An electric vehicle thermal management control system and method considering energy saving and battery life

Abstract

This invention discloses a thermal management control system and method for electric vehicles that considers energy conservation and battery life. The control system includes an operating condition prediction module, an energy efficiency ratio prediction module, a thermal management controller, a state monitoring module, and an execution module. The operating condition prediction module determines future operating conditions based on the current driving scenario; the state monitoring module acquires and stores the vehicle status; the energy efficiency ratio prediction module generates an energy efficiency ratio prediction sequence based on future driving conditions; and the thermal management controller uses thermal management energy consumption, battery life, and passenger compartment temperature as objective functions to make decisions on the control variables of the thermal management system and outputs control instructions to the execution module to implement a closed-loop control system. This invention linearizes the complex thermal management system and improves computational efficiency.

Description

Technical Field

[0001] The present invention belongs to the field of new energy vehicles, and in particular relates to a thermal management control system and method for electric vehicles that takes energy saving and battery life into consideration. Background Art

[0002] The operating current of a power battery is the primary factor contributing to battery temperature rise. Its magnitude is primarily influenced by motor output power, regenerative braking power, and air conditioning load. Improper current distribution can waste energy and reduce battery life. With the development of intelligent connected technology, a new range of high-value information can be acquired and integrated into the optimization process, enabling greater energy savings in electric vehicles. However, designing controllers to achieve efficient multi-objective optimization has become a research challenge.

[0003] In recent years, optimization-based control strategies have become a research hotspot due to their good dynamic performance, such as dynamic programming, Pontryagin's minimum principle, model predictive control, etc. Among them, model predictive control is very suitable for online optimization due to its rolling optimization characteristics. However, the complexity and strong coupling of thermal management systems make it difficult to represent the model linearly. Most studies use nonlinear models to model thermal management systems, resulting in a large amount of calculation and an inability to increase the prediction time domain. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the present invention provides a thermal management control system and method for an electric vehicle that takes energy saving and battery life into consideration.

[0005] The present invention achieves the above technical objectives through the following technical means.

[0006] A thermal management control method for electric vehicles considering energy saving and battery life:

[0007] S1, based on the future driving conditions V of the vehicle pre , use the prediction equation to predict the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b，1 、Nth single cell battery temperature T b，n and the passenger compartment temperature T cab ;

[0008] S2, based on the average temperature of the battery pack T mean , battery pack temperature difference ΔT pre and T cab , construct the objective function and constraints of the problem to be optimized, and solve to obtain the optimal output power sequence U(t) of the battery cooler and evaporator; the T mean By T b，1 and T b，n The average value is obtained, the ΔTpre By T b，1 and T b，n To obtain by making a difference;

[0009] S3, based on the temperature difference ΔT of the battery pack pre The optimal output power sequence U(t) is used to determine the coolant flow rate u(t+1) at the next moment. Based on the historical coolant flow rate, it is determined whether the coolant flow rate in the prediction equation needs to be updated. If it needs to be updated, return to S1, otherwise output the optimal output power sequence U(t).

[0010] Furthermore, the prediction equation is:

[0011]

[0012] in: is the heat transfer coefficient between coolant and battery, P r is the Prandtl number, λ is the thermal conductivity of the coolant, v f is the kinematic viscosity of the coolant, d is the characteristic length of the heat exchange surface; A is the heat exchange area, u is the coolant flow rate, c b and c f are the specific heat capacities of the battery pack and coolant, m b is the mass of the battery pack, M is the mass of the coolant in the liquid cooling plate, s is the cross-sectional area of ​​the cooling pipe, ρ f is the coolant density, T s is the sampling time, n is the total number of single batteries, c air is the specific heat capacity of air, m air is the air quality, T b，1 (t) is the current temperature of the No. 1 single cell battery, T b，n (t) is the current temperature of the nth single cell battery, T r (t) is the current temperature of the coolant outlet, T f (t) is the current temperature of the coolant inlet, T cab (t) is the current temperature of the passenger compartment, Predict the thermal load vector in the time domain for the battery pack, Predict the heat load vector in the time domain for the passenger compartment, P chil is the optimal output power of the battery cooler, P eva is the optimal output power of the evaporator.

[0013] Furthermore, the objective function and constraints of the problem to be optimized are:

[0014]

[0015] stT b_min <T mean(i|t)<T b_max +ξ(i|t),i=0:p

[0016] 0<ΔT pre (i|t)<ΔT max , i=0:p

[0017] 0<P chil (i|t)<P chil_max , i=0:m-1

[0018] 0<P eva (i|t)<P eva_max , i=0:m-1

[0019] Y(0|t)=Y(t); U(0|t)=U(t)

[0020] Among them: w1 is the weight parameter of energy consumption evaluation index, w2 is the weight parameter of battery life evaluation index, w3 is the weight parameter of comfort evaluation index, w4 is the weight parameter of relaxation factor in constraint condition, P consumption is the energy consumption evaluation index, P life is the battery life evaluation index, P comfort is the passenger cabin comfort evaluation index, ξ is the relaxation factor, T b_min is the lower limit of the average temperature of the battery pack, T b_max is the upper limit of the average temperature of the battery pack, ξ is the relaxation factor, ΔT max is the upper limit of the battery pack temperature difference, P chil_max is the upper limit of the battery cooler output power, P eva_max is the upper limit of the evaporator output power, p is the prediction time domain, t represents the current moment, m is the control time domain, and Y(t) is the output of the system at the current moment in the prediction time domain.

[0021] Furthermore, the ratio of w1 to w3 is adaptively adjusted according to the battery SOC and the passenger compartment temperature difference: when the battery SOC is smaller and the passenger compartment temperature difference is smaller, w1 / w3 is larger; when the battery SOC is larger and the passenger compartment temperature difference is larger, w1 / w3 is smaller.

[0022] Furthermore, the coolant flow rate u(t+1) at the next moment is determined as follows:

[0023] The maximum value of the vector element of the battery pack temperature difference ΔT pre_max Subtracting each element in the control parameter vector yields T = ΔT pre_max -T control =[T1, T2, ..., T i ,...,T j ], where T control =[T c1, T c2 ,...,T cj ] is the control parameter obtained offline, j is the number of elements in the vector composed of control parameters, T i is the i-th element in vector T;

[0024] Determine the size of each element in T. When T i With T i+1 T-compliant i+1 <0<T i , then the flow velocity at the next moment u(t+1)=U coo (i), where U coo =[u1, u2, ..., u j-1 ] is the vector composed of the determined flow velocities in each stage.

[0025] Furthermore, the control parameter T obtained offline control , specifically including:

[0026] (1) Build a battery cooling simulation platform including battery packs, cooling channels, and water pumps;

[0027] (2) Design a phased control strategy, initialize the battery pack temperature difference control parameters, define the battery pack temperature difference and water pump energy consumption as the fitness function of the particle swarm algorithm, and initialize the weights of the two;

[0028] (3) Run the simulation and calculate the fitness function based on the battery pack temperature difference sequence and the total energy consumption of the water pump returned by the simulation platform;

[0029] (4) Update the particle velocity and find the control parameters that minimize the fitness function through particle swarm iterative optimization;

[0030] (5) Change the weight and determine the control parameter T corresponding to the optimal weight based on the Pareto boundary curve of the battery pack temperature difference and the water pump energy consumption control =[T c1 , T c2 ,...,T cj ].

[0031] Furthermore, it is necessary to determine whether to update the coolant flow rate in the prediction equation, specifically: when the judgment signal flag u When ≥20, output the value of u(t+1) and update the coolant flow rate in the prediction equation.

[0032] Furthermore, the judgment signal is determined by comparing the magnitudes of u(t+1) and u(t):

[0033] When u(t+1)>u(t), record the judgment signal flag u =20;

[0034] When u(t+1)=u(t), record flag u =0;

[0035] When u(t+1)<u(t), let q decrease from q=t until q=t-τ, compare u(t+1) and u(q), when u(t+1)<u(q), mark flag u =tq, if u(t+1)<u(q) does not appear, record flag u =20; where τ is the number of historical coolant flow rates minus 1.

[0036] A thermal management control system for electric vehicles that considers energy saving and battery life, including:

[0037] The operating condition prediction module selects the vehicle speed planning submodule or the road data analysis submodule based on the current scenario to determine the vehicle's future operating conditions;

[0038] An energy efficiency ratio prediction module converts the future driving conditions and the heat load in the prediction time domain into a prediction sequence of air intake volume and a cooling power sequence, inputs the sequence into an energy efficiency ratio prediction model, and outputs an energy efficiency ratio prediction sequence;

[0039] The thermal management controller includes a state estimator, a multi-objective optimization solver and a temperature difference controller; the state estimator predicts the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b，1 、Nth single cell battery temperature T b，n and the passenger compartment temperature T cab The multi-objective optimization solver constructs the objective function and constraints of the problem to be optimized and solves to obtain the optimal output power sequence; the temperature difference controller is used to determine the coolant flow rate u(t+1) at the next moment and update the coolant flow rate in the prediction equation;

[0040] The state monitoring module includes a state acquisition unit and a state storage unit. The state acquisition unit records the sensor information of each monitoring point, and the state storage unit stores it;

[0041] The execution module converts the first element of the optimal output power sequence into a control signal of the compressor, the water pump and the flow valve, and acts on the compressor, the water pump and the flow valve.

[0042] In the above technical solution, the future driving conditions are converted into a prediction sequence of the air intake volume, specifically: Where δ is the relationship coefficient, V fan The air volume when the condenser fan is on and the vehicle speed is 0. is the predicted vehicle speed at the i-th second from the current moment.

[0043] The beneficial effects of the present invention are:

[0044] (1) The thermal management controller of the present invention establishes energy consumption evaluation indicators, battery life evaluation indicators, passenger compartment comfort evaluation indicators and safety constraints, constructs the objective function of the problem to be optimized and its constraints, solves the optimal output power sequence through the sequential quadratic programming algorithm, and completes the temperature control of the power system and passenger compartment through the action of the actuator; and updates the coolant flow rate in the prediction equation to achieve linearization of the complex system and improve computational efficiency.

[0045] (2) The phased control strategy of the present invention takes both the battery pack temperature difference and the water pump energy consumption into consideration during design. The control parameters are optimized offline by using the particle swarm algorithm and drawing the Pareto boundary curve, thereby maximizing the comprehensive benefits of the battery pack temperature difference and the water pump energy consumption.

[0046] (3) The energy efficiency ratio prediction module of the present invention takes into account the impact of vehicle speed and ambient temperature on the efficiency of the air-conditioning system when evaluating energy consumption, and is based on offline training of a feedforward neural network, which can improve the accuracy of energy consumption evaluation and further improve energy saving efficiency.

[0047] (4) The present invention incorporates future driving conditions into the design of the thermal management controller, and introduces energy efficiency ratio prediction, battery pack and passenger compartment state estimation into the multi-objective optimization solver to improve the control accuracy and optimization space of the thermal management controller. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a schematic diagram of the structure of the electric vehicle thermal management control system considering energy saving and battery life according to the present invention;

[0049] Figure 2 This is a working diagram of the thermal management controller of the present invention;

[0050] Figure 3 A comparison diagram of the energy consumption curve and battery life reduction curve of the thermal management system when the control method of the present invention is used compared with the traditional method;

[0051] Figure 4 This is a comparison chart of the change in average battery temperature over time when the control method of the present invention is used and the traditional method is used. DETAILED DESCRIPTION

[0052] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0053] like Figure 1As shown, a thermal management control system for an electric vehicle that takes energy saving and battery life into consideration includes an operating condition prediction module, an energy efficiency ratio prediction module, a thermal management controller, a state monitoring module, and an execution module.

[0054] The operating condition prediction module includes a vehicle speed planning submodule and a road data analysis submodule. The operating condition prediction method is selected according to the current scenario, and the vehicle speed planning submodule or the road data analysis submodule is used to perform online rolling prediction of the vehicle speed; the vehicle speed planning submodule uses the surrounding vehicle information and traffic facility information obtained by the vehicle network to plan the future speed of the vehicle in a specific scenario through an optimization algorithm to obtain the future driving condition of the vehicle. The process of planning the future speed is the existing technology; the road data analysis submodule obtains the location information and driving characteristic parameter information of the vehicle and surrounding vehicles through the vehicle network in non-specific scenarios, and predicts the future driving condition of the vehicle. The prediction process is the existing technology.

[0055] The state monitoring module includes a state acquisition unit and a state storage unit. The state acquisition unit records the sensor information of each monitoring point (including the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b，1 、Nth single cell battery temperature T b，n , passenger compartment temperature T cab and coolant flow rate u), and is stored by the state storage unit.

[0056] The energy efficiency ratio prediction module is based on the energy efficiency ratio prediction model obtained through offline training, and receives the future driving condition V of the vehicle in the predicted time domain output by the working condition prediction module. pre and the heat load in the predicted time domain output by the state estimator and And convert the heat load into the required cooling power P need pass Convert the future driving conditions into a prediction sequence of the air intake volume, where δ is the relationship coefficient, V fan is the intake air volume when the condenser fan is on and the vehicle speed is 0; the required cooling power sequence and the predicted intake air volume sequence are used as inputs of the energy efficiency ratio prediction model. The energy efficiency ratio prediction model outputs the energy efficiency ratio prediction sequence COP, which is used as the dynamic coefficient of the objective function in the thermal management controller.

[0057] The energy efficiency ratio prediction model obtained through offline training specifically includes the following steps:

[0058] S1: Build an air conditioning system performance test bench including a refrigeration cycle test circuit including a compressor, condenser, expansion valve, chiller (battery cooler) and a coolant cycle test circuit including a battery pack, liquid cold plate, water pump, chiller, etc., and maintain a stable ambient temperature and select an appropriate air inlet volume v air Interval and required cooling power P need interval.

[0059] S2: Keep the minimum required cooling power unchanged, and each time select a value from the smallest to the largest in the air volume range to start the test. When the air conditioning system reaches a steady state, record the energy efficiency ratio of the compressor in this state. Repeat until the air volume reaches the maximum value, and you can get the corresponding relationship between the compressor energy efficiency ratio and the air volume at the current ambient temperature.

[0060] S3: Increase the required cooling power by an equal amount and repeat S2 until the required cooling power reaches the maximum value. Then, n groups of corresponding relationships can be obtained.

[0061] S4: Interchange the required cooling power (constant) and the air inlet volume (variable) of S2 and S3 above, and repeat the test to obtain m groups of corresponding relationships.

[0062] S5: The obtained n×m relationship points are combined into a nonlinear mapping function COP=Φ(v air P need ), a feedforward neural network is used for training to approximate the mapping function, thereby establishing an energy efficiency ratio prediction model under the ambient temperature.

[0063] S6: Change the ambient temperature and repeat S1 to S5 to establish an energy efficiency ratio prediction model under different ambient temperatures.

[0064] The thermal management controller includes a state estimator, a multi-objective optimization solver and a temperature difference controller. The state estimator receives the future driving condition V of the vehicle. pre , and predict the system state variables (including the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b，1 、Nth single cell battery temperature T b，n , passenger compartment temperature T cab ), the temperature of the No. 1 single cell T b，1 and the temperature of the nth single cell battery T b，n Take the average and get the average temperature T of the battery pack mean , the temperature of the No. 1 single cell T b，1 and the temperature of the nth single cell battery Tb，n Calculate the difference and get the battery pack temperature difference ΔT pre ; The multi-objective optimization solver is based on T mean , ΔT pre and T cab , construct the objective function and constraints of the problem to be optimized, and solve to obtain the optimal output power sequence of chiller and evaporator Among them, P chil is the optimal output power of chiller, P eva is the optimal output power of the evaporator, m is the control time domain; the temperature difference controller is based on the temperature difference ΔT of the battery pack pre The coolant flow rate u(t+1) at the next moment is determined based on the optimal output power sequence U(t), and the coolant flow rate in the prediction equation is updated.

[0065] The execution module receives the optimal output power sequence output by the multi-objective optimization solver and converts the first element of the optimal output power sequence into the control signal of the compressor, water pump and flow valve by the PID controller (wherein the compressor speed is used to track the chiller output power P chil The water pump speed is used to track the coolant flow rate u, and the flow valve opening is used to track the evaporator output power P eva ) and acts on compressors, water pumps and flow valves.

[0066] like Figure 2 As shown in FIG, the working process of the thermal management controller specifically includes the following steps:

[0067] Step (1), the state estimator uses the temperature of the No. 1 single cell T b1 、Nth single cell battery temperature T bn , coolant inlet temperature T f , Coolant outlet temperature T r and the passenger compartment temperature T cab is the state quantity, with the average temperature of the battery pack T mean , battery pack temperature difference ΔT pre and the passenger compartment temperature T cab is the output; after receiving the future driving condition V output by the working condition prediction module pre , according to the optimal output power sequence obtained by the multi-objective optimization solver at the previous moment, the vehicle speed prediction sequence is converted into the thermal load of the battery pack and the passenger compartment, and then the system state quantity is predicted based on the prediction equation, and the output quantity Y in the prediction time domain is calculated based on the output equation; the thermal load vector of the battery pack in the prediction time domain is The calculation process is shown in formula (1):

[0068]

[0069] in, is the current used to calculate the thermal load of the battery pack, P com is the vector of the predicted sequence of compressor output power, P dri is the vector of the predicted sequence of the output power of the driving motor, U m is the battery pack voltage, R bat is the internal resistance of the battery pack, vector P com The calculation process and vector P dri The calculation process of each element in is:

[0070]

[0071]

[0072] in, is a vector consisting of the inverse of the energy efficiency ratio prediction sequence output by the energy efficiency ratio prediction module; U(t-1) is the optimal control sequence output at the previous moment; p is the prediction time domain; m is the control time domain; is the predicted vehicle speed at the i-th second from the current moment, V pre The i-th element in F r is the rolling resistance; F a is the air resistance; η is the efficiency of the traction system.

[0073] Thermal load vector in the passenger compartment prediction domain The calculation process is shown in formula (2):

[0074]

[0075] in, is the solar thermal radiation power, is the external heat exchange power, is the occupant's thermal radiation power, and The calculation process is:

[0076]

[0077] Among them, K b is the heat transfer coefficient of the vehicle body, A b is the heat transfer area of ​​the vehicle body, ΔT b is the temperature difference between the vehicle body and the passenger compartment, ε w is the transmittance of sunlight, A w is the heat transfer area of ​​the car window, I is the solar radiation intensity, K w is the window heat transfer coefficient, ΔT w is the temperature difference between the window surface and the passenger compartment;

[0078]

[0079] Among them, ρ air is the air density, λ air is the air circulation rate, h o is the enthalpy of the air outside the vehicle, h i is the enthalpy of the air inside the vehicle;

[0080]

[0081] Among them, I driver is the driver's thermal radiation power, N pas is the number of passengers, β pas is the clustering coefficient.

[0082] The prediction equation and output equation are shown in equations (3) and (4):

[0083]

[0084] in, is the heat transfer coefficient between coolant and battery, P r is the Prandtl number, λ is the thermal conductivity of the coolant, v f is the kinematic viscosity of the coolant, d is the characteristic length of the heat exchange surface; A is the heat exchange area, u is the coolant flow rate, c b and c f are the specific heat capacities of the battery pack and coolant, m b is the mass of the battery pack, M is the mass of the coolant in the liquid cooling plate, s is the cross-sectional area of ​​the cooling pipe, ρ f is the coolant density, T s is the sampling time, n is the total number of single batteries, c air is the specific heat capacity of air, m air For air quality.

[0085]

[0086] Formulas (3) and (4) can be simplified as:

[0087] ζ(t+1)=Aζ(t)+Bu(t)+Cd(t) (5)

[0088] y(t+1)=Dζ(t+1) (6)

[0089] Each matrix satisfies:

[0090]

[0091]

[0092]

[0093] Change formulas (5) and (6) to incremental mode for prediction:

[0094] Δζ(t+1)=AΔζ(t)+BΔu(t)+CΔd(t) (7)

[0095] y(t+1)=DΔζ(t+1)+y(t) (8)

[0096] Substituting formula (7) into formula (8), we can get the output prediction of the next p steps at the current time t as:

[0097]

[0098]

[0099]

[0100]

[0101] Therefore, the system output Y in the prediction time domain is represented as:

[0102]

[0103] Step (2): The multi-objective optimization solver establishes energy consumption evaluation index, battery life evaluation index, passenger compartment comfort evaluation index and safety constraint conditions, balances the above performance indicators under dynamic working conditions based on adaptive weight parameters, and solves the multi-objective problem and its constraints online through the sequential quadratic programming algorithm to obtain the optimal output power sequence U(t), and judges the signal flag. c Is it 0? If not, directly output the optimal output power sequence U(t) and make the signal flag c Reset to 0, otherwise go to the next step.

[0104] The energy consumption evaluation index is the sum of the battery pack SOC loss within the prediction time domain: in is the current used to calculate the battery pack SOC at time i, C nom is the capacity of the battery pack.

[0105] The battery life evaluation index is the sum of the battery pack capacity loss percentages within the prediction time domain: Among them, A0, B0, and z are fitting coefficients obtained from the battery cycle life test; E a and R is a constant.

[0106] The passenger compartment comfort evaluation index is the sum of the squares of the errors between the passenger compartment temperature and the reference temperature in the prediction time domain:

[0107]

[0108] The safety constraints are implemented by the following inequalities:

[0109] T b_min <T mean (i|t)<T b_max +ξ(i|t)

[0110] 0<ΔT pre (i|t)<<T max

[0111] 0<P chil (i|t)<P chil_max

[0112] 0<P eva (i|t)<P eva_max

[0113] Where: T b_min is the lower limit of the average temperature of the battery pack, T b_max is the upper limit of the average temperature of the battery pack, ξ is the relaxation factor, ΔT max is the upper limit of the battery pack temperature difference, P chil_max is the upper limit of chiller output power, P eva_max It is the upper limit of the evaporator output power.

[0114] The weights of each evaluation index are assigned by the adaptive weight parameters, and the resulting multi-objective optimization problem and its constraints are shown in the following formula:

[0115]

[0116] stT b_min <T mean (i|t)<T b_max +ξ(i|t),i=0:p

[0117] 0<ΔT pre (i|t)<ΔT max , i=0:p

[0118] 0<P chil (i|t)<P chil_max , i=0:m-1

[0119] 0<P eva (i|t)<P eva_max , i=0:m-1

[0120] Y(0|t)=Y(t); U(0|t)=U(t)

[0121] w1 is the weight parameter of the energy consumption evaluation index, w2 is the weight parameter of the battery life evaluation index, w3 is the weight parameter of the comfort evaluation index, and w4 is the weight parameter of the relaxation factor in the constraint condition; the ratio of w1 and w3 is adaptively adjusted according to the battery SOC and the temperature difference in the passenger compartment, and is input to the multi-objective optimization solver by the fuzzy controller. w2 and w4 are fixed values ​​obtained from multiple experiments.

[0122] The adaptive weight parameters are as follows: when the battery SOC is smaller and the temperature difference in the passenger compartment is smaller, w1 / w3 is larger; when the battery SOC is larger and the temperature difference in the passenger compartment is larger, w1 / w3 is smaller.

[0123] w1 / w3 are output by the fuzzy controller, which takes the passenger compartment temperature difference and the battery SOC as input. The fuzzy control rules are shown in Table 1, which uses a double-antecedent multi-rule statement, namely "If X is A and Y is B, then Z is C":

[0124] Table 1 Fuzzy control rules

[0125]

[0126] In step (3), the state estimator obtains the optimal output power sequence U(t) obtained by the multi-objective optimization solver, replaces the matrix U(t-1) in step (1) with U(t), and replaces the matrix D with the matrix E = [-1 1 0 0 0]. After repeating step (1), the predicted output quantity Y′ in the time domain (i.e., the vector of battery pack temperature difference) is obtained:

[0127]

[0128] In step (4), the temperature difference controller obtains the vector of the temperature difference of the battery pack in the predicted time domain obtained by the state estimator, determines the flow rate at the next moment, and compares it with the historical coolant flow rate sequence to decide whether the coolant flow rate has changed. If it has changed, the coolant flow rate in the prediction equation is updated and the optimal output power sequence is re-solved. If it has not changed, the optimal output power sequence is output. The specific decision-making process is as follows:

[0129] The maximum value of the elements in Y′ is ΔT pre_max , ΔT pre_max Subtracting each element in the control parameter vector yields T = ΔT pre_max -T control =[T1, T2, ..., T i ,...,T j ], where T control =[T c1 , T c2 ,...,T cj] is the control parameter obtained offline, j is the number of elements in the vector composed of control parameters, T i is the i-th element in vector T; judge the size of each element in T, if T i With T i+1 T-compliant i+1 <0<T i , then the flow velocity at the next moment u(t+1)=U coo (i), where U coo =[u1, u2, ..., u j-1 ] is the vector composed of the determined flow velocities in each stage.

[0130] Control parameter T obtained offline control , specifically including the following steps:

[0131] S1: Build a battery cooling simulation platform including battery pack, cooling channel, and water pump;

[0132] S2: Design a phased control strategy, initialize the battery pack temperature difference control parameters, define the battery pack temperature difference and water pump energy consumption as the fitness function of the particle swarm algorithm, and initialize the weights of the two;

[0133] S3: Run the simulation, and the particle swarm algorithm calculates the fitness function based on the battery pack temperature difference series and the total energy consumption of the water pump returned by the simulation platform;

[0134] S4: Update the particle velocity and find the control parameters that minimize the fitness function through particle swarm iterative optimization;

[0135] S5: Change the weight value and find the control parameter T corresponding to the optimal weight based on the Pareto boundary curve of the battery pack temperature difference and the water pump energy consumption. control =[T c1 , T c2 ,...,T cj ].

[0136] After obtaining the flow rate value at the next moment, read the historical coolant flow rate sequence [u(t-τ), ..., u(t-1), u(t)] stored in the state storage unit, where u(t) is the coolant flow rate at the current moment, u(t-τ) is the coolant flow rate τ moments ago, and satisfies τ>20; compare u(t+1) with u(t): when u(t+1)>u(t), record the judgment signal flag u =20; when u(t+1)=u(t), record flag u =0; when u(t+1)<u(t), let q decrease from q=t until q=t-τ, compare u(t+1) and u(q), when u(t+1)<u(q), record flag u=tq, if u(t+1)<u(q) does not appear, record flag u =20;

[0137] When flag u ≥20, output the value of u(t+1) and update the coolant flow rate in the prediction equation, repeat steps (1) and (2), when flag u <20, let u(t+1)=u(t) and output the optimal output power U(t) obtained in step (2).

[0138] Step (5): at the next sampling moment, repeat steps (1) to (4).

[0139] Figure 3 The figure shows the comparison of thermal management energy consumption and battery life when the control method of the present invention is used and the traditional method is used. It can be seen from the figure that when the control method of the present invention is used, the energy consumed by the thermal management system under the UDDS cycle condition is 2.699×10 5 J, the battery life is reduced from 100% to 99.7506%. When using PID control, the energy consumed by the thermal management system is 2.824×10 5 J, the battery life is reduced from 100% to 99.7502%; therefore, under one UDDS cycle condition, the control method of the present invention can save 1.25×10 4 J energy consumption, increasing the battery life by 0.0004%. The prediction time domain of the method of the present invention is 120s, while the prediction time domain of other methods is about 20s. It can be seen that the prediction time domain of the thermal management controller of the present invention is significantly increased.

[0140] Figure 4 This graph compares the average battery temperature over time using the inventive control method and conventional control methods. As can be seen, the inventive control method allows for pre-cooling of the battery pack before reaching the 200-second high-load condition of the UDDS cycle, ensuring that the average battery pack temperature remains below the temperature limit. Compared to conventional control methods, the inventive control method more quickly returns the battery pack to its ideal state after each temperature rise. Therefore, the inventive control method effectively controls battery temperature.

[0141] The embodiments described are preferred implementations of the present invention, but the present invention is not limited to the above-mentioned implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essence of the present invention are within the scope of protection of the present invention.

Claims

1. A thermal management control method for electric vehicles considering energy saving and battery life, characterized by: S1, based on the vehicle's future driving conditions V pre , use the prediction equation to predict the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b,1 、Nth single cell battery temperature T b,n and the passenger compartment temperature T cab ; S2, based on the average temperature of the battery pack T mean , battery pack temperature difference △T pre and T cab , construct the objective function and constraints of the problem to be optimized, and solve to obtain the optimal output power sequence U(t) of the battery cooler and evaporator; the T mean By T b,1 and T b,n The average is obtained, the △T pre By T b,1 and T b,n To obtain by making a difference; S3, according to the temperature difference △T of the battery pack pre The optimal output power sequence U(t) is used to determine the coolant flow rate u(t+1) at the next moment. Based on the historical coolant flow rate, it is determined whether the coolant flow rate in the prediction equation needs to be updated. If so, the system returns to S1. Otherwise, the optimal output power sequence U(t) is output. The prediction equation is: in: is the heat transfer coefficient between coolant and battery, P r is the Prandtl number, λ is the thermal conductivity of the coolant, v f is the kinematic viscosity of the coolant, d is the characteristic length of the heat exchange surface; A is the heat exchange area, u is the coolant flow rate, c b and c f are the specific heat capacities of the battery pack and coolant, m b is the mass of the battery pack, M is the mass of the coolant in the liquid cooling plate, s is the cross-sectional area of ​​the cooling pipe, ρ f is the coolant density, T s is the sampling time, n is the total number of single batteries, c air is the specific heat capacity of air, m air is the air quality, T b,1 (t) is the current temperature of the No. 1 single cell battery, T b,n (t) is the current temperature of the nth single cell battery, T r (t) is the current temperature of the coolant outlet, T f (t) is the current temperature of the coolant inlet, T cab (t) is the current temperature of the passenger compartment, Predict the thermal load vector in the time domain for the battery pack, Predict the heat load vector in the time domain for the passenger compartment, P chil is the optimal output power of the battery cooler, P eva is the optimal output power of the evaporator.

2. The electric vehicle thermal management control method according to claim 1, characterized in that: The objective function and constraints of the problem to be optimized are: s.t.T b_min <T mean (i|t)<T b_max +ξ(i|t),i=0:p 0<ΔT pre (i|t)<ΔT max ,i=0:p 0<P chil (i|t)<P chil_max ,i=0:m-1 0<P eva (i|t)<P eva_max ,i=0:m-1 Y(0|t)=Y(t); U(0|t)=U(t) Among them: ω1 is the weight parameter of the energy consumption evaluation index, ω2 is the weight parameter of the battery life evaluation index, ω3 is the weight parameter of the comfort evaluation index, ω4 is the weight parameter of the relaxation factor in the constraint condition, P consumption is the energy consumption evaluation index, P life is the battery life evaluation index, P comfort is the passenger cabin comfort evaluation index, ξ is the relaxation factor, T b_min is the lower limit of the average temperature of the battery pack, T b_max is the upper limit of the average temperature of the battery pack, ξ is the relaxation factor, ΔT max is the upper limit of battery pack temperature difference, P chil_max is the upper limit of the battery cooler output power, P eva_max is the upper limit of the evaporator output power, p is the prediction time domain, t represents the current moment, m is the control time domain, and Y(t) is the output of the system at the current moment in the prediction time domain.

3. The electric vehicle thermal management control method according to claim 2, characterized in that: The ratio of ω1 to ω3 is adaptively adjusted according to the battery SOC and the passenger compartment temperature difference: when the battery SOC is smaller and the passenger compartment temperature difference is smaller, ω1 / ω3 is larger; when the battery SOC is larger and the passenger compartment temperature difference is larger, ω1 / ω3 is smaller.

4. The electric vehicle thermal management control method according to claim 1, characterized in that: The coolant flow rate u(t+1) at the next moment is determined as follows: The maximum value of the vector element of the battery pack temperature difference ΔT pre_max Subtracting each element in the control parameter vector yields T = ΔT pre_max -T control =[T1,T2,…,T i ,…,T j ], where T control =[T c1 ,T c2 ,…,T cj ] is the control parameter obtained offline, j is the number of elements in the vector composed of control parameters, T i is the i-th element in vector T; Determine the size of each element in T. When T i With T i+1 Meet T i+1 <0<T i , then the flow velocity at the next moment u(t+1)=U coo (i), where U coo =[u1,u2,…,u j-1 ] is the vector composed of the determined flow velocities in each stage.

5. The electric vehicle thermal management control method according to claim 4, characterized in that: The control parameter T obtained offline control , specifically including: (1) Build a battery cooling simulation platform including battery packs, cooling channels, and water pumps; (2) Design a phased control strategy, initialize the battery pack temperature difference control parameters, define the battery pack temperature difference and water pump energy consumption as the fitness function of the particle swarm algorithm, and initialize the weights of the two; (3) Run the simulation and calculate the fitness function based on the battery pack temperature difference sequence and the total energy consumption of the water pump returned by the simulation platform; (4) Update the particle velocity and find the control parameters that minimize the fitness function through particle swarm iterative optimization; (5) Change the weight and determine the control parameter T corresponding to the optimal weight based on the Pareto boundary curve of the battery pack temperature difference and the water pump energy consumption control =[T c1 ,T c2 ,…,T cj ].

6. The electric vehicle thermal management control method according to claim 1, characterized in that: The need to determine whether to update the coolant flow rate in the prediction equation is specifically: when the judgment signal flag u When ≥20, output the value of u(t+1) and update the coolant flow rate in the prediction equation.

7. The electric vehicle thermal management control method according to claim 6, characterized in that: The judgment signal is determined by comparing the size of u(t+1) and u(t): When u(t+1)>u(t), record the judgment signal flag u =20; When u(t+1)=u(t), record flag u =0; When u(t + 1) < u(t), let q start decreasing from q = t until q = t - τ, and compare the magnitudes of u(t + 1) and u(q). When u(t + 1) < u(q) occurs, record flag u = t - q. If u(t + 1) < u(q) does not occur, record flag u = 20; where τ is the number of historical coolant flow rates minus 1.

8. A system for implementing the electric vehicle thermal management control method according to any one of claims 1 to 7, characterized in that: include: The operating condition prediction module selects the vehicle speed planning submodule or the road data analysis submodule based on the current scenario to determine the vehicle's future operating conditions; An energy efficiency ratio prediction module converts the future driving conditions and the heat load in the prediction time domain into a prediction sequence of air intake volume and a cooling power sequence, inputs the sequence into an energy efficiency ratio prediction model, and outputs an energy efficiency ratio prediction sequence; The thermal management controller includes a state estimator, a multi-objective optimization solver and a temperature difference controller; the state estimator predicts the coolant inlet temperature T f , Coolant outlet temperature T r 、No. 1 single cell temperature T b,1 、Nth single cell battery temperature T b,n and the passenger compartment temperature T cab The multi-objective optimization solver constructs the objective function and constraints of the problem to be optimized and solves to obtain the optimal output power sequence; the temperature difference controller is used to determine the coolant flow rate u(t+1) at the next moment and update the coolant flow rate in the prediction equation; The state monitoring module includes a state acquisition unit and a state storage unit. The state acquisition unit records the sensor information of each monitoring point, and the state storage unit stores it; The execution module converts the first element of the optimal output power sequence into a control signal of the compressor, the water pump and the flow valve, and acts on the compressor, the water pump and the flow valve.

9. The system according to claim 8, characterized in that The future driving conditions are converted into a prediction sequence of the intake air volume, specifically: Where δ is the relationship coefficient, V fan The air volume when the condenser fan is on and the vehicle speed is 0. is the predicted vehicle speed at the i-th second from the current moment.

Images