Battery fast charging control information prediction model training method and device, computer device, readable storage medium and program product

Abstract

The application relates to a battery fast charging control information prediction model training method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: taking state information at a current moment as target state information, inputting the target state information into a deep reinforcement learning network of a target battery to obtain charging control information of the target battery at the current moment; performing single-step simulation based on the charging control information at the current moment to determine state information of the target battery at a next moment and a target return value at the current moment, and updating the target state information based on the state information at the next moment; the target return value is generated based on negative electrode potential information of the target battery; and the deep reinforcement learning network is trained based on the state information, the target return value and the charging control information at each moment and state information at a next moment corresponding to each moment to obtain a battery fast charging control information prediction model. The method can realize fast charging.

Description

Technical Field

[0001] This application relates to the field of battery management technology, and in particular to a training method, apparatus, computer device, computer-readable storage medium, and computer program product for a battery fast charging control information prediction model. Background Technology

[0002] Traditional gasoline-powered vehicles can be recharged in just a few minutes, allowing for rapid energy replenishment. In contrast, electric vehicles still require significantly longer charging times, with slow charging taking up to several hours. This severely impacts the user's charging experience and hinders the widespread adoption of electric vehicles.

[0003] Therefore, there is an urgent need for a training method for a battery fast-charging control information prediction model that can achieve fast charging. Summary of the Invention

[0004] Therefore, it is necessary to provide a training method, apparatus, computer equipment, computer-readable storage medium, and computer program product for a battery fast charging control information prediction model that can achieve fast charging, addressing the aforementioned technical problems.

[0005] Firstly, this application provides a training method for a battery fast charging control information prediction model, including:

[0006] The current state information of the target battery is obtained and used as the target state information. The target state information is then input into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment.

[0007] A single-step simulation is performed based on the charging control information at the current moment to determine the state information of the target battery at the next moment and the target reward value at the current moment. The target state information is then updated based on the state information at the next moment, and the process returns to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment. The target reward value is generated based on the negative electrode potential information of the target battery.

[0008] Based on the state information, target reward value, and charging control information at each time point, as well as the state information of the next time point corresponding to each time point, the deep reinforcement learning network is trained to obtain a battery fast charging control information prediction model.

[0009] In one embodiment, the step of performing a single-step simulation based on the charging control information at the current moment to determine the state information of the target battery at the next moment and the target reward value at the current moment includes:

[0010] A single-step simulation is performed based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment;

[0011] Based on the state information and hidden information of the next moment, the target reward value of the target battery at the current moment is determined.

[0012] In one embodiment, the state information includes negative electrode potential information; or the hidden information includes negative electrode potential information; determining the target reward value of the target battery at the current moment based on the state information and hidden information at the next moment includes:

[0013] Based on the negative electrode potential information at the next moment, calculate the negative electrode potential violation penalty value of the target battery at the current moment;

[0014] Based on the state information and hidden information of the next moment, calculate the non-negative electrode potential violation penalty value of the target battery at the current moment;

[0015] The target reward value of the target battery at the current moment is calculated by weighting the penalty value for violation of the negative electrode potential and the penalty value for violation of the non-negative electrode potential.

[0016] In one embodiment, calculating the negative electrode potential violation penalty value of the target battery at the current moment based on the negative electrode potential information at the next moment includes:

[0017] If the charge of the target battery at the next moment is less than or equal to a preset charge threshold, the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment.

[0018] If the charge level of the target battery at the next moment is greater than a preset charge threshold, the preset penalty value will be used as the penalty value for the negative electrode potential violation of the target battery at the current moment.

[0019] In one embodiment, the deep reinforcement learning network includes a policy network, and the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment includes:

[0020] Based on the preset decay function, determine the exploration threshold corresponding to the current moment;

[0021] If the random number generated at the current moment is greater than or equal to the exploration threshold, then the target state information is input into the policy network to obtain the charging control information of the target battery at the current moment;

[0022] If the random number generated at the current moment is less than the exploration threshold, then the charging control information of the target battery at the current moment is generated based on the uniformized random action generation function.

[0023] In one embodiment, the deep reinforcement learning network includes a policy network and a Q-network. The step of training the deep reinforcement learning network based on the state information, target reward value, and charging control information at each time step, as well as the state information for the next time step corresponding to each time step, to obtain a battery fast charging control information prediction model includes:

[0024] For each moment in each charging process, a set of charging experience information is generated based on the state information, target reward value, and charging control information at that moment, as well as the state information of the next moment corresponding to that moment.

[0025] Under the premise of meeting the preset model training conditions, m sets of charging experience information are randomly selected every preset time period.

[0026] The m sets of charging experience information are input into the policy network to obtain the current predicted value of the charging control information corresponding to the m sets.

[0027] The strategy network is updated based on the m sets of charging experience information, the current predicted values ​​of the charging control information corresponding to the m sets, and the Q network.

[0028] When the preset training completion conditions are met, the current policy network is used as the battery fast charging control information prediction model.

[0029] Secondly, this application provides a battery fast charging method, including:

[0030] Obtain the status information of the target battery;

[0031] The state information is input into a pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery.

[0032] The target battery is charged according to the charging control information;

[0033] The battery fast charging control information prediction model is trained using the training method for the battery fast charging control information prediction model described in the first aspect above.

[0034] Thirdly, this application also provides a training device for a battery fast charging control information prediction model, comprising:

[0035] The first determining module is used to obtain the current state information of the target battery, use the current state information as the target state information, input the target state information into the deep reinforcement learning network of the target battery, and obtain the charging control information of the target battery at the current moment.

[0036] The second determining module is used to perform a single-step simulation based on the charging control information at the current moment, determine the state information of the target battery at the next moment and the target reward value at the current moment, update the target state information based on the state information at the next moment, and return to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment; the target reward value is generated based on the negative electrode potential information of the target battery;

[0037] The training module is used to train the deep reinforcement learning network based on the state information, target reward value, and charging control information at each time step, as well as the state information of the next time step corresponding to each time step, to obtain a battery fast charging control information prediction model.

[0038] Fourthly, this application also provides a battery fast charging device, comprising:

[0039] The acquisition module is used to acquire the status information of the target battery;

[0040] The prediction module is used to input the state information into a pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery.

[0041] A control module is used to control the charging of the target battery according to the charging control information;

[0042] The battery fast charging control information prediction model is trained using the training method for the battery fast charging control information prediction model described in the first aspect above.

[0043] Fifthly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps described in the first aspect or the second aspect above.

[0044] Sixthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps described in the first or second aspect above.

[0045] In a seventh aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps described in the first or second aspect above.

[0046] The training method, apparatus, computer equipment, computer-readable storage medium, and computer program product of the aforementioned battery fast charging control information prediction model acquire the current state information of the target battery and use the current state information as the target state information. The target state information is input into a deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment. Based on the charging control information at the current moment, a single-step simulation is performed to determine the state information of the target battery at the next moment and the target reward value at the current moment. The target state information is updated based on the state information at the next moment, and the process returns to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment. The target reward value is generated based on the negative electrode potential information of the target battery. The deep reinforcement learning network is trained according to the state information, target reward value, and charging control information at each moment, as well as the state information at the next moment corresponding to each moment, to obtain the battery fast charging control information prediction model. In this way, a closed-loop feedback framework is established for the interaction between the battery single-step simulation environment and the deep reinforcement learning network. The deep reinforcement learning network predicts the charging control information at the current moment based on the state information obtained from the battery single-step simulation environment. The battery single-step simulation environment simulates the state information of the battery at the next moment based on the charging control information output by the deep reinforcement learning network, and generates a target reward value based on the negative electrode potential information of the battery. Based on the state information, target reward value, charging control information, and the state information of the next moment corresponding to each moment, the deep reinforcement learning network is trained to obtain a battery fast charging control information prediction model. This model can predict the battery charging control information in real time based on the battery state information and adjust the battery charging rate in real time to achieve fast charging. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating the training method of a battery fast charging control information prediction model in one embodiment.

[0049] Figure 2 This is a flowchart illustrating the steps of a single-step simulation based on the charging control information at the current moment in one embodiment to determine the state information of the target battery at the next moment and the target reward value at the current moment.

[0050] Figure 3 This is a flowchart illustrating the steps of determining the target reward value of the target battery at the current moment based on the state information and hidden information of the next moment in one embodiment.

[0051] Figure 4 This is a flowchart illustrating the step of calculating the penalty value for a target battery's current negative electrode potential violation based on the negative electrode potential information at the next moment, as shown in one embodiment.

[0052] Figure 5 This is a flowchart illustrating the steps in one embodiment of inputting target state information into a deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment.

[0053] Figure 6 This is a flowchart illustrating the steps of training a deep reinforcement learning network to obtain a battery fast charging control information prediction model based on the state information, target reward value, and charging control information at each time point, as well as the state information of the next time point corresponding to each time point.

[0054] Figure 7 This is a schematic diagram of a power battery fast charging control strategy framework based on deep reinforcement learning in one embodiment;

[0055] Figure 8 This is a schematic diagram of the current change curve during the charging control process using a battery fast charging control information prediction model in one embodiment.

[0056] Figure 9 This is a schematic diagram of the voltage change curve during the charging control process using a battery fast charging control information prediction model in one embodiment.

[0057] Figure 10 This is a schematic diagram of the change curve of the negative electrode potential during the charging control process using a battery fast charging control information prediction model in one embodiment.

[0058] Figure 11 This is a schematic diagram of the SOC change curve during the charging control process using a battery fast charging control information prediction model in one embodiment.

[0059] Figure 12 This is a schematic diagram of the temperature change curve during the charging control process using a battery fast charging control information prediction model in one embodiment.

[0060] Figure 13 This is a flowchart illustrating a battery fast charging method in one embodiment;

[0061] Figure 14 This is a structural block diagram of a training device for a battery fast charging control information prediction model in one embodiment.

[0062] Figure 15 This is a structural block diagram of a battery fast charging device in one embodiment;

[0063] Figure 16 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0065] In one embodiment, such as Figure 1 As shown, a training method for a battery fast charging control information prediction model is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, or to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. In this embodiment, the method includes the following steps:

[0066] Step 101: Obtain the current state information of the target battery and use the current state information as the target state information. Input the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment.

[0067] In this embodiment, the target battery is a battery to be charged or fast-charged, and can be a battery for an electric vehicle. The target battery can be a lithium-ion battery, such as a lithium iron phosphate battery, a nickel-cobalt-manganese (NCM) series ternary battery, or a nickel-cobalt-aluminum (NCA) series ternary battery. The deep reinforcement learning network can be a Soft Actor-Critic (SAC) network, a Deep Deterministic Policy Gradient (DDPG) network, or a Twin Delayed Deep Deterministic Policy Gradient (TD3) network. The deep reinforcement learning network includes an Actor subnetwork and a Critic subnetwork. The Actor subnetwork can be a fully connected neural network, a time-series network, or a convolutional neural network. The Critic subnetwork can be a fully connected neural network, a time-series network, or a convolutional neural network. State information is used to represent the state of the target battery, including at least one of battery temperature, State of Charge (SOC), and terminal voltage, and may also include negative electrode potential information. Battery temperature refers to the internal temperature of the battery. Battery capacity indicates the percentage of usable charge relative to the nominal capacity, usually expressed as a percentage. Negative electrode potential information indicates the negative electrode potential state. Charging control information includes, but is not limited to, the charging rate, and may also include cooling flow rate. This enables fast charging control with coordinated electro-thermal response.

[0068] In one example, the terminal uses state information as the input state to the deep reinforcement learning network and charging control information as the action of the deep reinforcement learning network, thus obtaining a state-action pair for the deep reinforcement learning network. For instance, the terminal uses battery temperature, battery level, and terminal voltage as the input state to the deep reinforcement learning network and charging rate as the action of the deep reinforcement learning network, thus obtaining a state-action pair for the deep reinforcement learning network.

[0069] Step 102: Perform a single-step simulation based on the charging control information at the current moment to determine the state information of the target battery at the next moment and the target reward value at the current moment. Update the target state information based on the state information at the next moment and return to input the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment.

[0070] The target return value is generated based on the negative electrode potential information of the target battery.

[0071] In this embodiment, the terminal performs a single-step simulation based on the current charging control information within a pre-established single-step simulation environment of the target battery. This simulation determines the target battery's state information for the next moment and the target reward value for the current moment. Then, the terminal updates the target state information to the state information for the next moment and returns to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the current charging control information of the target battery.

[0072] In one example, the terminal builds a battery model of the target battery. Then, the terminal performs model state initialization processing on the battery model to obtain a state-initialized battery model. Next, based on the state-initialized battery model, the terminal establishes a single-step simulation environment for the target battery. Specifically, the terminal discretizes the state equations of the state-initialized battery model according to a preset single-step simulation time to obtain the single-step simulation environment for the target battery. The battery model includes, but is not limited to, neural network models, electrochemical-thermal-lithium plating coupling models, reduced-order electrochemical models, and hierarchical equivalent circuit models. The battery model includes a main reaction model and a side reaction model.

[0073] In one example, the terminal employs a data-driven model to fuse mechanistic information, establishing a battery negative electrode potential state estimation model based on a physically constrained neural network. This model serves as the battery model. By generating battery temperature, charge, terminal voltage, and negative electrode potential data through the mechanistic model, while satisfying kinetic constraints, the accuracy of the battery model and the battery state information obtained from simulation are improved. This, in turn, enhances the accuracy of the battery fast-charging control information prediction model, ultimately improving the accuracy and efficiency of battery charging.

[0074] In one example, the terminal fully charges the target battery at a preset initial ambient temperature, then discharges it completely, and then charges it back to its initial charge level. The terminal then allows the target battery to rest for a preset time, waiting for it to reach an equilibrium state. The terminal records the battery temperature, charge level, and terminal voltage at this point as the target battery's initial state information, while the negative electrode potential and whether the target charge level has been reached are hidden information. The initial ambient temperature includes, but is not limited to, low temperature, room temperature, and high temperature. The initial charge level includes, but is not limited to, 10%, 20%, 30%, 40%, and 50%.

[0075] In one embodiment, the terminal establishes an electrochemical-thermal-lithium plating coupling model for the high-energy-density lithium-ion battery LG INR 21700 M50, with a positive electrode composed of 80% nickel, 10% cobalt, and 10% manganese (NCM811) and a silicon-carbon (SiOx+C) negative electrode. The terminal then initializes the battery model's parameters using the publicly available parameter set provided by the Python Battery Mathematical Modelling (pybamm) platform. Finally, the terminal verifies the validity of the parameter-initialized battery model, obtaining a validated battery model.

[0076] In one example, the terminal tests the single-step simulation environment of the target battery to verify the effectiveness of establishing the single-step simulation environment of the target battery.

[0077] Step 103: Based on the state information, target reward value, and charging control information at each time point, as well as the state information of the next time point corresponding to each time point, train a deep reinforcement learning network to obtain a battery fast charging control information prediction model.

[0078] In this embodiment, for each moment in the charging process, the terminal generates a set of charging experience information based on the state information, target reward value, and charging control information at that moment, as well as the state information for the next moment. Under preset model training conditions, the terminal randomly selects m sets of charging experience information every preset time interval. Then, the terminal updates the deep reinforcement learning network based on the m sets of charging experience information. When a preset training completion condition is met, the terminal determines the battery fast charging control information prediction model based on the current deep reinforcement learning network. The training completion condition can be whether the number of network updates exceeds a preset network update threshold.

[0079] In the training method of the aforementioned battery fast-charging control information prediction model, a closed-loop feedback framework is established that integrates a single-step battery simulation environment and a deep reinforcement learning network. The deep reinforcement learning network predicts the charging control information for the current moment based on the state information obtained from the single-step battery simulation environment. The single-step battery simulation environment then simulates the battery's state information for the next moment based on the charging control information output by the deep reinforcement learning network, and generates a target reward value based on the battery's negative electrode potential information. Based on the state information, target reward value, charging control information, and the state information for the next moment at each moment, the deep reinforcement learning network is trained to obtain the battery fast-charging control information prediction model. This model can predict battery charging control information in real time based on battery state information and adjust the battery charging rate accordingly, achieving fast charging. Furthermore, this method employs a closed-loop feedback optimization method based on the interaction between the battery environment and the intelligent agent, adjusting the battery charging rate in real time, which is more consistent with actual conditions and improves the real-time performance and stability of the charging strategy, thus enhancing the accuracy of fast charging. Moreover, this method, based on the deep reinforcement learning framework and the self-learning capability of neural networks, optimizes the battery fast-charging strategy, improving the adaptive capability of the charging strategy. Furthermore, this method generates the target reward value based on the battery's negative electrode potential information. It considers not only external battery characteristics but also the impact of the internal negative electrode potential state on the charging rate, thus avoiding lithium plating side reactions and the accelerated capacity degradation caused by them. This improves the safety of fast charging and slows down battery capacity decay. The developed fast charging strategy ensures both charging safety and durability. Moreover, compared to traditional charging protocols developed for specific batteries under specific conditions, this method uses a deep reinforcement learning network—the charging agent—to interact with the battery environment, learning and optimizing the battery's fast charging strategy. This eliminates the need for extensive testing and calibration, solving the problems of long testing times, high costs, and lack of critical state information. It also has a wide range of applicable batteries and good transferability.

[0080] In one exemplary embodiment, such as Figure 2 As shown, the process of determining the target battery's state information at the next moment and the target reward value at the current moment through single-step simulation based on the charging control information at the current moment includes the following steps:

[0081] Step 201: Perform a single-step simulation based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment.

[0082] In this embodiment, the hidden information is also used to represent the state of the target battery, but it is not used as input to the deep reinforcement learning network. The state information includes negative electrode potential information, or the hidden information includes negative electrode potential information. For example, the state information includes battery temperature, charge level, terminal voltage, and negative electrode potential information, while the hidden information includes whether the target charge level has been reached. Alternatively, the state information includes battery temperature, charge level, and terminal voltage, while the hidden information includes negative electrode potential information and whether the target charge level has been reached.

[0083] Step 202: Determine the target reward value of the target battery at the current moment based on the state information and hidden information of the next moment.

[0084] In this embodiment, the terminal calculates the target reward value of the target battery at the current moment based on a preset reward function, the state information of the next moment, and hidden information. The target reward value at the current moment is the reward value for executing the charging control information at the current moment. The reward function includes, but is not limited to, a negative electrode potential violation penalty function.

[0085] In the training method of the aforementioned battery fast charging control information prediction model, a single-step simulation is performed based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment; based on the state information and hidden information at the next moment, the target reward value of the target battery at the current moment is determined. In this way, distinguishing the information used to calculate the target reward value into state information that is used as input to the deep reinforcement learning network and hidden information that is not used as input to the deep reinforcement learning network can reduce the size and computational cost of the deep reinforcement learning network.

[0086] In one exemplary embodiment, the state information includes negative electrode potential information. Alternatively, the hidden information may include negative electrode potential information. Figure 3 As shown, the specific process of determining the target reward value of the target battery at the current moment based on the state information and hidden information of the next moment includes the following steps:

[0087] Step 301: Calculate the penalty value for violation of the negative electrode potential of the target battery at the current moment based on the negative electrode potential information at the next moment.

[0088] In this embodiment of the application, the terminal calculates the negative electrode potential violation penalty value of the target battery at the current moment based on the negative electrode potential information at the next moment and the negative electrode potential violation penalty function included in the preset reward function.

[0089] Step 302: Calculate the penalty value for non-negative electrode potential violation of the target battery at the current moment based on the state information and hidden information of the next moment.

[0090] In this embodiment, the terminal calculates the non-negative electrode potential violation penalty value of the target battery at the current moment based on the state information and hidden information of the next moment, as well as the non-negative electrode potential violation penalty function included in the preset reward function. The non-negative electrode potential violation penalty function includes, but is not limited to, a temperature violation penalty function, a shortfall in target charge violation penalty function, and a terminal voltage violation penalty function. The non-negative electrode potential violation penalty value includes, but is not limited to, the temperature violation penalty function value, the shortfall in target charge violation penalty function value, and the terminal voltage violation penalty function value. The reward function includes, but is not limited to, the negative electrode potential violation penalty function, the temperature violation penalty function, the shortfall in target charge violation penalty function, and the terminal voltage violation penalty function. The non-negative electrode potential violation penalty function corresponds to the state information and hidden information.

[0091] In one example, the non-negative electrode potential violation penalty function value includes, but is not limited to, the temperature violation penalty function value, the shortfall in target charge violation penalty function value, and the terminal voltage violation penalty function value. The terminal calculates the current temperature violation penalty function value, shortfall in target charge violation penalty function, and terminal voltage violation penalty function value of the target battery based on the battery temperature, charge level, terminal voltage, and whether the target charge level has been reached at the next moment, as well as the temperature violation penalty function, shortfall in target charge violation penalty function, and terminal voltage violation penalty function included in the preset reward function.

[0092] Step 303: Calculate the target reward value of the target battery at the current moment by weighting the penalty value for violation of negative electrode potential and the penalty value for violation of non-negative electrode potential.

[0093] In this embodiment, the terminal performs a weighted summation based on the negative electrode potential violation penalty value, the negative electrode potential weight, the non-negative electrode potential violation penalty value, and the weight corresponding to each non-negative electrode potential violation penalty value, to obtain the target reward value of the target battery at the current moment.

[0094] In one embodiment, step 303 can be represented as:

[0095] reward=reward T *w T +reward F *w F +reward Vt *w Vt +reward phi *w phi

[0096] Where reward is the target return value, reward T w represents the temperature violation penalty function value. T The reward is the temperature weight corresponding to the temperature violation penalty function value. F For failing to reach the target power level, the penalty function value is violated, wT The reward is the weight of the amount of electricity for which the target amount of electricity is not reached, which corresponds to the penalty function value. Vt w is the value of the penalty function for terminal voltage violation. Vt The reward is the weight of the terminal voltage corresponding to the terminal voltage violation penalty function value. phi w is the penalty value for violating the negative electrode potential. phi The negative electrode potential weight is used. For example, the temperature weight is 1, the charge weight is 2, the terminal voltage weight is 1, and the negative electrode potential weight is 3.

[0097] In the training method of the aforementioned battery fast-charging control information prediction model, the penalty value for the current negative electrode potential violation of the target battery is calculated based on the negative electrode potential information at the next time step; the penalty value for the current non-negative electrode potential violation of the target battery is calculated based on the state information and hidden information at the next time step; and the target reward value for the current target battery is calculated by weighting the penalty values ​​for negative electrode potential violation and non-negative electrode potential violation. Thus, the target reward function includes multiple penalty functions, and the target reward value is generated not only based on the battery negative electrode potential information but also based on battery temperature, SOC, and terminal voltage. By comprehensively considering multiple state constraints, the fast-charging strategy of the battery is comprehensively optimized, solving the problem of incomplete constraints. This avoids thermal runaway and accelerated degradation caused by fast charging, ensuring charging speed while also guaranteeing battery safety and durability during fast charging, providing strong support for safe and rapid energy replenishment of electric vehicles. Furthermore, the negative electrode potential information can be used as both the input state of the charging agent in the deep reinforcement learning network and as penalty information hidden in the single-step charging reward. By using the negative electrode potential state as a hidden constraint condition for the charging agent, rather than as the agent's state input, excessive state information can be avoided from affecting policy training, thus reducing the size and computational cost of the deep reinforcement learning network.

[0098] In one exemplary embodiment, such as Figure 4 As shown, the specific process of calculating the penalty value for the current negative electrode potential violation of the target battery based on the negative electrode potential information at the next moment includes the following steps:

[0099] Step 401: If the charge of the target battery at the next moment is less than or equal to a preset charge threshold, calculate the negative electrode potential violation penalty value of the target battery at the current moment based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment.

[0100] In this embodiment, if the target battery's charge level at the next moment is less than or equal to a preset charge threshold, the terminal uses the absolute value of the difference between the negative electrode potential information at the next moment and the preset threshold value as the negative electrode potential critical distance. Then, the terminal uses the product of a preset negative electrode potential penalty coefficient and the negative electrode potential critical distance as the negative electrode potential violation penalty value of the target battery at the current moment. The target battery's state information at the next moment may include the target battery's charge level at the next moment. The threshold value can be a number slightly larger than 0, for example, 0.01. The charge threshold can be 0.45, and the negative electrode potential penalty coefficient can be -10.

[0101] In one embodiment, the terminal calculates the penalty value for the violation of the negative electrode potential of the target battery at the current moment, which can be expressed as: penalty value for violation of negative electrode potential = penalty coefficient * abs (negative electrode potential value - critical value), where abs() is the absolute value.

[0102] Step 402: If the charge of the target battery at the next moment is greater than the preset charge threshold, the preset penalty value is used as the penalty value for the violation of the negative electrode potential of the target battery at the current moment.

[0103] In this embodiment of the application, the penalty value is a fixed penalty value, which can be -1.

[0104] In the training method of the aforementioned battery fast charging control information prediction model, when the target battery's charge level at the next moment is less than or equal to a preset charge threshold, the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment. When the target battery's charge level at the next moment is greater than the preset charge threshold, the preset penalty value is used as the negative electrode potential violation penalty value of the target battery at the current moment. Thus, this method uses negative electrode potential driving in the low SOC range, achieving faster charging speeds when the negative electrode potential approaches the critical value. In the high SOC range, it uses target SOC driving, avoiding exceeding the cutoff voltage, forming a dual-drive fast charging strategy that further improves battery charging speed.

[0105] In one exemplary embodiment, such as Figure 5 As shown, the deep reinforcement learning network includes a policy network. The specific process of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment includes the following steps:

[0106] Step 501: Determine the exploration threshold corresponding to the current moment based on the preset decay function.

[0107] In this embodiment, the terminal determines the function value corresponding to the current moment based on a preset decay function, and uses this function value as the exploration threshold corresponding to the current moment. The decay function can be a linear decay function. The exploration threshold is used to determine the method of generating the deep reinforcement learning network action (i.e., charging control information).

[0108] Step 502: If the random number generated at the current moment is greater than or equal to the exploration threshold, then the target state information is input into the policy network to obtain the charging control information of the target battery at the current moment.

[0109] Step 503: If the random number generated at the current moment is less than the exploration threshold, then the charging control information of the target battery at the current moment is generated based on the uniformized random action generation function.

[0110] In the training method of the aforementioned battery fast-charging control information prediction model, an exploration threshold corresponding to the current moment is determined based on a preset decay function. If the random number generated at the current moment is greater than or equal to the exploration threshold, the target state information is input into the policy network to obtain the charging control information of the target battery at the current moment. If the random number generated at the current moment is less than the exploration threshold, the charging control information of the target battery at the current moment is generated based on a uniformly distributed random action generation function. This establishes a battery single-step simulation environment and a deep reinforcement learning network interaction framework. Initially, a uniformly distributed random action generation function is used to enhance the exploration capability of the deep reinforcement learning network. As the actor network in the deep reinforcement learning network is continuously optimized, the probability of randomly generated actions is reduced, while the probability of actions generated using the actor network is increased. This balances the exploration and utilization of the deep reinforcement learning network. Further optimization of the deep reinforcement learning network's action network can improve the training effect and accuracy of the battery fast-charging control information prediction model.

[0111] In one exemplary embodiment, such as Figure 6 As shown, the deep reinforcement learning network includes a policy network and a Q-network. The specific process of training the deep reinforcement learning network to obtain a battery fast-charging control information prediction model, based on the state information, target reward value, and charging control information at each time step, as well as the state information for the next time step, includes the following steps:

[0112] Step 601: For each moment in each charging process, generate a set of charging experience information based on the state information, target reward value, and charging control information at that moment, as well as the state information of the next moment corresponding to that moment.

[0113] In this embodiment, for each moment in each charging process, the terminal generates a set of charging experience information, including the state information, target reward value, and charging control information for that moment, as well as the state information for the next moment. The process of charging the battery from its initial charge to the target charge is called a charging process, also known as an episode. A charging process includes at least one charging time step.

[0114] In one example, for each moment in each charging process, the terminal combines the state information of that moment, the target reward value and charging control information, and the state information of the next moment corresponding to that moment into a set of charging experience information.

[0115] In another example, for each moment in each charging process, the terminal constructs a set of charging experience information, including the state information of that moment, the target reward value and charging control information, the state information of the next moment corresponding to that moment, and other information for that moment. This other information includes, but is not limited to, whether the current round has exceeded the preset maximum number of charging steps and whether the current round has charged to the target level.

[0116] Step 602: Under the condition of meeting the preset model training conditions, randomly select m sets of charging experience information every preset time period.

[0117] In this embodiment, after each set of charging experience information is generated, the terminal saves the set of charging experience information to the replay buffer. Under preset model training conditions, the terminal randomly selects m sets of charging experience information from the replay buffer every preset time period. Here, m is a positive integer. The model training condition can be that the number of sets of charging experience information reaches a preset threshold. The m sets of charging experience information can be charging experience information from different rounds.

[0118] Step 603: Input the charging experience information of m groups into the strategy network to obtain the current predicted value of the charging control information corresponding to m groups.

[0119] In this embodiment of the application, the policy network is the actor subnetwork and the Q network is the critic subnetwork.

[0120] Step 604: Update the strategy network based on the charging experience information of m groups, the current predicted value of the charging control information corresponding to m groups, and the Q network.

[0121] In this embodiment, the terminal calculates a target state action value based on m sets of charging experience information via a Q-network. The terminal also calculates a first state action value and a second state action value based on the m sets of charging experience information and the current predicted values ​​of the corresponding charging control information. The terminal then uses the minimum of the first and second state action values ​​as the predicted state action value. Finally, the terminal updates the policy network based on the difference between the target state action value and the predicted state action value.

[0122] Step 605: When the preset training completion conditions are met, the current policy network is used as the battery fast charging control information prediction model.

[0123] In this embodiment of the application, when the preset training completion conditions are met, the terminal stops updating the policy network and uses the current policy network as the battery fast charging control information prediction model.

[0124] In the training method of the aforementioned battery fast-charging control information prediction model, for each moment in the charging process, a set of charging experience information is generated based on the state information, target reward value, and charging control information at that moment, as well as the state information of the next moment corresponding to that moment. Under the condition of satisfying the preset model training conditions, m sets of charging experience information are randomly selected every preset time period. The m sets of charging experience information are input into the policy network to obtain the current predicted value of the charging control information corresponding to the m sets. Based on the m sets of charging experience information, the current predicted value of the charging control information corresponding to the m sets, and the Q network, the policy network is updated. When the preset training completion condition is met, the current policy network is used as the battery fast-charging control information prediction model. In this way, the charging control information prediction, the state information simulation of the next moment, and the calculation of the target reward value at the current moment are performed through the closed-loop feedback framework of battery single-step simulation environment and deep reinforcement learning network interaction, accumulating charging experience information. Under the condition of satisfying the preset model training conditions, m sets of charging experience information are periodically selected, combined with the value function, and the policy network is updated to realize the training of the battery fast-charging control information prediction model. The trained battery fast-charging control information prediction model has good real-time computing capabilities.

[0125] In one embodiment, a power battery fast-charging control strategy framework based on deep reinforcement learning, such as... Figure 7As shown, the fast-charging control strategy framework for the power battery includes a single-step simulation environment for the battery (referred to as the battery environment) and a deep reinforcement learning network. The deep reinforcement learning network is a flexible actor-critic network, comprising an actor sub-network (also known as the policy network) and a critic sub-network (also known as the Q network). The single-step simulation environment includes the battery's main reaction model and sub-reaction model. The Q network includes the Q1 network and the Q2 network. The terminal obtains the current state information of the target battery from the single-step simulation environment. The current state information is used as the target state information, which is then input into the deep reinforcement learning network of the target battery to obtain the current charging rate of the target battery. Then, the terminal calculates the charging rate based on the current charging rate. Perform a single-step simulation to determine the target battery's state information at the next moment. And the target reward value (also known as bonus) at the current moment. The process involves updating the target state information based on the state information at the next moment, and then returning the target state information to the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment. Target return value and charging rate The state information of the next moment corresponding to this moment. The terminal generates a set of charging experience information, including the information itself and other info at that moment. Each time a set of charging experience information is generated, the terminal saves it to the experience replay buffer (also known as the experience cache). Under preset model training conditions, the terminal randomly selects m sets of charging experience information from the experience replay buffer every preset time interval. Then, the terminal inputs these m sets of charging experience information into the policy network to obtain the current predicted values ​​of the charging control information corresponding to each of the m sets. Then, based on m sets of charging experience information, the current predicted values ​​of the corresponding charging control information for m sets, and the Q network, the terminal updates the Q1 network, Q2 network, and policy network, i.e., updates the deep reinforcement learning network. Specifically, the terminal updates the Q1 network... Updated to Q target network The network of Q2 Updated to Q target network The network of policy networks Update to policy target network .in, express The state at any given moment, express The charging rate at any given moment. Indicates in Perform actions at all times The return value, This indicates the state at time t+1. Info indicates other information, such as whether the current round has exceeded the preset maximum number of charging steps, or whether the current round has been charged to the target SOC. The charging strategy network needs to be optimized. The policy target network represents the network that plays a role in stabilizing the policy optimization process. and This represents the state-action value network. By using two networks, we can avoid overestimation of the state-action values. and This represents the target network of state action values, which can play a role in stabilizing the state action value network.

[0126] In one embodiment, the terminal analyzes and tests the trained Actor network, observing whether the battery temperature, terminal voltage, and negative electrode potential violate constraints at each charging step, and whether the battery ultimately charges from the initial SOC to the target SOC, thus judging the effectiveness of the battery fast charging control information prediction model. Then, the terminal sets different initial ambient temperatures and analyzes and tests the trained battery fast charging control information prediction model in a similar manner to the training method described above, judging the model's adaptability to different temperatures.

[0127] In one embodiment, 25°C is used as the initial ambient temperature and 20% as the initial State of Charge (SOC). During the charging control process using a battery fast-charging control information prediction model, the curves showing the changes in current, voltage, negative electrode potential, SOC, and temperature are respectively as follows: Figure 8-12 As shown, the charging time is 1260s, approximately 20 minutes, significantly reducing battery charging time. The battery temperature meets the maximum temperature constraint of 50℃, the battery voltage does not exceed 4.2V, the battery SOC is 0.2~0.8, reaching the set 80%, and the negative electrode potential is above the critical negative electrode potential of 0, avoiding accelerated capacity decay and safety issues triggered by lithium plating.

[0128] In one embodiment, such as Figure 13As shown, a battery fast charging method is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, or to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. It is understood that the application scenarios of this battery fast charging method and the training method of the battery fast charging control information prediction model described above can be the same or different, and this application does not limit them. In this embodiment, the method includes the following steps:

[0129] Step 1301: Obtain the status information of the target battery.

[0130] Step 1302: Input the state information into the pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery.

[0131] Step 1303: Control the charging of the target battery according to the charging control information.

[0132] The battery fast charging control information prediction model is trained using any one of the above-mentioned training methods for battery fast charging control information prediction models.

[0133] In this embodiment of the application, the terminal controls the target battery to charge at the current time according to the current charging rate.

[0134] In the above-mentioned fast charging method, the fast charging control information prediction model is trained using the training method of the above-mentioned fast charging control information prediction model to predict charging control information in real time, adjust the battery charging rate in real time, and control the target battery charging in real time, thereby achieving fast charging.

[0135] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0136] Based on the same inventive concept, this application also provides a training device for a battery fast charging control information prediction model, which is used to implement the training method for the battery fast charging control information prediction model described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more training device embodiments for the battery fast charging control information prediction model provided below can be found in the limitations of the training method for the battery fast charging control information prediction model described above, and will not be repeated here.

[0137] In one exemplary embodiment, such as Figure 14 As shown, a training device 1400 for predicting battery fast charging control information is provided, comprising: a first determining module 1410, a second determining module 1420, and a training module 1430, wherein:

[0138] The first determining module 1410 is used to obtain the current state information of the target battery, use the current state information as the target state information, input the target state information into the deep reinforcement learning network of the target battery, and obtain the charging control information of the target battery at the current moment.

[0139] The second determining module 1420 is used to perform a single-step simulation based on the charging control information at the current moment, determine the state information of the target battery at the next moment and the target reward value at the current moment, update the target state information based on the state information at the next moment, and return to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment; the target reward value is generated based on the negative electrode potential information of the target battery;

[0140] The training module 1430 is used to train the deep reinforcement learning network based on the state information, target reward value, and charging control information at each time point, as well as the state information of the next time point corresponding to each time point, to obtain a battery fast charging control information prediction model.

[0141] Optionally, the second determining module 1420 is specifically used for:

[0142] A single-step simulation is performed based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment;

[0143] Based on the state information and hidden information of the next moment, the target reward value of the target battery at the current moment is determined.

[0144] Optionally, the state information includes negative electrode potential information; or the hidden information includes negative electrode potential information; the second determining module 1420 is specifically used for:

[0145] Based on the negative electrode potential information at the next moment, calculate the negative electrode potential violation penalty value of the target battery at the current moment;

[0146] Based on the state information and hidden information of the next moment, calculate the non-negative electrode potential violation penalty value of the target battery at the current moment;

[0147] The target reward value of the target battery at the current moment is calculated by weighting the penalty value for violation of the negative electrode potential and the penalty value for violation of the non-negative electrode potential.

[0148] Optionally, the second determining module 1420 is specifically used for:

[0149] If the charge of the target battery at the next moment is less than or equal to a preset charge threshold, the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment.

[0150] If the charge level of the target battery at the next moment is greater than a preset charge threshold, the preset penalty value will be used as the penalty value for the negative electrode potential violation of the target battery at the current moment.

[0151] Optionally, the deep reinforcement learning network includes a policy network, and the first determining module 1410 is specifically used for:

[0152] Based on the preset decay function, determine the exploration threshold corresponding to the current moment;

[0153] If the random number generated at the current moment is greater than or equal to the exploration threshold, then the target state information is input into the policy network to obtain the charging control information of the target battery at the current moment;

[0154] If the random number generated at the current moment is less than the exploration threshold, then the charging control information of the target battery at the current moment is generated based on the uniformized random action generation function.

[0155] Optionally, the deep reinforcement learning network includes a policy network and a Q-network, and the training module 1430 is specifically used for:

[0156] For each moment in each charging process, a set of charging experience information is generated based on the state information, target reward value, and charging control information at that moment, as well as the state information of the next moment corresponding to that moment.

[0157] Under the premise of meeting the preset model training conditions, m sets of charging experience information are randomly selected every preset time period.

[0158] The m sets of charging experience information are input into the policy network to obtain the current predicted value of the charging control information corresponding to the m sets.

[0159] The strategy network is updated based on the m sets of charging experience information, the current predicted values ​​of the charging control information corresponding to the m sets, and the Q network.

[0160] When the preset training completion conditions are met, the current policy network is used as the battery fast charging control information prediction model.

[0161] Based on the same inventive concept, this application also provides a battery fast charging device for implementing the battery fast charging method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more battery fast charging device embodiments provided below can be found in the limitations of the battery fast charging method described above, and will not be repeated here.

[0162] In one exemplary embodiment, such as Figure 15 As shown, a battery fast charging device 1500 is provided, including: an acquisition module 1510, a prediction module 1520, and a control module 1530, wherein:

[0163] The acquisition module 1510 is used to acquire the status information of the target battery;

[0164] The prediction module 1520 is used to input the state information into a pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery.

[0165] The control module 1530 is used to control the charging of the target battery according to the charging control information;

[0166] The battery fast charging control information prediction model is trained using any one of the above-mentioned battery fast charging control information prediction model training methods.

[0167] The training device for the aforementioned battery fast charging control information prediction model and each module in the battery fast charging device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0168] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 16 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a training method for a battery fast-charging control information prediction model or a battery fast-charging method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0169] Those skilled in the art will understand that Figure 16The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0170] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0171] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0172] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0173] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0174] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0175] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0176] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A training method for a battery fast charging control information prediction model, characterized in that, The method includes: The current state information of the target battery is obtained and used as the target state information. The target state information is then input into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment. A single-step simulation is performed based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment; the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the negative electrode potential information at the next moment; the non-negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the state information and hidden information at the next moment; the target reward value of the target battery at the current moment is calculated by weighting the negative electrode potential violation penalty value and the non-negative electrode potential violation penalty value, and the target state information is updated based on the state information at the next moment. The process returns to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment; the state information includes negative electrode potential information; or the hidden information includes negative electrode potential information. Based on the state information, target reward value, and charging control information at each time point, as well as the state information of the next time point corresponding to each time point, the deep reinforcement learning network is trained to obtain a battery fast charging control information prediction model.

2. The method according to claim 1, characterized in that, The step of calculating the negative electrode potential violation penalty value of the target battery at the current moment based on the negative electrode potential information at the next moment includes: If the charge of the target battery at the next moment is less than or equal to a preset charge threshold, the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment. If the charge level of the target battery at the next moment is greater than a preset charge threshold, the preset penalty value will be used as the penalty value for the negative electrode potential violation of the target battery at the current moment.

3. The method according to claim 1, characterized in that, The deep reinforcement learning network includes a policy network. The step of inputting the target state information into the deep reinforcement learning network to obtain the current charging control information of the target battery includes: Based on the preset decay function, determine the exploration threshold corresponding to the current moment; If the random number generated at the current moment is greater than or equal to the exploration threshold, then the target state information is input into the policy network to obtain the charging control information of the target battery at the current moment; If the random number generated at the current moment is less than the exploration threshold, then the charging control information of the target battery at the current moment is generated based on the uniformized random action generation function.

4. The method according to claim 1, characterized in that, The deep reinforcement learning network includes a policy network and a Q-network. The process of training the deep reinforcement learning network based on the state information, target reward value, and charging control information at each time step, as well as the state information for the next time step, to obtain a battery fast charging control information prediction model includes: For each moment in each charging process, a set of charging experience information is generated based on the state information, target reward value, and charging control information at that moment, as well as the state information of the next moment corresponding to that moment. Under the premise of meeting the preset model training conditions, m sets of charging experience information are randomly selected every preset time period. The m sets of charging experience information are input into the policy network to obtain the current predicted value of the charging control information corresponding to the m sets. The strategy network is updated based on the m sets of charging experience information, the current predicted values ​​of the charging control information corresponding to the m sets, and the Q network. When the preset training completion conditions are met, the current policy network is used as the battery fast charging control information prediction model.

5. A method for fast charging a battery, characterized in that, The method includes: Obtain the status information of the target battery; The state information is input into a pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery. The target battery is charged according to the charging control information; The battery fast charging control information prediction model is trained using the training method for the battery fast charging control information prediction model according to any one of claims 1 to 4.

6. A training device for a battery fast charging control information prediction model, characterized in that, The device includes: The first determining module is used to obtain the current state information of the target battery, use the current state information as the target state information, input the target state information into the deep reinforcement learning network of the target battery, and obtain the charging control information of the target battery at the current moment. The second determining module is used to perform a single-step simulation based on the charging control information at the current moment to obtain the state information and hidden information of the target battery at the next moment; calculate the negative electrode potential violation penalty value of the target battery at the current moment based on the negative electrode potential information at the next moment; calculate the non-negative electrode potential violation penalty value of the target battery at the current moment based on the state information and hidden information at the next moment; calculate the target reward value of the target battery at the current moment by weighting the negative electrode potential violation penalty value and the non-negative electrode potential violation penalty value, and update the target state information based on the state information at the next moment, returning to the step of inputting the target state information into the deep reinforcement learning network of the target battery to obtain the charging control information of the target battery at the current moment; the target reward value is generated based on the negative electrode potential information of the target battery; the state information includes negative electrode potential information; or the hidden information includes negative electrode potential information; The training module is used to train the deep reinforcement learning network based on the state information, target reward value, and charging control information at each time step, as well as the state information of the next time step corresponding to each time step, to obtain a battery fast charging control information prediction model.

7. The apparatus according to claim 6, characterized in that, The second determining module is specifically used for: If the charge of the target battery at the next moment is less than or equal to a preset charge threshold, the negative electrode potential violation penalty value of the target battery at the current moment is calculated based on the preset negative electrode potential penalty coefficient and critical value, as well as the negative electrode potential information at the next moment. If the charge level of the target battery at the next moment is greater than a preset charge threshold, the preset penalty value will be used as the penalty value for the negative electrode potential violation of the target battery at the current moment.

8. A battery fast charging device, characterized in that, The device includes: The acquisition module is used to acquire the status information of the target battery; The prediction module is used to input the state information into a pre-trained battery fast charging control information prediction model to obtain the charging control information of the target battery. A control module is used to control the charging of the target battery according to the charging control information; The battery fast charging control information prediction model is trained using the training method for the battery fast charging control information prediction model according to any one of claims 1 to 4.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method described in any one of claims 1 to 4 or claim 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1 to 4 or claim 5.

Images