of various accidents that may occur during the charging process, including but not limited to electrical failures, overheating, short circuits and even battery damage. Potential risks, at the same time, this charging method can also guide the charging pile to directly apply the prediction results to the dynamic adjustment of the charging strategy, and intelligently adjust key parameters such as charging current, voltage, and charging time, so as to effectively avoid or reduce the probability of potential accidents, greatly improving the operating efficiency and overall safety of the charging pile.

[0026] In contrast, traditional charging management often relies on empirical judgment or fixed charging strategies, which are difficult to effectively deal with complex and changeable actual situations. The present invention, through the prediction model trained based on accident data, can make charging decisions more scientific and accurate, which not only reduces equipment damage and maintenance costs caused by accidents, but also saves capital investment in safety prevention from the source, achieving a dual improvement in economic and safety benefits.

[0027] Specifically, see Figure 2 As shown, adjusting the charging strategy of the connected vehicle also includes: S161: Determine whether the predicted accident rates are all lower than the first preset threshold value. If so, jump to step S162; if not, jump to step S163.

[0028] S162: Control the charging pile to charge the connected vehicle with a preset input power set by the connected vehicle.

[0029] S163: Control the charging pile to reduce the output power to the connected vehicle.

[0030] In this embodiment, the first preset threshold is determined by the vehicle model and the mileage. When a certain predicted accident rate is higher than or equal to the first preset threshold, it is preferred to reduce the output power to 80% of the maximum value, thereby reducing the possibility of accidents and ensuring the safety of the charging process.

[0031] Specifically, the accident data includes vehicle model, mileage, environmental parameters of the vehicle when the charging accident occurs, and the type of charging accident. The preprocessing of the accident data includes: (1) converting the vehicle model into a low-order dense vector. In this embodiment, the vehicle model is converted into a low-order dense vector using an embedding layer; (2) standardizing or normalizing the mileage to ensure that the numerical range is suitable for neural network input; (3) standardizing or normalizing the environmental parameters and adding timestamps so that the accident rate prediction model can understand the time dependence of the data in the environmental parameters; (4) converting the type of charging accident into a low-order dense vector. In this embodiment, the type of charging accident is converted into a low-order dense vector using an embedding layer. Among them, the environmental parameters also include data such as temperature and humidity when the accident occurs, and the types of charging accidents include battery fire, short circuit, etc.

[0032] Specifically, see Figure 3 As shown, in this embodiment, the training method of the accident rate prediction model includes: S21: In the input layer, a plurality of input nodes corresponding to the data in the accident data are created to obtain a plurality of input features. In this embodiment, the plurality of input nodes correspond to a plurality of data in the accident data (such as vehicle model, mileage, etc.).

[0033] S22: In the feature interaction layer, feature crossover is performed on the input features.

[0034] Feature crossing enables the accident rate prediction model to take into account the interaction between different input features at an early stage, which helps to improve the accident rate prediction model's ability to learn complex patterns.

[0035] S23: In the fully connected layer, high-level features are extracted from the input features after feature crossing, and dimensionality reduction is performed on the input features.

[0036] S24: In the output layer, the predicted accident rate corresponding to each type of charging accident is output through the activation function.

[0037] In this embodiment, for predicting the accident rate, the charging method generates a probability value between 0 and 1 through a single neuron output layer with a sigmoid activation function during the training process. For the classification of charging accident types, a new branch is added to the output layer, which is connected to a softmax activation function after several fully connected layers. The softmax activation function is used to output the probability distribution of each possible charging accident type.

[0038] Specifically, before deploying the accident rate prediction model to the charging pile, the accident rate prediction model needs to be optimized. The optimization methods include: (1) Using classification cross entropy as the loss function; (2) Applying regularization methods (such as Dropout or L2 regularization technology) to prevent overfitting of the accident rate prediction model; (3) Tuning the hyperparameters of the accident rate prediction model training process. Hyperparameters refer to training parameters such as learning rate and batch size. The best hyperparameter configuration can be found by grid search, random search or Bayesian optimization. In this embodiment, grid search is used by default for hyperparameter configuration.

[0039] Specifically, in this embodiment, the performance of the accident rate prediction model is also evaluated by K-fold cross validation, and when the accuracy of the accident rate prediction model is less than or equal to the second preset threshold or the recall rate of the accident rate prediction model is greater than or equal to the third preset threshold, the accident rate prediction model is retrained. The preferred value of the second preset threshold is 70%, and the preferred value of the third preset threshold is 30%.

[0040] Specifically, in this embodiment, when the connected vehicle finishes charging, the charging data generated by the connected vehicle during the charging process will also be saved, and the accident rate prediction model will be updated and optimized regularly (preferably every one month) based on the charging data and accident data. The training data set of the accident rate prediction model is updated through the collected actual charging data, and accident data is also regularly collected from the network to ensure that the accident rate prediction model always maintains the best performance.

[0041] The present invention also discloses a charging system capable of adjusting a charging strategy based on an accident rate, and the charging system operates based on the charging method described above.

[0042] The present invention also discloses another charging system that can adjust the charging strategy based on the accident rate, which includes one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the program includes instructions for executing the charging method as described above. The processor can adopt a general central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits to execute related programs to implement the functions required to be executed by the modules in the charging system of the embodiment of the present application, or to execute the charging method of the method embodiment of the present application.

[0043] The present invention also discloses a computer-readable storage medium, which includes a computer program, and the computer program can be executed by a processor to complete the charging method as described above. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium can be a read-only memory (ROM), or a random access memory (RAM), or a magnetic medium, such as a floppy disk, a hard disk, a tape, a magnetic disk, or an optical medium, such as a digital versatile disc (DVD), or a semiconductor medium, such as a solid state disk (SSD), etc.

[0044] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. Therefore, equivalent changes made according to the scope of the patent application of the present invention are still within the scope covered by the present invention.

Claims

1. A charging method capable of adjusting the charging strategy based on the accident rate, characterized in that: include: Acquire accident data of several different vehicle models, preprocess the accident data, and perform training based on the preprocessed accident data to obtain an accident rate prediction model, wherein the accident rate prediction model is used to predict the accident rate of vehicles connected to the charging pile; Deploying the accident rate prediction model to the charging pile; Obtaining the vehicle model parameters and mileage of the connected vehicle; Based on the vehicle type parameters and the accident rate prediction model, the probability of each type of charging accident occurring to the connected vehicle is calculated to obtain a number of predicted accident rates corresponding to the charging accidents; Based on the predicted accident rate, a charging strategy for the connected vehicle is adjusted.

2. The charging method capable of adjusting the charging strategy based on the accident rate according to claim 1, characterized in that: Adjusting the charging strategy of the connected vehicle includes: When the predicted accident rates are all lower than a first preset threshold, controlling the charging pile to charge the connected vehicle with a preset input power set by the connected vehicle; When a certain predicted accident rate is higher than or equal to the first preset threshold, the charging pile is controlled to reduce the output power to the connected vehicle.

3. The charging method capable of adjusting the charging strategy based on the accident rate according to claim 1, characterized in that: The training method of the accident rate prediction model includes: In the input layer, a plurality of input nodes corresponding to the data in the accident data are created to obtain a plurality of input features; In the feature interaction layer, feature crossover is performed on the input features; In the fully connected layer, high-level features are extracted from the input features after feature crossing, and dimensionality reduction is performed on the input features; In the output layer, the predicted accident rate corresponding to each type of charging accident is output through an activation function.

4. The charging method capable of adjusting the charging strategy based on the accident rate according to claim 1, characterized in that: Before deploying the accident rate prediction model to the charging pile, the accident rate prediction model needs to be optimized.

5. The charging method capable of adjusting the charging strategy based on the accident rate according to claim 1, characterized in that: The performance of the accident rate prediction model is also evaluated through K-fold cross validation. When the accuracy of the accident rate prediction model is less than or equal to the second preset threshold or the recall rate of the accident rate prediction model is greater than or equal to the third preset threshold, the accident rate prediction model is retrained.

6. The charging method capable of adjusting the charging strategy based on the accident rate according to claim 1, characterized in that: When the connected vehicle finishes charging, the charging data generated by the connected vehicle during the charging process is saved, and the accident rate prediction model is regularly updated and optimized based on the charging data and the accident data.

7. A charging system capable of adjusting charging strategy based on accident rate, characterized in that: The charging system operates based on the charging method of any one of claims 1 to 6, wherein the charging strategy can be adjusted based on the accident rate.

8. A charging system capable of adjusting charging strategy based on accident rate, characterized in that: include: one or more processors; Memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, the programs including instructions for executing the charging method of any one of claims 1 to 6 that can adjust the charging strategy based on the accident rate.

9. A computer-readable storage medium, characterized in that: The invention comprises a computer program which can be executed by a processor to implement the charging method capable of adjusting the charging strategy based on the accident rate as claimed in any one of claims 1 to 6.

Images