A layered cooperative control method of electric vehicle heat pump and PTC heating

Abstract

The application discloses a layered collaborative control method for electric vehicle heat pump and PTC heating, and belongs to the technical field of electric vehicle power battery thermal management, and the method comprises the following steps: constructing an electric vehicle battery thermal management system; defining an internal state vector, an actuator control vector and an external disturbance vector of the thermal management system as a thermal state in a power battery heating process; adopting a layered iterative dynamic programming algorithm in an upper controller to plan a battery temperature trajectory; adopting an SAC algorithm in a lower controller to perform multi-actuator collaborative trajectory tracking and output a control amount of the thermal management system; and executing the control amount output by the lower controller by the thermal management system. The application combines long-term temperature trajectory planning of the upper HIDP with continuous actuator control of the lower SAC, so that the controller has long-term planning capability and complex nonlinear actuator control capability, the control precision of the battery temperature is improved, and the control stability under a low-temperature working condition is improved.

Description

Technical Field

[0001] This invention relates to the field of low-temperature thermal management and intelligent control technology for electric vehicle power batteries, and in particular to a hierarchical coordinated control method for electric vehicle heat pumps and PTC heating. Background Technology

[0002] With the rapid development of the electric vehicle industry, the power battery, as the core component for energy storage and power output in vehicles, has a direct impact on the vehicle's range, charging and discharging efficiency, power performance, and lifespan due to its operating temperature. In low-temperature environments, the internal electrochemical reaction rate of lithium-ion power batteries decreases, leading to a reduction in usable capacity and power output. Therefore, rapid, stable, and low-energy-consumption heating of power batteries in low-temperature environments is a crucial task for the thermal management system of electric vehicles.

[0003] Currently, low-temperature heating methods for electric vehicle power batteries mainly include PTC electric heating, heat pump heating, liquid cooling / liquid heating circuits, and multi-heat source synergistic heating. Among these, PTC electric heaters have fast response speed, simple structure, and strong heating capacity, but they directly consume electrical energy, and continuous high-power operation will significantly increase the overall vehicle energy consumption. Heat pump systems can utilize ambient heat or system waste heat for heating, and have high energy efficiency, but their heating capacity and response speed decrease under low-temperature conditions, especially in environments at -10℃ and below, where their heating effect alone may be insufficient. To balance heating speed and energy consumption, a synergistic approach of heat pumps and PTC electric heaters is often used to achieve low-temperature heating of the power battery.

[0004] In terms of control methods, existing battery thermal management control strategies mainly include rule-based methods, classical control theory-based methods, model optimization-based methods, and reinforcement learning-based methods. Rule-based methods typically determine the operating states of the heat pump, PTC, and water pump based on ambient temperature, battery temperature, and preset thresholds. They are simple in structure and easy to implement in engineering, but the thresholds and control parameters largely rely on human experience, making them difficult to adapt to different ambient temperatures and complex operating conditions. Methods based on classical control such as PID can adjust actuator outputs according to temperature errors, but they struggle to globally coordinate the strong nonlinear coupling relationships between the heat pump, battery circuit, and PTC, and have limited ability to handle thermal inertia and hysteresis characteristics. Methods based on model predictive control (MPC) or dynamic programming (DP) can predict future states and optimize using system models, but directly incorporating multiple actuators into the optimization space results in a large computational load, posing challenges for online real-time applications and engineering deployment.

[0005] In recent years, deep reinforcement learning has gained increasing attention in the fields of electric vehicle energy management and thermal management due to its ability to handle continuous action spaces and nonlinear system control problems. The SAC algorithm achieves continuous action control through a policy network, a dual-commentator network, a target commentator network, and a maximum entropy mechanism, exhibiting good exploratory capabilities and training stability. However, if a single-layer SAC controller is used and trained directly with the battery target temperature as the control objective, the controller needs to simultaneously learn both "long-term temperature trajectory planning" and "actuator instantaneous control," resulting in significant training difficulty and a tendency for overly conservative or aggressive policies under low-temperature conditions. Therefore, there is an urgent need to design a hierarchical collaborative control method for electric vehicle heat pumps and PTC heating to address the problems in existing low-temperature battery thermal management, such as insufficient adaptability of rule-based control, difficulty in coordinating multi-actuator coupling, excessive learning burden on single-layer reinforcement learning controllers, high online computational cost of traditional optimization methods, and insufficient utilization of the complete thermal state. Summary of the Invention

[0006] This invention addresses the shortcomings of existing technologies by providing a hierarchical collaborative control method for heat pumps and PTC heating in electric vehicles. It combines upper-level HIDP long-term temperature trajectory planning with lower-level SAC continuous actuator control. The upper level utilizes a complete thermal state and parameter-updated thermal management environment model to perform full-condition temperature trajectory planning, outputting the target battery temperature trajectory. The lower layer uses this trajectory as a reference to perform multi-actuator collaborative tracking, enabling the controller to have both long-term planning capabilities and complex nonlinear actuator control capabilities, thereby improving battery temperature control accuracy, reducing heating energy consumption, and improving control stability and interpretability under low-temperature conditions.

[0007] The specific technical solution is as follows: A hierarchical coordinated control method for heat pump and PTC heating in electric vehicles includes: Step 1: Construct an electric vehicle battery thermal management system, including: a heat pump circuit, a battery heating circuit, a PTC electric heater, a power battery, a vehicle longitudinal dynamics heat generation module, and a controller module. The controller module includes an upper-level controller and a lower-level controller. Step 2: Define the internal state vector, actuator control vector, and external disturbance vector of the thermal management system as the thermal state during the power battery heating process; Step 3: The upper-level controller adopts a hierarchical iterative dynamic programming algorithm. Based on the future NEDC operating condition information of the electric vehicle, the ambient temperature and the thermal state of the current thermal management system, it plans a target battery temperature trajectory with the optimization goal of reducing energy consumption and temperature deviation. Step 4: The lower-level controller uses the SAC algorithm to track the temperature trajectory of the target battery. Combined with the current thermal state of the thermal management system, it outputs the control quantities of the thermal management system. The control quantities include: compressor speed, water pump speed and PTC heating power. Step 5: The thermal management system executes the control input output by the lower-level controller and feeds back the real-time status to the upper-level controller and the lower-level controller.

[0008] Furthermore, in step 1, the heat pump circuit includes at least a compressor, a condensing heat exchange component, an expansion component, and an evaporating heat exchange component; the battery heating circuit includes a battery circuit water pump, a heat exchange plate or a battery cold plate, a coolant pipeline, and a liquid replenishment / expansion component; a PTC electric heater is arranged in the battery heating circuit to provide auxiliary heating in low-temperature environments; the power battery and the battery heating circuit exchange heat through the battery heat exchange plate.

[0009] Furthermore, in step 2, the internal state vector of the thermal management system is defined as: ; in, For the core temperature of the battery, The surface temperature of the battery. This refers to the inlet temperature of the battery heating circuit. This refers to the outlet temperature of the battery heating circuit. The average temperature of the battery heating circuit. This is the mixing point temperature after heat exchange between the heat pump and the battery circuit. To take into account the dynamic heat supply after considering the dynamic response of the heat source; The actuator control vector is defined as: ; in, This refers to the compressor speed. The water pump speed, PTC heating power; The external disturbance vector is defined as: ; in, For ambient temperature, For vehicle speed.

[0010] Furthermore, in step 3, the stage cost of the hierarchical iterative dynamic programming algorithm is defined as: in, The target temperature for the battery. For the first Step-by-step electrical energy consumption , and These are respectively the weighting for temperature deviation, energy consumption, and terminal temperature deviation; The hierarchical iterative dynamic programming algorithm satisfies the executor amplitude constraint during planning: ; ; ; Simultaneously satisfy the actuator rate of change constraint: ; ; ; in, , , These represent the maximum compressor speed, maximum water pump speed, and maximum PTC heating power, respectively. , , These represent the maximum rate of change of compressor speed, the maximum rate of change of water pump speed, and the maximum rate of change of PTC heating power, respectively.

[0011] Furthermore, the hierarchical iterative dynamic programming algorithm uses a complete state node for merging during pruning and merging, and the complete state node is represented as: ; in, , and These represent the compressor speed, water pump speed, and PTC heating power at the previous moment, respectively.

[0012] Furthermore, the complete observations of the SAC algorithm are as follows: ; in, Indicates basic observations, This represents the normalized value of the upper-layer target temperature trajectory. This represents the observed temperature tracking error. ; Each normalized component is defined as follows: , , , , A symbol indicating that the electric heater is permitted to participate in heating; ; ; This represents the temperature value on the target battery temperature trajectory at the current moment.

[0013] Furthermore, the reward function of the SAC algorithm adopts a two-term structure, including the temperature trajectory tracking error and the total electrical power consumption, specifically expressed as: ; in, For temperature trajectory tracking weights, As power consumption weight, This represents the total power consumption of the compressor, water pump, and PTC electric heater.

[0014] Furthermore, the three-dimensional continuous action output by the SAC algorithm is as follows: ; Normalized control values ​​representing compressor speed, water pump speed, and heating power respectively. Control action; The three-dimensional continuous motion is mapped into control quantities for the compressor, water pump, and PTC electric heater: ; ; ; in, , These are the maximum and minimum compressor speeds, respectively. , These are the maximum and minimum values ​​of the water pump speed, respectively; , These represent the maximum and minimum values ​​of the PTC heating power, respectively.

[0015] Furthermore, the SAC algorithm employs two expert policy networks. When the ambient temperature When selecting the first expert strategy network, the first expert strategy network only allows the heat pump to heat alone and prohibits the PTC electric heater from starting. When the ambient temperature At ≤-10℃, the second expert strategy network is selected, which allows the PTC electric heater and the heat pump to participate in battery heating together.

[0016] This invention also provides a hierarchical collaborative control system for electric vehicle heat pumps and PTC heating, used to implement the above-mentioned hierarchical collaborative control method, comprising: The system comprises the following modules: a data acquisition module for acquiring the current thermal state, ambient temperature, and future NEDC operating condition information of the power battery thermal management system; an upper-level HIDP trajectory planning module for planning a target battery temperature trajectory based on a hierarchical iterative dynamic programming algorithm, taking into account future NEDC operating condition information, ambient temperature, and current thermal state, with the optimization goal of reducing energy consumption and temperature deviation; a lower-level SAC tracking control module for outputting control quantities for controlling the compressor, water pump, and PTC electric heater, using the target battery temperature trajectory as the tracking target and combining it with the current thermal state; and an execution module for applying the control quantities to the thermal management system and feeding back the real-time status to the upper-level HIDP trajectory planning module and the lower-level SAC tracking control module.

[0017] Compared with the prior art, the beneficial technical effects of the present invention are as follows: The upper-level controller of this invention employs a hierarchical iterative dynamic programming algorithm to plan a reasonable target battery temperature trajectory based on complete operating conditions and a thermal model. The lower-level controller uses the SAC algorithm to track the target battery temperature trajectory as a reference, effectively solving the problem of long-term deviation of battery temperature from the target under low-temperature conditions and further improving the tracking accuracy of the target temperature. Compared to directly using the SAC algorithm to learn the final actuator control from the real-time state, this invention introduces an upper-level reference temperature trajectory, providing a clear intermediate target for the SAC algorithm, allowing the lower-level controller to primarily learn the multi-actuator tracking rules, thus reducing training complexity. This invention also selects heat pump heating alone or heat pump and PTC mixed heating based on the ambient temperature, and continuously coordinates the power of the compressor, water pump, and PTC through the lower-level SAC, which is beneficial for improving the heating capacity at low temperatures and reducing unnecessary PTC energy consumption at 0℃. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a diagram of the overall architecture of the power battery layered collaborative heating control according to an embodiment of the present invention; Figure 2 This is a flowchart of the hierarchical collaborative control method according to an embodiment of the present invention; Figure 3 This is a flowchart of battery temperature trajectory planning according to an embodiment of the present invention; Figure 4 This is a flowchart of the temperature trajectory tracking training process according to an embodiment of the present invention; Figure 5This is a comparison chart of temperature and power consumption results in an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] The proposed power battery layered collaborative heating control architecture is as follows: Figure 1 As shown, this architecture includes future NEDC (New European Driving Cycle) operating condition information, current thermal state, control objectives, an upper-level HIDP battery temperature trajectory planner, a lower-level SAC multi-actuator trajectory tracking controller, and an electric vehicle battery thermal management system. The upper-level HIDP plans the temperature trajectory based on vehicle driving conditions, ambient temperature, and the complete thermal state; the lower-level SAC plans the temperature trajectory based on the target temperature trajectory output from the upper layer. The system outputs control quantities to the actuators based on real-time thermal status; the battery thermal management system executes control commands and provides feedback on the status.

[0023] The control method of the above-mentioned power battery layered collaborative heating control architecture is as follows: Figure 2 As shown, it specifically includes: Step 1: Build an electric vehicle battery thermal management system.

[0024] The electric vehicle battery thermal management system to which this invention applies mainly includes a heat pump circuit, a battery heating circuit, a PTC electric heater, a power battery, a vehicle longitudinal dynamics heat generation module, and a controller module. The heat pump circuit includes at least a compressor, a condensation heat exchange component, an expansion component, and an evaporation heat exchange component; the battery heating circuit includes a battery circuit water pump, a heat exchange plate or battery cold plate, coolant piping, and a coolant replenishment / expansion component; the PTC electric heater is arranged in the battery heating circuit to provide auxiliary heating in low-temperature environments; the power battery and the battery heating circuit exchange heat through the battery heat exchange plate.

[0025] The actuator output of the controller includes the compressor speed. Pump speed and PTC heating power Among them, when the ambient temperature When the temperature is >-10℃, the system preferably uses a heat pump-only heating mode; when When the temperature is ≤-10℃, the system preferably adopts a combined heating mode of heat pump and PTC electric heater.

[0026] Step 2: Define the internal state vector, actuator control vector, and external disturbance vector of the thermal management system as the thermal state during the power battery heating process.

[0027] The internal state vector of the thermal management system is defined as follows: ; in, For the core temperature of the battery, The surface temperature of the battery. This refers to the inlet temperature of the battery heating circuit. This refers to the outlet temperature of the battery heating circuit. The average temperature of the battery heating circuit. This is the mixing point temperature after heat exchange between the heat pump and the battery circuit. To take into account the dynamic heat supply after considering the dynamic response of the heat source; The actuator control vector is defined as: ; in, This refers to the compressor speed. The water pump speed, PTC heating power; The external disturbance vector is defined as: ; in, For ambient temperature, For vehicle speed.

[0028] The power battery thermal management system can also be abstracted as the following nonlinear discrete system: ; In this embodiment, the heat exchange between the battery and the environment can be expressed as: ; The change in battery temperature can be approximated as: ; in, For battery quality, For the specific heat capacity of the battery, The heat transferred from the battery heating circuit to the battery. The battery itself generates heat during vehicle operation. For simulation or control step size.

[0029] Step 3: Upper-layer HIDP battery temperature trajectory planning method.

[0030] The upper-level controller of this invention employs the HIDP method for battery temperature trajectory planning. Its purpose is not to directly generate final actuator control commands, but rather, based on a complete NEDC operating condition and parameter-updated thermal management environment model, to search for a superior battery temperature trajectory that balances heating effect and energy consumption, and to use this trajectory as a tracking reference for the lower-level SAC. Figure 3 As shown.

[0031] HIDP first reads future NEDC operating condition information, ambient temperature, and initial thermal state, and then predicts the thermal state based on a set of candidate actuator actions. Each candidate action is advanced in the battery thermal management environment model with updated parameters to obtain the next stage's thermal state, energy consumption, and temperature deviation. Subsequently, the cost of candidate nodes is calculated according to the objective function, and low-cost nodes are retained using complete state node discretization and pruning mechanisms to finally obtain the optimal battery temperature trajectory under all operating conditions. .

[0032] The stage cost of the upper-level HIDP can be defined as: in, The target temperature for the battery. For the first Step-by-step electrical energy consumption , and These are the weights for temperature deviation, energy consumption, and terminal temperature deviation, respectively.

[0033] The upper-level HIDP planning satisfies the actuator amplitude constraints: Simultaneously satisfy the actuator rate of change constraint: in, , , These represent the maximum compressor speed, maximum water pump speed, and maximum PTC heating power, respectively. , , These represent the maximum compressor speed change rate, the maximum water pump speed change rate, and the maximum PTC heating power change rate, respectively. Limiting the actuator change rate can reduce sudden changes in control quantities and improve system stability.

[0034] In this embodiment, the upper-layer HIDP uses complete state nodes for pruning and merging, rather than merging nodes solely based on the battery core temperature. A complete state node can be represented as: ; , and These represent the compressor speed, water pump speed, and PTC heating power at the previous moment, respectively, by introducing... , , , , By comparing the actuator state with the previous moment, we can avoid mistaking nodes with the same battery core temperature but different liquid circuit thermal states and heat source states as equivalent nodes, thereby improving the accuracy and stability of upper-level planning.

[0035] The target battery temperature trajectory output by the upper HIDP layer is defined as follows: ; in, This represents the optimal battery temperature trajectory obtained by HIDP during full-condition planning. In this embodiment, the lower-level SAC controller does not directly use the compressor, water pump, and PTC actuator sequence obtained by the upper-level HIDP, but only uses... This serves as a reference for temperature trajectory tracking. This avoids lower-level controllers mechanically replicating the discrete search actions of the upper-level controllers, instead allowing them to learn how to achieve trajectory tracking under actual continuous control and disturbance conditions.

[0036] Step 4: Lower-level SAC temperature trajectory tracking control method.

[0037] The lower-level controller uses the SAC algorithm for multi-actuator cooperative trajectory tracking. Its inputs are the real-time observed status of the thermal management system and the reference temperature trajectory output from the upper-level HIDP. The outputs are the control variables of the thermal management system, including compressor speed, water pump speed, and PTC heating power. The lower-level SAC training process is as follows: Figure 4 As shown.

[0038] The training process is as follows: first, read the reference temperature trajectory of the upper layer, then collect the current system state, then use the Actor network to output the corresponding control action, the thermal management system updates the state and obtains a reward, stores the samples generated by each step of interaction into the experience replay pool and samples for training, and then updates the network inside the SAC. If the number of training times reaches the set number or the reward value obtained is less than the set threshold, the training ends and the SAC control strategy after training is obtained.

[0039] Specifically, during the reading of the reference temperature trajectory, the complete observations can be represented as follows: ; in, Indicates basic observations, This represents the normalized value of the upper-layer target temperature trajectory. This represents the observed temperature tracking error. ; Each normalized component is defined as follows: , , , , ; ; This represents the temperature value on the target battery temperature trajectory at the current moment.

[0040] When a cryogenic expert network needs to differentiate between -10°C and -20°C operating conditions, ambient temperature indicators can be added to the observations, for example: ; The lower-level SAC action space consists of three-dimensional continuous actions, and the SAC policy network outputs three-dimensional normalized actions. It can be represented as: ; in, These represent the normalized control actions for compressor speed, water pump speed, and heating power, respectively. The value is calculated in real-time by the Actor policy network of the SAC algorithm based on the current observation state, and is not manually set. The Actor policy network is compressed using the tanh limiting function, and the output action components satisfy... This range.

[0041] Map normalized actions to physical executor commands: ; ; ; in, , These are the maximum and minimum compressor speeds, respectively. , These are the maximum and minimum values ​​of the water pump speed, respectively; , These represent the maximum and minimum values ​​of the PTC heating power, respectively.

[0042] The lower-level SAC reward function uses a two-term structure, including temperature trajectory tracking error and total electrical power consumption: ; in, For temperature trajectory tracking weights, As power consumption weight, Total power consumption for compressor, water pump and PTC electric heater: ; These are the power consumption figures for the compressor, water pump, and PTC electric heater, respectively.

[0043] The aforementioned reward function enables the lower-level SAC to prioritize reducing the deviation between the actual battery temperature and the upper-level reference trajectory. Furthermore, since the reward function includes a total power consumption term, the lower-level SAC can suppress the ineffective energy consumption of the compressor, water pump, and PTC while tracking the reference temperature trajectory.

[0044] The policy network in the SAC algorithm can be represented as: ; in, For Actor policy networks, Here are the parameters for the policy network. Two commentator networks estimate the state-action value respectively: ; The target Q value can be constructed as follows: ; in, and For the target critic network, As a discount factor, Let be the entropy coefficient. The commentator network loss function is: ; In the formula, This represents the network parameters of the i-th commentator. The target Q-value used for training the critic network; The policy network loss function is: ; The target commentator network uses soft updates: ; in, This is the soft update coefficient.

[0045] In this embodiment, a two-expert lower-layer SAC structure is adopted. The first expert network is used for the 0°C heat pump heating condition alone, and the second expert network is used for the -10°C and -20°C low-temperature mixed heating condition.

[0046] The expert selection criteria are as follows: in, This indicates a network of expert strategies for heat pump-specific heating. This represents a heat pump-PTC synergistic heating expert strategy network. The corresponding PTC permission flag is: exist At temperatures above -10℃, the PTC electric heater is constrained to not participate in heating; At ≤-10℃, the PTC electric heater and the heat pump system are allowed to participate in battery heating together.

[0047] In one implementation, the upper-layer HIDP outputs target temperature trajectory files for ambient temperatures of 0°C, -10°C, and -20°C, respectively. The lower-layer SAC reads the corresponding files during training. The system is designed to perform trajectory tracking training within a full NEDC cycle or extended simulation duration. By training separate expert strategy networks for heat pump-only heating and heat pump-PTC co-heating, and by mixing -10℃ and -20℃ operating conditions in the low-temperature expert network, the controller can take into account the differences in heat pump efficiency, heat loss, and PTC auxiliary heating requirements under different ambient temperatures.

[0048] The lower-level SAC training uses an experience replay pool and a policy update mechanism, with each interaction generating samples: ; in, This indicates the end of the round.

[0049] The samples generated at each interaction step are stored in the experience replay pool. During training, mini-batch data is randomly sampled from the experience replay pool to update the dual critic network, policy network, and target critic network.

[0050] In this embodiment, the lower-layer SAC uses a multilayer perceptron structure with two hidden layers and 256 neurons per layer. The experience replay pool capacity is 400,000, the initial learning steps are 20,000, the batch size is 256, and the discount factor is [not specified]. Target network soft update coefficient The training frequency is once per step, and the entropy coefficient is automatically adjusted. These parameters can be adjusted according to different vehicle platforms and thermal management system parameters.

[0051] After training, only the Actor policy network parameters need to be exported for online control and Simulink / AVL deployment; the critic network and target critic network are only used during the training phase. The exported models include parameter files for the heat pump-only heating Actor and the heat pump-PTC co-heating Actor. In the vehicle controller or simulation platform, the corresponding expert network is first selected based on the ambient temperature. Then, the Actor outputs normalized actions based on real-time observations. After action mapping and rate-of-change constraints, the actual actuator commands are obtained. For example... Figure 5 As shown, the battery heating speed is significantly improved, and its energy-saving performance is also better.

[0052] Step 5: The thermal management system executes the control inputs from the lower-level controller and feeds back the real-time status to both the upper-level and lower-level controllers.

[0053] This invention also provides a hierarchical collaborative control system for electric vehicle heat pumps and PTC heating, used to implement the above-mentioned hierarchical collaborative control method, comprising: The system comprises the following modules: a data acquisition module for acquiring the current thermal state, ambient temperature, and future NEDC operating condition information of the power battery thermal management system; an upper-level HIDP trajectory planning module for planning a target battery temperature trajectory based on a hierarchical iterative dynamic programming algorithm, taking into account future NEDC operating condition information, ambient temperature, and current thermal state, with the optimization goal of reducing energy consumption and temperature deviation; a lower-level SAC tracking control module for outputting control quantities for controlling the compressor, water pump, and PTC electric heater, using the target battery temperature trajectory as the tracking target and combining it with the current thermal state; and an execution module for applying the control quantities to the thermal management system and feeding back the real-time status to the upper-level HIDP trajectory planning module and the lower-level SAC tracking control module.

[0054] Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. Furthermore, those skilled in the art will recognize that, based on the ideas of this invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A hierarchical coordinated control method for heat pump and PTC heating in electric vehicles, characterized in that, Includes the following steps: Step 1: Construct an electric vehicle battery thermal management system, including: a heat pump circuit, a battery heating circuit, a PTC electric heater, a power battery, a vehicle longitudinal dynamics heat generation module, and a controller module. The controller module includes an upper-level controller and a lower-level controller. Step 2: Define the internal state vector, actuator control vector, and external disturbance vector of the thermal management system as the thermal state during the power battery heating process; Step 3: The upper-level controller adopts a hierarchical iterative dynamic programming algorithm. Based on the future NEDC operating condition information of the electric vehicle, the ambient temperature and the thermal state of the current thermal management system, it plans a target battery temperature trajectory with the optimization goal of reducing energy consumption and temperature deviation. Step 4: The lower-level controller uses the SAC algorithm to track the temperature trajectory of the target battery. Combined with the current thermal state of the thermal management system, it outputs the control quantities of the thermal management system. The control quantities include: compressor speed, water pump speed and PTC heating power. Step 5: The thermal management system executes the control input output by the lower-level controller and feeds back the real-time status to the upper-level controller and the lower-level controller; In step 2, the internal state vector of the thermal management system is defined as follows: ； in, For the core temperature of the battery, The surface temperature of the battery. This refers to the inlet temperature of the battery heating circuit. This refers to the outlet temperature of the battery heating circuit. The average temperature of the battery heating circuit. This is the mixing point temperature after heat exchange between the heat pump and the battery circuit. To take into account the dynamic heat supply after considering the dynamic response of the heat source; The actuator control vector is defined as: ； in, This refers to the compressor speed. The water pump speed, PTC heating power; The external disturbance vector is defined as: ； in, For ambient temperature, For vehicle speed; In step 3, the stage cost of the hierarchical iterative dynamic programming algorithm is defined as: in, The target temperature for the battery. For the first Step electrical energy consumption, , and These are respectively the temperature deviation weight, energy consumption weight, and terminal temperature deviation weight; The hierarchical iterative dynamic programming algorithm satisfies the executor amplitude constraint during planning: ； ； ； Simultaneously satisfy the actuator rate of change constraint: ； ； ； in, , , These represent the maximum compressor speed, maximum water pump speed, and maximum PTC heating power, respectively. , , These represent the maximum rate of change of compressor speed, the maximum rate of change of water pump speed, and the maximum rate of change of PTC heating power, respectively.

2. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 1, characterized in that, In step 1, the heat pump circuit includes at least a compressor, a condensing heat exchange component, an expansion component, and an evaporating heat exchange component; the battery heating circuit includes a battery circuit water pump, a heat exchange plate or a battery cold plate, a coolant pipeline, and a liquid replenishment / expansion component; a PTC electric heater is arranged in the battery heating circuit to provide auxiliary heating in low-temperature environments; the power battery and the battery heating circuit exchange heat through the battery heat exchange plate.

3. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 1, characterized in that, The hierarchical iterative dynamic programming algorithm uses a complete state node for pruning and merging, and the complete state node is represented as follows: ； in, , and These represent the compressor speed, water pump speed, and PTC heating power at the previous moment, respectively.

4. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 3, characterized in that, In step 4, the complete observables of the SAC algorithm are: ； in, Indicates basic observations, This represents the normalized value of the upper-layer target temperature trajectory. This represents the observed temperature tracking error. ； Each normalized component is defined as follows: , ， ， , A symbol indicating that the electric heater is permitted to participate in heating; ； ； This represents the temperature value on the target battery temperature trajectory at the current moment.

5. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 4, characterized in that, The reward function of the SAC algorithm adopts a two-term structure, including temperature trajectory tracking error and total electrical power consumption, specifically expressed as: ； in, For temperature trajectory tracking weights, As power consumption weight, This represents the total power consumption of the compressor, water pump, and PTC electric heater.

6. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 5, characterized in that, The three-dimensional continuous motion output by the SAC algorithm is: ； Normalized control values ​​representing compressor speed, water pump speed, and heating power respectively. Control action; The three-dimensional continuous motion is mapped into control quantities for the compressor, water pump, and PTC electric heater: ； ； ； in, , These are the maximum and minimum compressor speeds, respectively. , These are the maximum and minimum values ​​of the water pump speed, respectively; , These represent the maximum and minimum values ​​of the PTC heating power, respectively.

7. The hierarchical coordinated control method for electric vehicle heat pump and PTC heating according to claim 6, characterized in that, The SAC algorithm employs two expert policy networks. When the ambient temperature When selecting the first expert strategy network, the first expert strategy network only allows the heat pump to heat alone and prohibits the PTC electric heater from starting. When the ambient temperature ≤ At that time, a second expert strategy network is selected, which allows the PTC electric heater and the heat pump to participate in battery heating together.

8. A hierarchical collaborative control system for electric vehicle heat pump and PTC heating, used to implement the hierarchical collaborative control method for electric vehicle heat pump and PTC heating as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire the current thermal state, ambient temperature, and future NEDC operating condition information of the power battery thermal management system. The upper-level HIDP trajectory planning module is used to plan a target battery temperature trajectory based on the hierarchical iterative dynamic programming algorithm, taking into account future NEDC operating conditions, ambient temperature, and current thermal state, with the optimization goal of reducing energy consumption and temperature deviation. The lower-level SAC tracking control module is used to take the target battery temperature trajectory as the tracking target and, in combination with the current thermal state, output control quantities for controlling the compressor, water pump and PTC electric heater; The execution module is used to apply the control quantity to the thermal management system and feed back the real-time status to the upper-level HIDP trajectory planning module and the lower-level SAC tracking control module.

Images