Electric vehicle battery temperature control method based on self-adaptive dynamic programming

Abstract

The invention discloses an electric vehicle battery temperature control method based on self-adaptive dynamic programming, which comprises the following steps of: 1, respectively establishing a battery model, a refrigerant cooling model and a battery electrothermal coupling model, and outputting a battery temperature and a battery state of charge (SOC) according to the battery electrothermal coupling model; 2, establishing an electric vehicle battery thermal management global optimization problem; 21) establishing a dynamic programming mathematical model, and determining constraint conditions of an optimization problem; 22) discretizing the driving condition, the state variable and the decision variable to obtain a state variable set of any stage; 23) reversely solving an optimal objective function and a corresponding decision variable of each stage; 24) giving an initial state variable, and positively solving a globally optimal solution; step 3, establishing an online thermal management optimization controller; 31) constructing an optimal thermal management strategy data set under various working conditions; and 32) training a neural network, and constructing to obtain an online thermal management optimization controller.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery thermal management, and specifically relates to an electric vehicle battery temperature control method based on adaptive dynamic programming. Background Art

[0002] With the rapid growth in sales of new energy vehicles, lithium-ion batteries are widely used as carriers for storing and utilizing electrical energy. However, lithium-ion batteries generate a lot of heat during the charging and discharging process, which will cause the battery temperature to rise. Excessive temperature may cause accidents such as internal short circuits in the battery, shell ruptures, and battery leakage and fire. In addition, if lithium-ion batteries work under high temperature conditions for a long time, the performance of lithium batteries will be greatly reduced. Therefore, in order to maximize the performance of lithium-ion batteries and ensure the optimal charging and discharging efficiency, their temperature must be maintained within the range of 15 to 40°C.

[0003] In order to control the battery temperature, a battery thermal management system is required. However, since the energy of the battery thermal management system comes from the power of the power battery, while ensuring that the power battery operates within the optimal temperature range, it also increases the power-consuming components of the vehicle. And often the energy consumed when adjusting the temperature of the power battery is greater than the power increased by temperature adjustment, resulting in a reduction in the driving range of the power battery.

[0004] Traditional control strategies for battery thermal management include logic threshold strategy, fuzzy logic control strategy and global optimal energy management strategy. The logic threshold control strategy has a simple and stable structure, but it cannot achieve the optimal energy distribution of the power battery pack. After exceeding the threshold management range, the energy utilization rate of the power battery pack decreases. The use of fuzzy control strategy can optimize the energy distribution to a certain extent, but the designation of fuzzy control rules is often affected by the designer's own subjective factors, which is prone to some loopholes. When the efficiency loss of system components changes during use, the control strategy cannot adjust the corresponding rules and parameters in time, which is easy to cause local optimality. Coupled with the long-term system cumulative error, the control effect is not ideal. In the control strategy considering dynamic programming optimization, conventional dynamic programming can only optimize specific driving conditions, cannot consider all driving conditions of the car, and cannot run online in real time. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide an electric vehicle battery temperature control method based on adaptive dynamic programming, which can not only ensure that the battery temperature is reduced to a reasonable range, but also reduce the energy consumption of the thermal management system.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for controlling battery temperature of an electric vehicle based on adaptive dynamic programming comprises the following steps:

[0008] Step 1: Build a control-oriented electric vehicle battery thermal management system model

[0009] Establish a battery model, a refrigerant cooling model and a battery electrothermal coupling model respectively, and output the battery temperature and battery state of charge SOC according to the battery electrothermal coupling model;

[0010] Step 2: Establish the global optimization problem of electric vehicle battery thermal management

[0011] 21) Establish a dynamic programming mathematical model and determine the constraints of the optimization problem;

[0012] 22) Discretize the driving conditions, state variables and decision variables respectively to obtain a set of state variables at any stage;

[0013] 23) Inversely solve the optimal objective function and corresponding decision variables at each stage;

[0014] 24) Given the initial state variables, forward solve the global optimal solution;

[0015] Step 3: Establish an online thermal management optimization controller

[0016] 31) Construct the optimal thermal management strategy data set under various working conditions

[0017] According to different driving conditions, based on the dynamic programming optimization problem of electric vehicle battery thermal management, the optimal decision sequence and optimal state sequence of control variables are obtained, and the optimal thermal management strategy data set is constructed with the optimal decision sequence and optimal state sequence;

[0018] 32) Training a Neural Network

[0019] The constructed optimal thermal management strategy data set is used to train the neural network and construct an online thermal management optimization controller.

[0020] Furthermore, in the step 1, in the refrigerant cooling model, the heat transferred from the refrigerant to the battery pack It is expressed as:

[0021]

[0022] m r =λV c η c n c ρ c / 60

[0023] Where: C r is the heat capacity of the refrigerant; m ris the refrigerant flow rate into the battery circuit; h n,in is the inlet enthalpy of the battery cold plate; h h,out is the outlet enthalpy of the battery cold plate; λ is the opening of the electronic expansion valve of the battery circuit; V c is the displacement of the compressor; η c is the volumetric efficiency of the compressor; n c is the compressor speed; ρ c is the density of the refrigerant.

[0024] Further, in the step 1, in the battery model, the charge and discharge state SOC is expressed as:

[0025]

[0026] Among them, U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of thermal management system; R bat is the internal resistance; C bat is the total charge of the battery.

[0027] Furthermore, in the step 1, in the battery electrothermal coupling model, the influence of the heat generated and dissipated by the refrigerant flowing around the battery module on the battery temperature is used to obtain the thermodynamic balance equation of the battery electrothermal coupling model:

[0028]

[0029] Where: m bat is the battery mass; C bat is the total charge of the battery; is the temperature change of the battery; T bat is the battery temperature; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; bat is the charge and discharge current; R bat is the internal resistance; U oc is the external voltage.

[0030] Furthermore, in step 21), the dynamic programming mathematical model is constructed to minimize the cumulative energy consumption of battery thermal management within the global time range under the system terminal constraint:

[0031] J * =minJ=min{∑L(x(k),u(k))}

[0032] L(x(k),u(k))=ω1P hp +ω2|T bat ―T * bat |

[0033] Among them: J * is the optimal objective function; J is the objective function; ω1 and ω2 are weight coefficients; P hp is the energy consumption of thermal management system; T bat is the battery temperature; is the target battery temperature; x(k) and u(k) are the state variable and control variable respectively, and:

[0034]

[0035] The constraints of the optimization problem are:

[0036]

[0037] Where: Q bc_max The maximum cooling capacity allocated to the battery circuit for the air conditioning system; SOC min and SOC max is the lower and upper limits of the battery SOC; T bat_min and T bat_max are the lower and upper limits of the battery temperature.

[0038] Furthermore, in step 22), the driving condition is discretized into N stages by time, and the battery temperature T in the state variable is bat and battery state of charge SOC are discretized respectively, and the number of discretized grids is a and b respectively; then the state variable set of any k stage is:

[0039]

[0040] Where: X k represents the set of state variables at stage k; represents a state variable in the kth stage; T bat_i represents the battery temperature at the i-th grid point; SOC j represents the battery state of charge at the jth grid point.

[0041] Further, in step 23), the optimal objective function of each stage is reversely solved starting from stage N-1;

[0042] When k = N-1 stage, the objective function is:

[0043]

[0044] When k∈[1,…,N―2] stage, the objective function is:

[0045]

[0046] Where: F(·) is the state transfer function.

[0047] Further, in step 24), the method for forward solving the global optimal solution is:

[0048] When k=1, according to the result of the reverse solution, the decision variable u(1) of the initial state is found and applied to the state variable. The battery temperature T of the next stage is calculated using the state transfer equation. bat and battery state of charge SOC;

[0049] When k∈[2,…,N―1] stage, the state variable T of the current stage calculated in the previous stage is bat (k) and SOC(k), find the corresponding optimal decision variable; if the optimal decision variable found is not the state variable of the current stage, the interpolation method is used to obtain the optimal decision variable u(k) of the current stage; this is gradually recursively carried out to the final stage to obtain the optimal control sequence {u * (1),u * (2),...,u * (k)} and the optimal state sequence {x * (1),x * (2),...,x * (k)}, that is, the optimal battery temperature trajectory and optimal battery state trajectory SOC * .

[0050] Furthermore, the battery temperature T bat and the battery state of charge SOC as state variables, with the battery circuit cooling capacity Q bc (k) is used as the control variable, and the state transfer equation is obtained as follows:

[0051]

[0052] Among them: U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of the thermal management system, which is determined by the vehicle speed v and the battery circuit cooling capacity Q bc The values ​​of the two are obtained by looking up the table; R bat is the internal resistance; C bat is the total charge of the battery; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; m bat is the battery mass; C bat is the total charge of the battery; T bat is the battery temperature.

[0053] The beneficial effects of the present invention are:

[0054] The electric vehicle battery temperature control method based on adaptive dynamic programming of the present invention improves the economy of the battery thermal management system compared with the temperature control strategy based on rules, and solves the problem of lack of real-time performance compared with the temperature control strategy based on dynamic programming algorithm; by designing a controller based on neural network and dynamic programming to adjust the cooling capacity of the battery circuit, and utilizing the characteristics of battery and heat coupling, the temperature requirements and energy-saving requirements of the battery can be met at the same time, which not only ensures that the battery temperature is reduced to a reasonable range, but also reduces the energy consumption of the thermal management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0056] Figure 1 This is a flow chart of an electric vehicle battery temperature control method based on adaptive dynamic programming of the present invention;

[0057] Figure 2 This is a schematic diagram of a battery thermal management system for an electric vehicle;

[0058] Figure 3 It is the steady-state model and working point cloud diagram of the air-conditioning system;

[0059] Figure 4 It is the battery SOC curve diagram;

[0060] Figure 5 Schematic diagram of adaptive dynamic programming control for electric vehicle battery thermal management;

[0061] Figure 6 It is a vehicle speed curve diagram of the selected vehicle driving cycle (WLTC);

[0062] Figure 7 Battery temperature optimization trajectory diagram for global optimization of electric vehicle battery thermal management;

[0063] Figure 8 A zoomed-in view of the battery temperature optimization trajectory for global optimization of electric vehicle battery thermal management. DETAILED DESCRIPTION

[0064] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0065] like Figure 1 As shown, the electric vehicle battery temperature control method based on adaptive dynamic programming in this embodiment includes the following steps.

[0066] Step 1: Build a control-oriented electric vehicle battery thermal management system model

[0067] like Figure 2 As shown, the electric vehicle battery thermal management system includes a battery thermal management system for providing refrigerant to the battery, and the battery thermal management system includes a condenser, an evaporator, a compressor, an expansion valve, etc. In this embodiment, when establishing a control-oriented electric vehicle battery thermal management system model, it is necessary to establish a battery model, a refrigerant cooling model, and a battery electrothermal coupling model respectively, and output the battery temperature and the battery state of charge SOC according to the battery electrothermal coupling model.

[0068] (1) Refrigerant cooling model

[0069] like Figure 3 As shown in the figure, the steady-state model and working point cloud diagram of the air-conditioning system are shown in the figure. Specifically, in the refrigerant cooling model, the refrigerant flow rate m flowing into the battery circuit is r for:

[0070] m r =λV c η c n c ρ c / 60

[0071] The heat transferred from the refrigerant to the battery pack It is expressed as:

[0072]

[0073] Where: C r is the heat capacity of the refrigerant; m r is the refrigerant flow rate into the battery circuit; h n,in is the inlet enthalpy of the battery cold plate; h h,out is the outlet enthalpy of the battery cold plate; λ is the opening of the electronic expansion valve of the battery circuit; V c is the displacement of the compressor; η c is the volumetric efficiency of the compressor; η c is the compressor speed; ρ c is the density of the refrigerant.

[0074] (2) Battery model

[0075] like Figure 4 As shown in the figure, it is a battery SOC curve. Specifically, in the battery model, the equivalent circuit model consists of the external voltage U oc and internal resistance R bat Composition, where the charge and discharge current I bat It is expressed as:

[0076] I bat =P bat / Ubat

[0077] Where: I bat is the charge and discharge current; P bat Indicates battery power; U bat is the external voltage;

[0078] Battery power P bat It is expressed as:

[0079] P bat =P mot +P hp

[0080] Where: P mot is the motor power; P hp Energy consumption for thermal management system;

[0081] External voltage U bat It is expressed as:

[0082] U bat =U oc ―I bat R bat

[0083] Where: External voltage U oc and R bat Affected by battery temperature and SOC, their values ​​can be obtained through offline testing, that is, U is obtained by table interpolation. oc and R bat The actual value of .

[0084] Combining the above formula, the charge and discharge current I bat It can be expressed as:

[0085]

[0086] The charge and discharge state SOC is expressed as:

[0087]

[0088] Then the charge and discharge state SOC can be expressed as:

[0089]

[0090] Among them, U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of thermal management system; R bat is the internal resistance; C bat is the total charge of the battery.

[0091] (3) Battery electrothermal coupling model

[0092] In the battery electrothermal coupling model, the heat generated inside the battery pack is for:

[0093]

[0094] The influence of heat generated and dissipated by the refrigerant flowing around the battery module on the battery temperature leads to the thermal balance equation of the battery electrothermal coupling model:

[0095]

[0096] Where: m bat is the battery mass; C bat is the total charge of the battery; is the temperature change of the battery; T bat is the battery temperature; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; bat is the charge and discharge current; R bat is the internal resistance; U oc is the external voltage.

[0097] Step 2: Establish the global optimization problem of electric vehicle battery thermal management. Select appropriate control input, establish the thermal management optimization problem, and determine the state quantity and control quantity constraints of the optimization problem;

[0098] 21) Establish a dynamic programming mathematical model and determine the constraints of the optimization problem.

[0099] In this embodiment, the dynamic programming mathematical model is constructed to minimize the cumulative energy consumption of battery thermal management within the global time range under the system terminal constraint:

[0100] J * =min J=min{∑L(x(k),u(k))}

[0101] L(x(k),u(k))=ω1P hp +ω2|T bat ―T * bat |

[0102] Among them: J * is the optimal objective function; J is the objective function; ω1 and ω2 are weight coefficients; P hp is the energy consumption of thermal management system; T bat is the battery temperature; is the target battery temperature; x(k) and u(k) are the state variable and control variable respectively, and:

[0103]

[0104] The constraints of the optimization problem are:

[0105]

[0106] Where: Q bc_max The maximum cooling capacity allocated to the battery circuit for the air conditioning system; SOC min and SOC max is the lower and upper limits of the battery SOC; T bat_min and T bat_max are the lower and upper limits of the battery temperature.

[0107] 22) The driving conditions, state variables and decision variables are discretized respectively to obtain the state variable set at any stage.

[0108] Specifically, the driving condition is discretized into N stages according to time. For example, if the time interval of the discretization is defined as 1 second, then a driving condition with N seconds will have N stages. bat and battery state of charge SOC are discretized respectively, and the number of discretized grids is a and b respectively. In this way, a N×a×b time-varying state variable grid is formed. Then the state variable set of any k stage is:

[0109]

[0110] Where: X k represents the set of state variables at stage k; represents a state variable in the kth stage; T bat_i represents the battery temperature at the i-th grid point; SOC j represents the battery state of charge at the jth grid point.

[0111] In this embodiment, the state quantity and the control quantity are divided into grids. The state quantity of the electric vehicle battery SOC increases from 0 to 100, and the number of grids is 100; the state quantity of the battery temperature T bat From 32℃ to 36℃, the number of grids is 1000; the control variable Q bc From 0w to 1000w, the number of grids is 5.

[0112] 23) Reverse solve the optimal objective function and corresponding decision variables at each stage.

[0113] The reverse solution is to calculate the optimal objective function and decision variables of each stage from the final stage to the initial stage. The reverse solution of dynamic programming regards the cost of the final stage as 0, and reversely solves the optimal objective function of each stage starting from the N-1 stage;

[0114] When k = N-1 stage, the objective function is:

[0115]

[0116] When k∈[1,…,N―2] stage, the objective function is:

[0117]

[0118] Where: F(·) is the state transfer function.

[0119] Specifically, when k = N-1 stage, all discrete decision variables Q bc Acting on each state variable battery temperature T bat and battery SOC, calculate the optimal objective function, and record the corresponding decision variable sequence, denoted as u(N―1).

[0120] When k∈[1,…,N―2] stage, all decision variables are traversed and the objective function of all state variable grid points is calculated; different from the k=N―1 stage, the objective function not only includes the objective function of the kth stage, but also needs to add the optimal objective function of the k+1th stage. In this way, the optimal objective function of the entire stage can be calculated, and the optimal objective function of the system in the entire stage can also be calculated. After the calculation is completed, the sequence of decision variables taken needs to be recorded, recorded as u(k).

[0121] 24) Given the initial state variables, forward solve the global optimal solution.

[0122] Forward calculation is to first give the initial value of the state variable, use the optimal decision variable obtained by reverse calculation, calculate from the initial stage to the final stage, and finally obtain the global optimal state variable sequence. Specifically, the method of forward solving the global optimal solution is:

[0123] When k=1, according to the result of the reverse solution, the decision variable u(1) of the initial state is found and applied to the state variable. The battery temperature T of the next stage is calculated using the state transfer equation. bat and the battery state of charge SOC. In this embodiment, the initial state variable battery temperature T is set bat (0) is 35°C, and the battery SOC(0) is 0.9.

[0124] When k∈[2,…,N―1] stage, the state variable T of the current stage is calculated according to the state transfer equation of the previous stage. bat (k) and SOC(k), find the corresponding optimal decision variable; if the optimal decision variable found is not the state variable of the current stage, the interpolation method is used to obtain the optimal decision variable u(k) of the current stage; this is gradually recursively carried out to the final stage to obtain the optimal control sequence {u* (1),u * (2),...,u * (k)} and the optimal state sequence {x * (1),x * (2),...,x * (k)}, that is, the optimal battery temperature trajectory and optimal battery state trajectory SOC * .

[0125] In this embodiment, the battery temperature T bat and the battery state of charge SOC as state variables, with the battery circuit cooling capacity Q bc (k) is used as the control variable, and the state transfer equation is obtained as follows:

[0126]

[0127] Among them: U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of the thermal management system, which is determined by the vehicle speed v and the battery circuit cooling capacity Q bc The numerical values ​​of the two are obtained by looking up the table, and the MAP diagram is obtained by identifying the high-simulation model; R bat is the internal resistance; C bat is the total charge of the battery; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; m bat is the battery mass; C bat is the total charge of the battery; T bat is the battery temperature.

[0128] Step 3: Establish an online thermal management optimization controller

[0129] 31) Construct the optimal thermal management strategy data set under various working conditions

[0130] like Figure 5 As shown in the figure, the offline optimization program is solved for different representative driving cycle data sets to obtain the input and output required for training the neural network. Specifically, first obtain the standard driving cycle data set, including UDDS, NEDC, WLTC and other driving conditions. For each driving cycle of different driving conditions, the dynamic programming optimization problem of electric vehicle battery thermal management is carried out and the optimal decision sequence and optimal state sequence of the control variables are obtained. Then, the optimal thermal management strategy data set is constructed based on the driving condition data, the optimal decision sequence and the optimal state sequence. Figure 6 As shown, it is a vehicle speed curve diagram of the vehicle driving cycle (WLTC) selected in this embodiment.

[0131] 32) Training a Neural Network

[0132] The constructed optimal thermal management strategy data set is used to train the neural network and construct an online thermal management optimization controller.

[0133] In this embodiment, driving condition data is used as the input of the neural network model, and the optimal state quantity sequence is used as the output of the neural network model. The neural network model is designed and trained using neural network training software. Specifically, in order to train the neural network, this embodiment adopts the Levenberg-Marquardt algorithm because it has the characteristics of fast convergence and robustness; the neural network model uses the mean quadratic error (MSE) as the fitness function. The trained and verified neural network model is used as an intelligent online thermal management controller. By acquiring real-time data of the neural network input, the neural network outputs the optimal trajectory of the battery temperature and the optimal trajectory of the battery SOC, such as Figure 7-8 shown.

[0134] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for controlling the battery temperature of an electric vehicle based on adaptive dynamic programming, characterized in that: The steps include: Step 1: Build a control-oriented electric vehicle battery thermal management system model Establish a battery model, a refrigerant cooling model and a battery electrothermal coupling model respectively, and output the battery temperature and battery state of charge SOC according to the battery electrothermal coupling model; Step 2: Establish the global optimization problem of electric vehicle battery thermal management 21) Establish a dynamic programming mathematical model and determine the constraints of the optimization problem; 22) Discretize the driving conditions, state variables and decision variables respectively to obtain a set of state variables at any stage; 23) Inversely solve the optimal objective function and corresponding decision variables at each stage; 24) Given the initial state variables, forward solve the global optimal solution; Step 3: Establish an online thermal management optimization controller 31) Construct the optimal thermal management strategy data set under various working conditions According to different driving conditions, based on the dynamic programming optimization problem of electric vehicle battery thermal management, the optimal decision sequence and optimal state sequence of control variables are obtained, and the optimal thermal management strategy data set is constructed with the optimal decision sequence and optimal state sequence; 32) Training a Neural Network The constructed optimal thermal management strategy data set is used to train the neural network and construct an online thermal management optimization controller.

2. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In the step 1, in the refrigerant cooling model, the heat transferred from the refrigerant to the battery pack It is expressed as: m r =λV c or c n c r c / 60 Where: C r is the heat capacity of the refrigerant; m r is the refrigerant flow rate into the battery circuit; h h,in is the inlet enthalpy of the battery cold plate; h h,out is the outlet enthalpy of the battery cold plate; λ is the opening of the electronic expansion valve of the battery circuit; V c is the displacement of the compressor; η c is the volumetric efficiency of the compressor; n c is the compressor speed; ρ c is the density of the refrigerant.

3. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In the step 1, in the battery model, the charge and discharge state SOC is expressed as: Among them, U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of thermal management system; R bat is the internal resistance; C bat is the total charge of the battery.

4. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In the step 1, in the battery electrothermal coupling model, the influence of the heat generated and dissipated by the refrigerant flowing around the battery module on the battery temperature is used to obtain the thermodynamic balance equation of the battery electrothermal coupling model: Where: m bat is the battery mass; C bat is the total charge of the battery; is the temperature change of the battery; T bat is the battery temperature; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; bat is the charge and discharge current; R bat is the internal resistance; U oc For external voltage.

5. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In step 21), the dynamic programming mathematical model is constructed to minimize the cumulative energy consumption of battery thermal management within the global time range under the system terminal constraint: L(x(k),u(k))=ω1P hp +ω2|T bat -T * bat | Among them: J * is the optimal objective function; J is the objective function; ω1 and ω2 are weight coefficients; P hp is the energy consumption of thermal management system; T bat is the battery temperature; is the target battery temperature; x(k) and u(k) are the state variable and control variable respectively, and: The constraints of the optimization problem are: Where: Q bc_max The maximum cooling capacity allocated to the battery circuit for the air conditioning system; SOC min and SOC max is the lower and upper limits of the battery SOC; T bat _ min and T bat _ max are the lower and upper limits of the battery temperature.

6. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In step 22), the driving condition is discretized into N stages by time, and the battery temperature T in the state variable is bat and battery state of charge SOC are discretized respectively, and the number of discretized grids is a and b respectively; then the state variable set of any k stage is: Where: X k represents the set of state variables at stage k; represents a state variable in the kth stage; T bat_i represents the battery temperature at the i-th grid point; SOC j represents the battery state of charge at the jth grid point.

7. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In the step 23), the optimal objective function of each stage is reversely solved starting from the N-1 stage; When k = N-1 stage, the objective function is: When k∈[1,…,N―2] stage, the objective function is: Where: F(·) is the state transfer function.

8. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 1 is characterized in that: In step 24), the method for forward solving the global optimal solution is: When k=1, according to the result of the reverse solution, the decision variable u(1) of the initial state is found and applied to the state variable. The battery temperature T of the next stage is calculated using the state transfer equation. bat and battery state of charge SOC; When k∈[2,…,N―1] stage, the state variable T of the current stage calculated in the previous stage is bat (k) and SOC(k), find the corresponding optimal decision variable; if the optimal decision variable found is not the state variable of the current stage, the interpolation method is used to obtain the optimal decision variable u(k) of the current stage; this is gradually recursively carried out to the final stage to obtain the optimal control sequence {u * (1),u * (2),...,u * (k)} and the optimal state sequence {x * (1),x * (2),...,x * (k)}, that is, the optimal battery temperature trajectory and optimal battery state trajectory SPC * .

9. The electric vehicle battery temperature control method based on adaptive dynamic programming according to claim 7 or 8, characterized in that: The battery temperature T bat and the battery state of charge SOC as state variables, with the battery circuit cooling capacity Q bc (k) is used as the control variable, and the state transfer equation is obtained as follows: Among them: U oc is the external voltage; P bat Indicates battery power; P hp is the energy consumption of the thermal management system, which is determined by the vehicle speed v and the battery circuit cooling capacity Q bc The values ​​of the two are obtained by looking up the table; R bat is the internal resistance; C bat is the total charge of the battery; The heat generated inside the battery pack; The heat transferred from the refrigerant to the battery pack; m bat is the battery mass; C bat is the total charge of the battery; T bat is the battery temperature.

Images