Dynamic control method of battery thermal management system based on deep learning

Abstract

The invention relates to the field of batteries, in particular to a battery thermal management system dynamic control method based on deep learning, which comprises the following steps: training a preset neural network based on battery management data collected in history, and obtaining error distribution in the training process; formulating a fuzzy control model; the real-time battery temperature is input into a preset neural network to obtain a battery temperature predicted value, battery temperature predicted value distribution is obtained according to the error distribution and the battery temperature predicted value, and the real-time cooling liquid flow and the real-time fan rotating speed are integrated according to the error distribution to obtain a cooling liquid flow corrected value and a fan rotating speed corrected value; and the control of battery thermal management is completed. According to the technical scheme, the battery temperature error range can be reduced, the precision of a battery temperature detection result is improved, and then the control precision of a thermal management system is improved.

Description

Technical Field

[0001] The present invention relates to the field of batteries. More specifically, the present invention relates to a dynamic control method for a battery thermal management system based on deep learning. Background Art

[0002] As a core component of electric vehicles, the performance of batteries directly determines the vehicle's endurance, safety, and overall user experience. In the battery system, the thermal management problem has always been a challenge that cannot be ignored. A large amount of heat is generated during the charging and discharging process of the battery. Especially under high-power workloads, the temperature of the battery will rise significantly. If the temperature of the battery cannot be effectively controlled, it will affect the performance and lifespan of the battery, and may even lead to safety hazards such as fire and explosion. Therefore, the design and optimization of the battery thermal management system have become an essential part of electric vehicle technology.

[0003] Currently, battery thermal management technologies usually adopt liquid cooling or air cooling systems, which absorb and remove the heat generated by the battery by adjusting the flow of the cooling medium. However, traditional thermal management systems often have problems such as untimely adjustment, uneven heat distribution, and system response lag in practical applications, resulting in inaccurate temperature control of the battery under different working conditions.

[0004] There is a certain error in predicting the temperature, which means that the battery thermal management system designed based on the predicted temperature may not be able to fully and accurately respond to the actual temperature changes. Existing battery thermal management systems are usually designed based on the predicted temperature without considering the impact of prediction errors, which may lead to the risk of overheating or overcooling of the battery during actual operation, thus affecting the battery performance, lifespan, and even safety. Summary of the Invention

[0005] To solve the technical problem of inaccurate control of the battery temperature management system caused by not considering prediction errors, the present invention provides a dynamic control method for a battery thermal management system based on deep learning. The method includes: training a preset neural network based on battery management data collected in history and obtaining the error distribution during the training process; formulating a fuzzy control model, including: using the battery temperature, predicted temperature, and ambient temperature as control inputs, and using the coolant flow rate and fan speed as control outputs, formulating control rules based on the control inputs and control outputs to complete the construction of the fuzzy control model; inputting the real-time battery temperature into the preset neural network to obtain a battery temperature prediction value, obtaining a battery temperature prediction value distribution according to the error distribution and the battery temperature prediction value, inputting the real-time battery temperature, battery temperature prediction value distribution, and real-time ambient temperature into the fuzzy control model, outputting the real-time coolant flow rate and real-time fan speed, and integrating the real-time coolant flow rate and real-time fan speed respectively according to the error distribution to obtain a coolant flow rate correction value and a fan speed correction value to complete the control of the battery thermal management.

[0006] By combining historical battery management data and real-time monitoring information, precisely control the battery's thermal management system to ensure the battery operates stably within the optimal temperature range. Predict the battery temperature through a neural network, and finely adjust the coolant flow rate and fan speed in combination with the error distribution, which can dynamically adapt to the temperature fluctuations of the battery under different working conditions. It can not only respond to temperature changes in real time to avoid battery performance degradation or safety issues caused by overheating or too low temperature, but also adjust the system operation parameters according to environmental changes, optimize energy consumption, and improve the battery's lifespan, efficiency, and safety.

[0007] Preferably, the preset neural network is a BP network. The input of the BP network is battery management data, which includes the battery's current, voltage, soc value, and ambient temperature. The output of the BP network is the predicted value of the battery temperature, and the loss function of the model is the mean square error loss.

[0008] Preferably, the fuzzy control model includes: a battery temperature function, a predicted temperature function, an ambient temperature function, a coolant flow rate function, and a fan speed function. Among them, the battery temperature function includes a low-temperature membership function, a suitable-temperature membership function, and a high-temperature membership function.

[0009] Preferably, the low-temperature membership function satisfies the relational expression: , represents the low-temperature membership function, represents the temperature, represents the preset first temperature threshold.

[0010] Preferably, the suitable-temperature membership function satisfies the relational expression: , represents the suitable-temperature membership function, represents the temperature, represents the preset first temperature threshold, represents the preset second temperature threshold, represents the preset third temperature threshold, represents the preset fourth temperature threshold.

[0011] Preferably, the high-temperature membership function satisfies the relational expression: , represents the high-temperature membership function, represents the temperature, represents the preset fourth temperature threshold, represents the preset fifth temperature threshold, represents the preset sixth temperature threshold.

[0012] Preferably, the error distribution includes: obtaining all loss values during the training of a preset neural network, and for any loss value, taking the ratio of the number of occurrences of the loss value to the total number of training times as the loss value probability, traversing to obtain the loss value probability of each loss value, and obtaining the error distribution.

[0013] The error distribution can reflect the error characteristics and stability of the neural network at different training stages, thereby providing a more accurate error control basis for subsequent temperature prediction. By effectively utilizing the error distribution, targeted adjustments can be made for different training losses, thereby optimizing the accuracy and stability of battery temperature prediction, enhancing the adaptability of the battery thermal management system to temperature fluctuations, ensuring that the battery always remains within the ideal operating temperature range, avoiding overheating or low-temperature operation caused by prediction errors, and improving the overall performance and safety of the battery.

[0014] Preferably, the error distribution includes: obtaining the training set of a preset neural network, and performing clustering according to the fluctuations of the training set to obtain several clustering clusters, and one clustering cluster corresponds to one error distribution.

[0015] Through clustering, the error characteristics under different working conditions can be obtained, and then in actual applications, the most suitable error distribution can be automatically matched according to the real-time battery temperature, optimizing the adjustment of coolant flow rate and fan speed. This not only enables the battery thermal management system to respond more precisely to temperature changes in the face of various dynamic working conditions, avoiding temperature control instability caused by error fluctuations, but also effectively improves the operating efficiency and safety of the battery.

[0016] Preferably, the integrating the real-time coolant flow rate and the real-time fan speed according to the error distribution to obtain the coolant flow rate correction value and the fan speed correction value includes: using the loss value probability of each loss value in the error distribution as a weight, integrating the control data of the real-time coolant flow rate and the real-time fan speed to obtain a control correction value, and the control correction value includes the coolant flow rate correction value and the fan speed correction value; , represents the control correction value, represents the loss value of the loss value probability, represents the loss value corresponding to the control data of the real-time coolant flow rate and the real-time fan speed.

[0017] The weighted adjustment method based on the error distribution enables the system to more precisely control the addition of coolant and the response of the fan speed system in the face of temperature fluctuations and working condition changes, ensuring that the battery operates within the optimal operating temperature range, avoiding the occurrence of extreme situations such as overheating or low temperature, and improving the operating stability and safety of the battery.

[0018] Preferably, the integration of the real-time coolant flow rate and the real-time fan speed according to the error distribution to obtain the coolant flow rate correction value and the fan speed correction value includes: matching the real-time battery temperature to the clustering clusters, and using the error distribution of the clustering cluster containing the real-time battery temperature as the weight to integrate the control data of the real-time coolant flow rate and the real-time fan speed to obtain the coolant flow rate correction value and the fan speed correction value.

[0019] It can respond to the change of battery temperature in real time, enabling the cooling system to dynamically adjust the coolant flow rate and the fan speed according to the current temperature state and error characteristics, so as to achieve more accurate temperature control management. This not only optimizes the heat dissipation effect of the battery, prevents performance degradation caused by overheating or overcooling, but also improves the overall reliability and energy utilization efficiency of the system, ensuring that the battery can maintain the best operating temperature in various working environments.

[0020] Advantages of the present invention: The present invention uses historical data to train a neural network, which can accurately predict the battery temperature, and dynamically adjusts the coolant flow rate and the fan speed according to the real-time temperature, environmental conditions and temperature prediction distribution to ensure that the battery works within the best temperature range. Through the fuzzy control model, accurate adjustment of the battery temperature can be achieved, and reasonable responses can be made under different temperature states. In addition, the introduction of the error distribution enables the system to be corrected according to the probability of the loss value during the historical training process, thereby further improving the control effect and reducing the risk of temperature fluctuation and overheating. The present invention not only improves the accuracy and response speed of battery thermal management, but also enhances the adaptability and stability of the system under complex working conditions. Description of the drawings

[0021] Figure 1 is a flowchart of the dynamic control method of the battery thermal management system based on deep learning in the embodiment of the present invention. Detailed implementation manners

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments.

[0023] Next, the specific implementation manners of the present invention will be described in detail with reference to the drawings.

[0024] Refer to Figure 1 , the dynamic control method of the battery thermal management system based on deep learning includes steps S1 - S3, specifically as follows: S1: Train a preset neural network based on the battery management data collected in history, and obtain the error distribution during the training process.

[0025] In one embodiment, historical battery data is collected, including relevant data such as current, voltage, SOC (state of charge), ambient temperature, and battery temperature. In the data preprocessing stage, these data are first normalized to eliminate the differences between different dimensions and ensure the balanced influence of each feature on model training. Then, a neural network based on backpropagation (BP) is selected as the prediction model. The inputs of the BP network include the current, voltage, SOC value, and ambient temperature of the battery, and the output is the predicted value of the battery temperature.

[0026] To optimize the training effect of the model, the mean squared error loss is used as the loss function, and the parameters of the model are updated through the gradient descent algorithm. During the training process, when the loss value of the model drops below the set threshold or reaches the maximum number of training times, the training process will stop. Exemplarily, when the number of training times reaches 150 or the loss value is less than 0.0001, the training stops.

[0027] After the training is completed, the loss values of the BP network during each training are statistically analyzed to obtain the error distribution.

[0028] S2: Formulate a fuzzy control model.

[0029] In one embodiment, the battery temperature, predicted temperature, and ambient temperature are used as control inputs, and the coolant flow rate and fan speed are used as control outputs. Control rules are formulated based on the control inputs and control outputs. Exemplarily, when the predicted temperature is high, increase the coolant flow rate and fan speed; when the ambient temperature is high and the predicted temperature is moderate, do not adjust the coolant flow rate and fan speed.

[0030] There are membership functions for both control inputs and control outputs in the fuzzy control model, including the battery temperature function, predicted temperature function, ambient temperature function, coolant flow rate function, and fan speed function. Among them, the battery temperature function includes a low-temperature membership function, a suitable-temperature membership function, and a high-temperature membership function.

[0031] The low-temperature membership function satisfies the relational expression: , represents the low-temperature membership function, represents the temperature, represents the preset first temperature threshold.

[0032] The suitable-temperature membership function satisfies the relational expression: , represents the suitable-temperature membership function, represents the temperature, represents the preset first temperature threshold, represents the preset second temperature threshold, represents the preset third temperature threshold, Represents a preset fourth temperature threshold.

[0033] The high-temperature membership function satisfies the relational expression: , Represents the high-temperature membership function, Represents temperature, Represents a preset fourth temperature threshold, Represents a preset fifth temperature threshold, Represents a preset sixth temperature threshold.

[0034] The temperature range of the battery is well-known technology to those skilled in the art. Therefore, the first temperature threshold, the second temperature threshold, the third temperature threshold, the fourth temperature threshold, the fifth temperature threshold, and the sixth temperature threshold can all be set by those skilled in the art.

[0035] The predicted temperature function, the ambient temperature function, the coolant flow rate function, and the fan speed function can be directly constructed using existing membership functions.

[0036] By defining the membership functions for different temperature ranges, the model can more accurately reflect the battery conditions under different temperature ranges. The low-temperature, suitable-temperature, and high-temperature membership functions enable the battery management system to make appropriate responses to different temperature intervals, avoiding over-cooling or insufficient cooling situations. In the low-temperature interval, by reasonably adjusting the coolant flow rate and the fan speed, it is ensured that the battery is not affected under low-temperature conditions; in the suitable-temperature interval, it ensures that the battery is within the optimal operating temperature range and maintains stable performance; while in the high-temperature interval, it is automatically adjusted according to the preset thresholds to prevent the battery from overheating and extend the battery life.

[0037] Through fuzzy control, it can handle the complexity of temperature prediction errors and external environment changes during actual operation, improving the robustness and stability of the system. In addition, adopting the fuzzy control method can avoid the sudden response that may exist in traditional control methods, reduce unnecessary system fluctuations, optimize battery thermal management, and ensure the safety and performance of the battery under various working conditions.

[0038] S3: Input the real-time battery temperature into a preset neural network to obtain the battery temperature prediction value, obtain the battery temperature prediction value distribution according to the error distribution and the battery temperature prediction value, input the real-time battery temperature, the battery temperature prediction value distribution, and the real-time ambient temperature into the fuzzy control model, output the real-time coolant flow rate and the real-time fan speed, and respectively integrate the real-time coolant flow rate and the real-time fan speed according to the error distribution to obtain the coolant flow rate correction value and the fan speed correction value, completing the control of battery thermal management.

[0039] In one embodiment, for all the loss values obtained according to step S1, the ratio of the number of occurrences of a loss value to the total number of training times is used as the loss value probability. By traversing, the loss value probability of each loss value is obtained, and an error distribution is obtained.

[0040] By analyzing the loss value distribution of the neural network during the training process, the error characteristics and stability of the model at different training stages can be deeply understood. The acquisition of the error distribution makes the performance evaluation of the model more intuitive, which can guide the subsequent parameter tuning and optimization process, thereby improving the generalization ability and practical application effect of the model.

[0041] The real-time battery temperature, the distribution of battery temperature prediction values, and the real-time ambient temperature are input into the fuzzy control model to output the real-time coolant flow rate and the real-time fan speed. The loss value probability of each loss value within the error distribution is used as a weight to integrate the control data of the real-time coolant flow rate and the real-time fan speed, and a control correction value is obtained. The control correction value includes a coolant flow rate correction value and a fan speed correction value.

[0042] , denotes the control correction value, denotes the loss value of the loss value probability, denotes the loss value corresponding to the control data of the real-time coolant flow rate and the real-time fan speed.

[0043] Using the loss value probability of each loss value in the error distribution as a weight to perform weighted integration on the control data of the coolant flow rate and the fan speed can more accurately reflect the error characteristics during the training process, so as to perform real-time refined adjustment on the cooling system according to the actual operating conditions. In this way, the model can provide appropriate control correction values under various working conditions, optimize the coolant flow rate and the fan speed, and ensure that the battery operates within a safe and efficient working temperature range.

[0044] Control the battery temperature according to the obtained coolant flow rate correction value and fan speed correction value.

[0045] In another embodiment, a training set of a preset neural network is obtained, and several clustering clusters are obtained by clustering according to the fluctuations of the training set. One clustering cluster corresponds to one error distribution.

[0046] Match the real-time battery temperature with the clustering clusters, and use the error distribution of the clustering cluster containing the real-time battery temperature as a weight to integrate the control data of the real-time coolant flow rate and the real-time fan speed to obtain a coolant flow rate correction value and a fan speed correction value.

[0047] Clustering is performed based on the fluctuations of a preset neural network training set, thereby establishing a corresponding error distribution for each cluster, and matching the most suitable cluster according to the real-time battery temperature. This method enables the control of the coolant flow rate and fan speed to be dynamically adjusted more precisely according to the changes in the battery temperature. By using the error distribution of the cluster as a weight for controlling data integration, real-time correction can be performed according to the error characteristics under different working conditions, optimizing the performance of the battery cooling system.

[0048] It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent shall be subject to the appended claims.

Claims

1. A dynamic control method for a battery thermal management system based on deep learning, characterized in that: include: Train a preset neural network based on historically collected battery management data and obtain the error distribution during the training process; Formulate a fuzzy control model, including: taking battery temperature, predicted temperature and ambient temperature as control inputs, taking coolant flow and fan speed as control outputs, formulating control rules based on control inputs and control outputs, and completing the construction of the fuzzy control model; The real-time battery temperature is input into the preset neural network to obtain the battery temperature prediction value, and the battery temperature prediction value distribution is obtained according to the error distribution and the battery temperature prediction value. The real-time battery temperature, the battery temperature prediction value distribution and the real-time ambient temperature are input into the fuzzy control model, and the real-time coolant flow and the real-time fan speed are output. The real-time coolant flow and the real-time fan speed are respectively integrated according to the error distribution to obtain the coolant flow correction value and the fan speed correction value, thereby completing the control of the battery thermal management.

2. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1, characterized in that: The preset neural network is a BP network, the input of the BP network is battery management data, the battery management data includes the battery current, voltage, soc value and ambient temperature, the output of the BP network is the battery temperature prediction value, and the loss function of the model is mean square error loss.

3. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1, characterized in that: The fuzzy control model includes: A battery temperature function, a predicted temperature function, an ambient temperature function, a coolant flow function and a fan speed function, wherein the battery temperature function includes a low temperature membership function, a suitable temperature membership function and a high temperature membership function.

4. The method for dynamic control of a battery thermal management system based on deep learning according to claim 3, characterized in that: The low temperature membership function satisfies the relationship: , represents the low temperature membership function, Indicates temperature, Indicates the preset first temperature threshold.

5. The method for dynamic control of a battery thermal management system based on deep learning according to claim 3, characterized in that: The suitable temperature membership function satisfies the relationship: , represents the suitable temperature membership function, Indicates temperature, Indicates the preset first temperature threshold, Indicates the preset second temperature threshold, Indicates the preset third temperature threshold, Indicates the preset fourth temperature threshold.

6. The method for dynamic control of a battery thermal management system based on deep learning according to claim 3, characterized in that: The high temperature membership function satisfies the relationship: , represents the high temperature membership function, Indicates temperature, Indicates the preset fourth temperature threshold, Indicates the preset fifth temperature threshold, Indicates the preset sixth temperature threshold.

7. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1, characterized in that: The error distribution includes: Get all the loss values ​​of the preset neural network during the training process. For any loss value, take the ratio of the number of times the loss value appears to the total number of training times as the loss value probability, traverse to obtain the loss value probability of each loss value, and obtain the error distribution.

8. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1, characterized in that: The error distribution includes: A training set of a preset neural network is obtained, and clustering is performed according to the fluctuation of the training set to obtain several clusters, where one cluster corresponds to one error distribution.

9. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1 or 7, characterized in that: The step of integrating the real-time coolant flow rate and the real-time fan speed according to the error distribution to obtain the coolant flow rate correction value and the fan speed correction value comprises: The loss value probability of each loss value in the error distribution is used as a weight, and the control data of the real-time coolant flow rate and the real-time fan speed are integrated to obtain a control correction value, which includes a coolant flow correction value and a fan speed correction value; , represents the control correction value, Indicates the loss value The probability of loss value, Indicates the loss value Corresponding control data of real-time coolant flow and real-time fan speed.

10. The method for dynamic control of a battery thermal management system based on deep learning according to claim 1 or 8, characterized in that: The step of integrating the real-time coolant flow rate and the real-time fan speed according to the error distribution to obtain the coolant flow rate correction value and the fan speed correction value comprises: The real-time battery temperature is matched to the cluster, the error distribution of the cluster containing the real-time battery temperature is used as the weight, and the control data of the real-time coolant flow rate and the real-time fan speed are integrated to obtain the coolant flow rate correction value and the fan speed correction value.

Images