Thermal management optimization method, system and equipment for power battery of intelligent networked automobile and medium

Abstract

The invention belongs to the technical field of new energy automobile battery thermal management, and particularly relates to an intelligent networked automobile power battery thermal management optimization method, system and device and a medium. Based on a model prediction control design controller, the influence of the length of a prediction range and future vehicle speed distribution on the performance of a battery thermal management system is analyzed; the optimal cooling power consumption of the thermal management system and the prediction time domain length required during temperature constraint are determined, multi-source space-time information on road traffic is obtained according to an intelligent network connection technology, a multi-source speed predictor is constructed based on a long-short-term memory neural network, and accurate prediction of the future speed of an automobile on a driving road is achieved. The thermal management system optimization method based on the self-adaptive optimization time domain is provided by combining a multi-time scale method according to the fact that a long prediction range is needed when the heat yield of a battery is large in the prediction range, so that the performance of different optimization time domains and different time scales is reasonably coordinated, and the calculation time is shortened under the condition that the optimization performance is not affected.

Description

Technical Field

[0001] The present invention belongs to the technical field of thermal management of new energy vehicle batteries, and in particular relates to a method, system, equipment and medium for optimizing thermal management of power batteries of intelligent networked vehicles. Background Art

[0002] With the continuous advancement of core technologies related to electric vehicles and the enhancement of necessary infrastructure, batteries have become a promising means to meet transportation needs. However, batteries generate heat during use, and high temperatures can significantly affect the performance, life and safety of the battery. In severe cases, they can even cause fires, explosions and other thermal runaway events. The battery temperature is actively controlled by the battery thermal management system, and a battery cooling control strategy needs to be carefully designed to achieve real-time, efficient and energy-saving cooling performance. Therefore, there is an urgent need for an effective electric vehicle battery thermal management system that can regulate the battery temperature and keep it within a safe range to extend battery life and improve vehicle performance. At the same time, the rapid development of intelligent networking and autonomous driving technologies has provided progress in different aspects of transportation. Through communication technologies such as vehicle to vehicle (V2V) and vehicle to infrastructure (V2I), cars can obtain the current motion status and collect traffic environment information. Based on this information, cars can achieve autonomous planning and prediction of future car speeds and road grades, thereby greatly reducing congestion, driving time and energy consumption, which is called eco-driving. However, most recent research on eco-driving has focused on reducing traction-related losses, while relatively little research has been conducted on battery thermal optimization related to traction power.

[0003] The current battery thermal management system is generally combined with the air conditioning system to achieve high cooling efficiency, which involves complex refrigerant phase change and electrothermal coupling processes, so a control-oriented model with sufficient accuracy and acceptable complexity is required. In addition, as the only power source of electric vehicles, the battery pack provides traction power and thermal management power. In high temperature environments, a large amount of thermal management energy is consumed when cooling the battery pack, which can reduce the range of electric vehicles by up to 59%. Therefore, optimizing the battery thermal management system is crucial to improving the energy efficiency of the entire vehicle. There have been many studies on battery cooling strategies, which can be divided into offline optimization methods and online optimization methods, such as lookup tables, fuzzy control, and model predictive control. Among them, predictive control has been widely used as an online optimization method that can utilize future information, and the development of intelligent networking and autonomous driving technologies has also promoted this choice.

[0004] In addition, thermal systems usually exhibit slow thermodynamics, that is, the battery thermal response has a large thermal inertia, and a long prediction range is required to obtain optimal performance. However, when the prediction range increases, the computational cost increases significantly. Therefore, in order to realize the online application of the control strategy, it is necessary to find a balance between thermal management performance and computational cost. The hierarchical framework is the main method to solve this problem. Through the different time scales and sampling times of the upper and lower layers, optimization within a longer prediction range can be achieved while having a smaller computational cost. However, for the hierarchical framework, due to the differences in time scales and equipment constraints, there is a potential risk that the upper-level scheduling results cannot be achieved by the lower-level scheduling tasks.

[0005] In summary, how connected electric vehicles can efficiently utilize intelligent network information and integrate it with battery pack thermal management to ensure that the battery is within a safe range while reducing the energy consumption of electric vehicles; how to determine the prediction range and key prediction information required by the thermal management system to improve the efficiency of the cooling system; and how to deal with the trade-off between long prediction range and computational cost to achieve real-time application of control strategies are issues that need to be urgently addressed. Summary of the invention

[0006] In view of the problems existing in the background technology, the purpose of the present invention is to provide a thermal management optimization method for power batteries of intelligent networked vehicles, based on multi-source speed predictors and adaptive optimization time domain model control under multiple time scales, to achieve minimum energy consumption and maintain battery temperature, so as to improve the mileage and driving safety of electric vehicles. The above purpose is achieved by the following technical solutions:

[0007] A method for optimizing thermal management of a power battery of an intelligent network-connected vehicle comprises the following steps:

[0008] Through intelligent networking, traffic information on the road is obtained, and a future vehicle speed prediction model based on the Long Short Term Memory (LSTM) neural network is constructed.

[0009] Step 1: Taking the compressor speed and water pump speed as the control variables, the battery temperature and coolant temperature as the state variables, the prediction range of the controller based on the battery thermal management system model predictive control when the controller obtains the best performance, the objective function is expressed as

[0010]

[0011] Among them, (k|t) represents k+i△t at time k s Prediction of time; △t s is the sampling time; H p =N△t s is the length of the prediction range; N represents the step length within the prediction range; T bat,min and T bat,maxare the upper and lower limits of the battery temperature; ε is the relaxation factor; τ is the weight of the relaxation factor; n com,min and n com,max is the minimum and maximum value of the compressor speed; n pump,min and n pump,max are the minimum and maximum values ​​of the pump speed.

[0012] Step 2: Input the historical actual speed and acceleration of the target car within the prediction range, the speed sequence on the future road section, the number of cars ahead, and the traffic light status at the intersection ahead into the LSTM-based speed predictor, and use the future car speed sequence as the output. The speed predictor is represented as

[0013] v pre (t+1:t+N p )=LSTM(X) (2)

[0014]

[0015] Among them, v pre is the predicted speed sequence; v is the historical speed sequence; a is the historical acceleration sequence; v p N is the speed sequence converted into time domain on the future road section; V is the number of cars ahead; L is the state sequence of the traffic lights at the intersection ahead;

[0016] Step 3: Calculate the future battery heat generation power according to the future vehicle speed. When a higher heat generation occurs within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is adopted. The cost function established in the optimization time domain is as follows:

[0017]

[0018] Among them, (k|t) represents k+i△t at time k s Prediction of the moment; H p,1 =N p,1 △t s,1 is the short time scale time domain length for accurately predicting vehicle speed; N p,1 is the total step length in the short time domain; △t s,1 is the sampling time of the short time scale; H p,2 =N p,2 △t s,2 is the long time scale time domain length of long-term vehicle speed prediction; N p,2 is the total step length in the long time domain, △t s,2 is the sampling time in the long time domain; H p,1 and H p,2 The total length is determined by the future heat production power.

[0019] As a better technical solution of the present invention: the battery thermal management system model is expressed as:

[0020]

[0021] Among them, T bat is the battery pack temperature; I bat is the battery pack current; R bat is the internal resistance of the battery pack; c c is the specific heat capacity of the coolant; is the mass flow rate of coolant in the battery pack cooling pipe; T c,in and T c,out The coolant temperature at the inlet and outlet of the battery pack cooling pipe; is the heat exchange rate between the battery pack and the external high temperature environment; m bat is the mass of the battery pack; c bat is the specific heat capacity of the battery pack, h b is the heat transfer coefficient between the coolant and the battery pack; A b is the heat exchange area between the battery pack and the coolant.

[0022] As a better technical solution of the present invention, the LSTM-based speed predictor optimization algorithm is the Adam algorithm, the number of hidden layers is set to 2, the number of hidden layer neurons is determined to be 10 according to empirical formulas and trial and error methods, the initial learning rate is set to 0.01 according to past experience, and the learning rate is set to gradually decrease.

[0023] As a more optimal technical solution of the present invention, the LSTM-based speed predictor selects the root mean square (RMSE) as an evaluation index of prediction performance.

[0024] As a more optimal technical solution of the present invention, the input data of the LSTM-based speed predictor is converted from distance to time by the following formula:

[0025] v(tt b :t+t f )=[v(tt b :t-1),v(t:t+t f )] (6)

[0026]

[0027]

[0028] Among them, v(tt b :t-1) is defined as the historical speed sequence; v(t:t+t f ) is the predicted future velocity series; is the speed in the future road section converted from the spatial domain to the time domain, d i , and are the position, time and speed of the marked points on the road section respectively.

[0029] Another object of the present invention is to provide a thermal management optimization system for power batteries of intelligent networked vehicles, comprising

[0030] Model predictive controller, based on the battery thermal management system model, analyzes the impact of the thermal management system on vehicle speed and predicts range under different driving conditions,

[0031] Multi-source speed predictor, which builds a speed prediction model based on network information. The historical actual speed and acceleration of the target car in the prediction range, the speed sequence on the future road section, the number of cars ahead and the traffic light status at the intersection ahead are input into the LSTM-based speed predictor, and the future car speed sequence is used as the output;

[0032] The battery heat generation power calculation module is used to calculate the battery heat generation power according to the future vehicle speed. When a higher heat generation occurs within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is adopted.

[0033] Another object of the present invention is to provide an electronic device, characterized in that it includes a memory and a processor, wherein the memory is used to store programs; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the above-mentioned intelligent connected vehicle power battery thermal management optimization method.

[0034] Another object of the present invention is to provide a computer-readable storage medium, characterized in that it is used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the above-mentioned intelligent connected vehicle power battery thermal management optimization method.

[0035] The present invention provides a battery thermal management system optimization method based on an LSTM multi-source speed predictor and a time-domain model predictive control with multi-time scale adaptive optimization to achieve minimum energy consumption and maintain battery temperature so as to improve the mileage and driving safety of electric vehicles.

[0036] Compared with the prior art, the present invention has the following advantages:

[0037] 1) For multi-input and multi-output battery thermal management systems, the controller is designed based on model predictive control to analyze the impact of the length of the prediction range and the future vehicle speed distribution on the performance of the battery thermal management system, and to determine the prediction time domain length required for the optimal cooling power consumption and temperature constraints of the thermal management system, so that the thermal management system during the driving of the electric vehicle can exhaust the battery temperature as much as possible and maintain it to improve the energy efficiency of the battery pack.

[0038] 2) Based on the intelligent network technology, multi-source spatiotemporal information on road traffic is obtained, and a multi-source speed predictor is constructed based on the long short-term memory neural network to accurately predict the future speed of the car on the driving road, reduce the impact of future speed prediction uncertainty on the optimization of the thermal management system, enable the controller to have a more accurate judgment on the system state, improve the controller's prediction ability and control effect, and realize the real-time performance of the optimization method.

[0039] 3) Based on the fact that a longer prediction range is required when the battery heat generation is large within the prediction range, a thermal management system optimization method based on adaptive optimization time domain is proposed in combination with a multi-time scale method. By adopting multiple time scales in the optimization time domain of model predictive control, the performance of different optimization time domains and different time scales can be reasonably coordinated to reduce the calculation time without affecting the optimization performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be described in detail below with reference to the following drawings, wherein:

[0041] Figure 1 This is a schematic diagram of the thermal management structure of an electric vehicle;

[0042] Figure 2 A flow chart of the thermal management optimization method for connected electric vehicle batteries;

[0043] Figure 3 This is the principle diagram of multi-source vehicle speed prediction based on LSTM neural network;

[0044] Figure 4 A schematic diagram of the adaptive optimization time domain with multiple time scales;

[0045] Figure 5 It is the curve diagram of vehicle speed change under UDDS working condition;

[0046] Figure 6 The battery temperature change comparison curve of this method and other methods;

[0047] Figure 7 This is the energy consumption comparison curve. DETAILED DESCRIPTION

[0048] The following is a detailed description of the embodiments of the present invention through specific examples, and those skilled in the relevant art can easily understand the advantages and innovations of the present invention based on the description. In addition, in order to demonstrate the effectiveness of the present invention, MATLAB / Simulink simulation is performed, and the results fully demonstrate the efficient performance of the control strategy.

[0049] In order to improve the energy utilization rate and the ability to maintain the battery temperature of the electric vehicle battery thermal management system in a high temperature environment, the present invention adopts an active liquid cooling system to cool the battery pack and an air conditioning system to reduce the temperature of the coolant. The structure of the battery thermal management system is as follows Figure 1 As shown in the figure, it mainly includes battery packs, cooling pipes, water pumps, compressors and other air conditioning system components. During operation, the air conditioning system realizes the cooling cycle of the refrigerant through the compressor, and realizes the heat exchange between the refrigerant and the coolant in the plate heat exchanger to reduce the coolant temperature. After that, the low-temperature coolant enters the battery pack cooling pipe through the water pump to absorb the heat of the battery to reduce the battery temperature.

[0050] Considering the large inertia of the battery thermal system and the impact of vehicle speed information on battery heat generation and thermal management system performance, a cooling optimization method for the battery thermal management system of a connected electric vehicle is proposed to improve optimization performance and reduce computational costs. First, based on the perfect prediction information under offline conditions, the sensitivity of the thermal management system to different prediction ranges and vehicle speeds is studied, and the prediction range length required to maintain optimal performance is determined; secondly, in order to cope with the uncertainty of long-term predicted speeds, based on future traffic information obtained by intelligent connected communication technology, a multi-source speed predictor based on long short-term memory neural network is designed to achieve accurate prediction of future speeds and obtain future battery heat generation power distribution. Finally, a battery thermal management system optimization method with multi-time scale adaptive optimization time domain model predictive control is designed to improve the energy economy of the battery thermal management system cooling process while reducing computational costs. The specific process is as follows: Figure 2 shown.

[0051] The technical solution proposed by the present invention is further described and illustrated below in conjunction with the accompanying drawings:

[0052] 1. Establish a control-oriented battery thermal management system model;

[0053] The battery of a connected electric vehicle will continue to generate heat during driving. According to the law of conservation of energy, the dynamic change of battery temperature can be expressed as

[0054]

[0055] Among them, T bat is the battery pack temperature; I bat is the battery pack current; Rbat is the internal resistance of the battery pack; c c is the specific heat capacity of the coolant; is the mass flow rate of coolant in the battery pack cooling pipe; T c,in and T c,out The coolant temperature at the inlet and outlet of the battery pack cooling pipe; is the heat exchange rate between the battery pack and the external high temperature environment; m bat is the mass of the battery pack; c bat is the specific heat capacity of the battery pack.

[0056] The coolant temperature at the cooling channel outlet can be calculated using the uniform wall temperature model.

[0057]

[0058] Among them, h b is the heat transfer coefficient between the coolant and the battery pack; A b is the heat exchange area between the battery pack and the coolant.

[0059] The coolant is driven by a water pump, and the mass flow rate of the coolant in the pipeline is

[0060]

[0061] Among them, V p is the volume of the pump; η p is the volumetric efficiency of the pump; n pump is the speed of the water pump; ρ co is the density of the coolant. The power consumption of the water pump can be calculated as

[0062]

[0063] Among them, P pump,m is the mechanical power, η m is the power transfer coefficient of the water pump, △p pump is the pressure drop across the pump.

[0064] The coolant exchanges heat with the refrigerant in the plate heat exchanger, thereby reducing the coolant temperature.

[0065]

[0066] Among them, m clnt It is expressed as the total mass of the coolant in the pipeline during the circulation process. Indicates the cooling capacity of the air conditioning system.

[0067] The present invention assumes that the cooling capacity of the air-conditioning system can be obtained from the compressor speed, and does not model the air-conditioning system in detail. The power consumption of the air-conditioning system can be calculated by fitting the cooling capacity and the coefficient of performance (COP) obtained from experimental data.

[0068]

[0069] Among them, P AC is the power consumption of the air conditioning system; COP(n com ) is the coefficient of performance of the air conditioning system.

[0070] As the only power source for electric vehicles, the power battery needs to provide the energy consumption of the thermal management system and the energy consumption of traction power. Therefore, the load current of the battery can be calculated as

[0071]

[0072] Among them, P bat Output power of the battery pack; U bat is the battery terminal voltage; P trac is the traction power of the connected electric vehicle, which can be calculated as

[0073]

[0074] Among them, v veh , and m veh are the speed, acceleration and mass of the electric vehicle respectively; F r and F a It is rolling friction resistance and air friction resistance.

[0075] Based on the Rint model, the terminal voltage of the battery pack can be expressed as

[0076] U bat =U oc -I bat R bat (17)

[0077] Among them, U oc is the open circuit voltage of the battery.

[0078] Based on formula (7) and formula (9), we can get

[0079]

[0080] It can be seen from formula (10) that the change of vehicle speed has an impact on the battery current, which in turn affects the heat generation and temperature change of the battery. In summary, the established control-oriented battery thermal management system model can be expressed as

[0081]

[0082] Among them, the battery temperature and coolant temperature are state variables, the compressor speed and water pump speed are control variables, and the vehicle speed is the external disturbance variable.

[0083] 2. Analyze the sensitivity of thermal management system to vehicle speed and prediction horizon length based on model predictive control;

[0084] The purpose of this invention is to keep the battery pack temperature of electric vehicles within a safe range and reduce the energy consumption of the thermal management system. The vehicle speed has an important influence on the battery temperature. In addition, due to the slow dynamic response of the battery thermal system, the thermal management system needs to be optimized in the long time domain to obtain the best control performance. However, a longer prediction range increases the computational cost of model predictive control. Therefore, it is necessary to analyze the sensitivity of the battery thermal management system to the vehicle speed and prediction range in order to improve the control performance of the system. The objective function of the designed controller prediction range based on model predictive control can be expressed as

[0085]

[0086] Among them, (k|t) represents k+i△t at time k s Prediction of time; △t s is the sampling time; H p =N△t s is the length of the prediction range; N represents the step length within the prediction range; T bat,min and T bat,max are the upper and lower limits of the battery temperature, which are 20°C and 30°C respectively. At this time, the battery has better performance and lower aging rate; ε is the relaxation factor, which allows the battery temperature T bat Temporary violation of the constraint; τ is the weight of the relaxation factor, which is used to penalize the battery temperature constraint violation, and its value is 10 6 ;n com,min and n com,max are the minimum and maximum values ​​of the compressor speed respectively; n pump,min and n pump,max are the minimum and maximum values ​​of the pump speed respectively.

[0087] Based on formula (11), the sensitivity of the thermal management system based on model predictive control to vehicle speed and prediction range is analyzed using perfect speed information under offline conditions. By adopting different driving conditions and designing different prediction range lengths, the impact of the prediction range length on the execution of energy consumption and temperature constraints is understood.

[0088] 3. Establish a multi-source speed predictor based on long short-term memory neural network;

[0089] The accuracy of future speed is crucial to the performance of thermal management system based on model predictive control, which determines the future heat generation of the battery. Accurate prediction speed can enable model predictive control to make more efficient decisions in advance and achieve ecological cooling. Considering the complexity of the traffic environment in actual driving environment, combined with intelligent network technology, a multi-source speed predictor based on LSTM neural network is designed. Its structure is as follows: Figure 3 As shown. Unlike conventional methods that only use speed sequences to train the network, the present invention uses multi-source data such as historical actual speed and acceleration, average speed on future road sections, number of cars ahead, and status of traffic lights at intersections to accurately predict future speeds. This method selects the Adam algorithm as the optimization algorithm for the LSTM neural network, the number of hidden layers is set to 2, the number of hidden layer neurons is determined to be 10 based on empirical formulas and trial and error, the initial learning rate is set to 0.01 based on past experience, and the learning rate is set to gradually decrease. The root mean square (RMSE) is selected as the evaluation indicator of prediction performance, which can be expressed as

[0090]

[0091] in, is the predicted value, y i is the actual value.

[0092] Since the average speed on the future road section is provided in the spatial domain, and the input and output of the established LSTM multi-source speed predictor are both time series. In order to ensure the consistency of the data, the present invention converts the traffic data from distance to time, which can be achieved by the following formula:

[0093] v(tt b :t+t f )=[v(tt b :t-1),v(t:t+t f )] (twenty two)

[0094]

[0095]

[0096] Among them, v(tt b :t-1) is defined as the historical speed sequence; v(t:t+t f ) is the predicted future velocity series; is the speed in the future road section converted from the spatial domain to the time domain, d i , and are the position, time and speed of the marked points on the road section respectively.

[0097] The designed LSTM-based multi-source speed predictor takes multi-source data as input and future speed sequences as output, which can be expressed as

[0098] v pre (t+1:t+N p )=LSTM(X) (25)

[0099]

[0100] Among them, LSTM is the established LSTM neural network speed predictor; v pre is the predicted speed sequence; v is the historical speed sequence; a is the historical acceleration sequence; v p N is the speed sequence converted into time domain on the future road section; V is the number of cars ahead; L is the state sequence of the traffic lights at the intersection ahead.

[0101] In addition, the future traction power prediction sequence of electric vehicles can be obtained by Newton's law of motion as shown in the following formula

[0102]

[0103]

[0104] 4. Thermal management system optimization strategy based on adaptive optimization time domain and multi-time scale

[0105] Although a longer prediction range brings better performance, it also increases the computational cost. In addition, it can be known from step two that when the battery heat generation in the future is large, a longer prediction range is needed to cool the battery in advance before the peak heat generation arrives to prevent the battery temperature from violating the constraint. When the heat generation is low, a longer prediction range and a shorter prediction range will have similar optimization effects. On the other hand, different sampling time scales will affect the calculation time and optimization effect. When the prediction time length is the same, a larger sampling time will have a smaller computational cost, but its optimization effect will be reduced, which has the opposite result when the sampling time is short. The present invention proposes a model predictive control optimization method based on an adaptive optimization time domain with multiple time scales, and its structure is as follows: Figure 2 Formula. The future vehicle speed is predicted based on the intelligent network information, and then the future battery heat generation power is calculated. When a higher heat generation occurs within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is selected. At the same time, the optimization time domain is composed of multiple time scales, and the time scale increases as the distance from the starting time increases, extending the length of the prediction range without increasing the relevant calculation burden of the controller. Specifically, in the previous H p,1 A shorter time scale △t is used in the s,1 , after Hp,2 A longer time scale △t s,2 Therefore, according to the above analysis, in order to ensure the cooling performance of the thermal management system and reduce the calculation cost, the objective function established in the optimization time domain is as follows:

[0106]

[0107] Among them, (k|t) represents k+i△t at time k s Prediction of the moment; H p,1 =N p,1 △t s,1 is the short time scale time domain length for accurately predicting vehicle speed; N p,1 is the total step length in the short time domain; △t s,1 is the sampling time of the short time scale; H p,2 =N p,2 △t s,2 is the long time scale time domain length of long-term vehicle speed prediction; N p,2 is the total step length in the long time domain, △t s,2 is the sampling time in the long time scale domain.

[0108] In order to better implement the intelligent networked vehicle power battery thermal management optimization method in the embodiment of the present invention, correspondingly, the embodiment of the present invention also provides an intelligent networked vehicle power battery thermal management optimization system, including a model prediction controller, which analyzes the impact of the thermal management system on the vehicle speed and prediction range under different driving conditions based on the battery thermal management system model; and a multi-source speed predictor, which establishes a vehicle speed prediction model based on network information, and inputs the historical actual speed and acceleration of the target vehicle within the prediction range, the speed sequence on the future road section, the number of vehicles ahead and the traffic light status at the intersection ahead into the LSTM-based speed predictor, and outputs the future vehicle speed sequence; and a battery heat generation power calculation module, which is used to calculate the battery heat generation power according to the future vehicle speed. When a higher heat generation appears within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is adopted.

[0109] The present invention also provides an electronic device accordingly. The electronic device includes a processor, a memory and a display. In some embodiments, the processor may be a central processing unit (CPU), a microprocessor or other data processing chip, which is used to run program codes stored in the memory or process data.

[0110] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions in the lithium-ion battery thermal management strategy optimization method provided in the above-mentioned method embodiments can be implemented.

[0111] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0112] The effectiveness of the present invention is further illustrated by simulation results on MATLAB. The Urban Dynamometer Driving Schedule (UDDS) in the United States is selected to verify the proposed optimization strategy. The speed curve of UDDS is as follows: Figure 5 Comparing the optimization method proposed in the present invention with the traditional MPC, the temperature change of the battery pack is shown in Figure 6 As shown, the energy consumption of the cooling system is Figure 7 As shown. Figure 6 It can be seen that the traditional MPC method exceeds the temperature constraint limit at 200s and then continues to rise rapidly. This is because the car is running at a high speed at this time, the battery heat generation power is high, and the cooling system cannot reduce the battery temperature in time. The method proposed in the present invention can cool the battery in advance and maintain the battery temperature well within the constraint range, thereby improving the temperature constraint execution capability. Figure 7 It can be seen that the energy consumption of traditional MPC is 314.61, while the energy consumption of MPC based on multi-source speed prediction and time-scale adaptive optimization is 209.12, and the energy consumption is reduced by 33.5%. In summary, the optimization strategy of the battery thermal management system of the networked electric vehicle proposed in the present invention can better maintain the battery temperature within the constraint range, reduce the energy consumption of the cooling system, and provide guarantee for the normal operation of the networked electric vehicle.

[0113] The optimization method, device, equipment and medium provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for optimizing thermal management of power batteries for intelligent networked vehicles, characterized in that: The steps include: Step 1: Taking the compressor speed and water pump speed as the control variables, the battery temperature and coolant temperature as the state variables, the prediction range of the controller based on the battery thermal management system model predictive control when the controller obtains the best performance, the objective function is expressed as Among them, (k|t) represents k+i△t at time k s Prediction of time; △t s is the sampling time; H p =N△t s is the length of the prediction range; N represents the step length within the prediction range; T bat,min and T bat,max are the upper and lower limits of the battery temperature; ε is the relaxation factor; τ is the weight of the relaxation factor; n com,min and n com,max is the minimum and maximum value of the compressor speed; n pump,min and n pump,max are the minimum and maximum values ​​of the pump speed. Step 2: Input the historical actual speed and acceleration of the target car within the prediction range, the speed sequence on the future road section, the number of cars ahead, and the traffic light status at the intersection ahead into the LSTM-based speed predictor, and use the future car speed sequence as the output. The speed predictor is represented as v pre (t+1:t+N p )=LSTM(X) (2) Among them, v pre is the predicted speed sequence; v is the historical speed sequence; a is the historical acceleration sequence; v p N is the speed sequence converted into time domain on the future road section; V is the number of cars ahead; L is the state sequence of the traffic lights at the intersection ahead; Step 3: Calculate the future battery heat generation power according to the future vehicle speed. When a higher heat generation occurs within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is adopted. The cost function established in the optimization time domain is as follows: Among them, (k|t) represents k+i△t at time k s Prediction of the moment; H p,1 =N p,1 △t s,1 is the short time scale time domain length for accurately predicting vehicle speed; N p,1 is the total step length in the short time domain; △t s,1 is the sampling time of the short time scale; H p,2 =N p,2 △t s,2 is the long time scale time domain length of long-term vehicle speed prediction; N p,2 is the total step length in the long time domain, △t s,2 is the sampling time in the long time domain; H p,1 and H p,2 The total length is determined by the future heat production power.

2. The intelligent networked vehicle power battery thermal management optimization method according to claim 1, characterized in that: The battery thermal management system model is expressed as Among them, T bat is the battery pack temperature; I bat is the battery pack current; R bat is the internal resistance of the battery pack; c c is the specific heat capacity of the coolant; is the mass flow rate of coolant in the battery pack cooling pipe; T c,in and T c,out The coolant temperature at the inlet and outlet of the battery pack cooling pipe; is the heat exchange rate between the battery pack and the external high temperature environment; m bat is the mass of the battery pack; c bat is the specific heat capacity of the battery pack, h b is the heat transfer coefficient between the coolant and the battery pack; A b is the heat exchange area between the battery pack and the coolant.

3. The intelligent networked vehicle power battery thermal management optimization method according to claim 1, characterized in that: The LSTM-based speed predictor optimization algorithm is the Adam algorithm, the number of hidden layers is set to 2, the number of hidden layer neurons is determined to be 10 based on empirical formulas and trial and error methods, the initial learning rate is set to 0.01 based on past experience, and the learning rate is set to gradually decrease.

4. The intelligent networked vehicle power battery thermal management optimization method according to claim 1, characterized in that: The LSTM-based speed predictor selects the root mean square RMSE as the evaluation indicator of prediction performance.

5. The intelligent networked vehicle power battery thermal management optimization method according to claim 1, characterized in that: The input data of the LSTM-based speed predictor is converted from distance to time by the following formula: v(t-t b :t+t f )=[v(t-t b :t-1),v(t:t+t f )] (6) Among them, v(tt b :t-1) is defined as the historical speed sequence; v(t:t+t f ) is the predicted future speed sequence; v(t di :t di+n ) is the speed in the future section converted from the spatial domain to the time domain, d i , t di and v di are the position, time and speed of the marked points on the road section respectively.

6. A thermal management optimization system for power batteries of intelligent networked vehicles, comprising: Model predictive controller, based on the battery thermal management system model, analyzes the impact of the thermal management system on vehicle speed and predicts range under different driving conditions, Multi-source speed predictor, which builds a speed prediction model based on network information. The historical actual speed and acceleration of the target car in the prediction range, the speed sequence on the future road section, the number of cars ahead and the traffic light status at the intersection ahead are input into the LSTM-based speed predictor, and the future car speed sequence is used as the output; The battery heat generation power calculation module is used to calculate the battery heat generation power according to the future vehicle speed. When a higher heat generation occurs within the prediction range, a longer optimization time domain is adopted, otherwise a shorter optimization time domain is adopted.

7. An electronic device, characterized in that: It includes a memory and a processor, wherein the memory is used to store programs; the processor is coupled to the memory and is used to execute the programs stored in the memory to implement the steps in the method for optimizing thermal management of power batteries of intelligent connected vehicles as claimed in claim 1.

8. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the intelligent connected vehicle power battery thermal management optimization method as claimed in claim 1.

Images