A hybrid electric vehicle energy management method and system based on multi-objective optimization

Abstract

This disclosure belongs to the field of energy management technology, specifically to a hybrid electric vehicle energy management method and system based on multi-objective optimization. The method comprises the following steps: obtaining hybrid electric vehicle operating data and constructing a hybrid electric vehicle model; establishing a multi-objective optimized hybrid electric vehicle energy management model based on the constructed hybrid electric vehicle model, and solving the weight coefficients between the multiple objectives using a sum-of-squares optimal method; and solving the hybrid electric vehicle energy management model using a proximal strategy optimization algorithm to implement hybrid electric vehicle energy management. This disclosure fully considers the temperature of the power battery and introduces an energy management framework based on a proximal strategy optimization algorithm to improve fuel economy, maintain a stable power battery state of charge (SOC), and reduce battery temperature rise.

Description

Technical Field

[0001] The present disclosure belongs to the field of energy management technology, and in particular relates to a hybrid electric vehicle energy management method and system based on multi-objective optimization. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Traditional internal combustion engine vehicles consume large amounts of non-renewable energy and produce a large amount of polluting gases. Electric vehicles offer an effective way to alleviate these issues, but due to technical limitations of batteries, the development of pure electric vehicles has reached a bottleneck. Conversely, hybrid vehicles have played a key role in promoting the electrification of the automotive industry.

[0004] Hybrid electric vehicles (HEVs) have complex powertrains and additional degrees of freedom, making their energy management more complex. Machine learning algorithms, with their adaptability and fast computational speed, can achieve continuous, real-time control. Consequently, learning-based methods, particularly reinforcement learning algorithms, have gained widespread application in recent years. Typical reinforcement learning algorithms not only address the decision-making optimization problem of continuous actions but also take into account long-term rewards and have a long-term perspective. However, existing reinforcement learning algorithms for HEV energy management have numerous hyperparameters, which can easily lead to training instability.

[0005] According to the inventors, existing energy management methods, when designing objective functions, give more consideration to the vehicle's fuel economy and SOC stability, while ignoring the temperature of the power battery. The battery's charge and discharge capacity, cycle period, and safety are all affected by its temperature. In addition, when solving multi-objective optimization problems, the optimal solution for one objective may be mutually exclusive with the optimal solutions for other objectives, so it is necessary to continuously adjust the weight coefficients to achieve a balance among multiple objectives. For the weight coefficients of multiple objectives, researchers usually use empirical methods to determine them. This method has too many subjective factors and cannot achieve the optimal trade-off. Summary of the Invention

[0006] In order to solve the above problems, the present disclosure proposes a hybrid electric vehicle energy management method and system based on multi-objective optimization, which fully considers the temperature of the power battery and introduces an energy management framework based on a proximal strategy optimization algorithm to improve fuel economy, maintain the stability of the power battery state of charge (SOC) and reduce battery temperature rise; at the same time, in order to achieve the optimal trade-off between multiple objectives, the square sum optimal solution is introduced to allocate the weight coefficients between each objective.

[0007] According to some embodiments, a first solution of the present disclosure provides a hybrid electric vehicle energy management method based on multi-objective optimization, which adopts the following technical solutions:

[0008] A hybrid electric vehicle energy management method based on multi-objective optimization includes the following steps:

[0009] Obtain hybrid vehicle operating data and build a hybrid vehicle model;

[0010] Based on the constructed hybrid vehicle model, a multi-objective optimized energy management model of the hybrid vehicle is established, and weight coefficients between the multiple objectives are solved by a sum-of-squares optimal method;

[0011] The energy management model of the hybrid vehicle is solved by a proximal strategy optimization algorithm to realize the energy management of the hybrid vehicle.

[0012] As a further technical limitation, the operating modes of the hybrid vehicle include a pure electric mode, an engine-only working mode, a motor and engine mixed action mode, and a regenerative braking mode.

[0013] As a further technical limitation, the operating data of the hybrid vehicle includes vehicle speed, initial temperature of the power battery and state of charge of the power battery.

[0014] Furthermore, based on the acquired vehicle speed, the acceleration and required torque of the vehicle are obtained; the vehicle speed, acceleration, power battery state of charge and power battery initial temperature are used as inputs to the proximal strategy optimization algorithm to obtain the engine torque at the current moment, and then solve the motor torque; according to the vehicle model, the power battery state of charge and power battery temperature after taking the torque are calculated.

[0015] As a further technical limitation, the reward value corresponding to the current engine torque is calculated based on the vehicle's fuel consumption, the power battery charge state at the next moment, and the power battery temperature. The optimal result is output based on the principle that the greater the reward, the better, to achieve energy management.

[0016] As a further technical limitation, the objective function of energy management includes fuel economy, stability of the power battery state of charge, and temperature of the power battery, and the optimal weight coefficient is determined according to the square sum optimal method;

[0017] That is, a single objective is set as the objective function, and the optimal torque distribution result and the corresponding first target value under the current objective function are calculated; then a new evaluation function is constructed based on the optimal target value solved by the single objective, and the corresponding optimal torque distribution result and the corresponding second target value under the constructed new evaluation function are solved; according to the weight coefficient calculation formula, the weight coefficient values corresponding to the three objective functions are calculated.

[0018] As a further technical limitation, the reward function is determined according to the objective function of the energy management model. By setting the reward function, the reward value corresponding to each action value is calculated. According to the principle of finding actions with increasingly larger reward values and the parameter update formula, the parameters of the neural network are continuously updated, and the output reward value converges to the maximum action value, that is, the optimal result that satisfies the objective function.

[0019] According to some embodiments, a second solution of the present disclosure provides a hybrid electric vehicle energy management system based on multi-objective optimization, which adopts the following technical solutions:

[0020] A hybrid electric vehicle energy management system based on multi-objective optimization, comprising:

[0021] an acquisition module configured to acquire operating data of the hybrid vehicle and construct a hybrid vehicle model;

[0022] a modeling module configured to establish a multi-objective optimized energy management model for the hybrid vehicle based on the constructed hybrid vehicle model, and to solve weight coefficients among the multiple objectives by using a sum-of-squares optimal method;

[0023] The energy management module is configured to solve the energy management model of the hybrid vehicle through a proximal strategy optimization algorithm to achieve energy management of the hybrid vehicle.

[0024] According to some embodiments, a third solution of the present disclosure provides a computer-readable storage medium, which adopts the following technical solution:

[0025] A computer-readable storage medium stores a program, which, when executed by a processor, implements the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in the first aspect of the present disclosure.

[0026] According to some embodiments, a fourth solution of the present disclosure provides an electronic device, which adopts the following technical solution:

[0027] An electronic device comprises a memory, a processor and a program stored in the memory and executable on the processor. When the processor executes the program, the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in the first aspect of the present disclosure are implemented.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The present disclosure fully considers the temperature of the power battery, introduces an energy management framework based on the proximal strategy optimization algorithm, solves the weight coefficients between multiple objectives based on the square sum optimal method, improves fuel economy, maintains the stability of the power battery SOC and reduces the battery temperature rise. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings, which constitute a part of the present disclosure, are used to provide a further understanding of the present disclosure. The exemplary embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation to the present disclosure.

[0031] Figure 1 is a flow chart of a hybrid electric vehicle energy management method based on multi-objective optimization in the first embodiment of the present disclosure;

[0032] Figure 2 is a structural diagram of the hybrid vehicle system configuration in the first embodiment of the present disclosure;

[0033] Figure 3 This is the engine fuel consumption map in the first embodiment of the present disclosure;

[0034] Figure 4 This is the motor efficiency map in the first embodiment of the present disclosure;

[0035] Figure 5 is a schematic diagram of the structure of the energy management model in the first embodiment of the present disclosure;

[0036] Figure 6 is an interaction diagram of the agent and the environment in the first embodiment of the present disclosure;

[0037] Figure 7 This is a diagram of the energy management framework based on the proximal strategy optimization algorithm in the first embodiment of the present disclosure;

[0038] Figure 8 is a comparative schematic diagram of the battery temperature rise curve in the first embodiment of the present disclosure;

[0039] Figure 9 This is a structural block diagram of a hybrid vehicle energy management system based on multi-objective optimization in the second embodiment of the present disclosure. DETAILED DESCRIPTION

[0040] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0041] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present disclosure belongs.

[0042] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0043] In the absence of conflict, the embodiments of the present disclosure and the features thereof may be combined with each other.

[0044] Example 1

[0045] The first embodiment of the present disclosure introduces a hybrid electric vehicle energy management method based on multi-objective optimization.

[0046] like Figure 1 A hybrid electric vehicle energy management method based on multi-objective optimization is shown, comprising the following steps:

[0047] Obtain hybrid vehicle operating data and build a hybrid vehicle model;

[0048] Based on the constructed hybrid vehicle model, a multi-objective optimized energy management model of the hybrid vehicle is established, and weight coefficients between the multiple objectives are solved by a sum-of-squares optimal method;

[0049] The energy management model of the hybrid vehicle is solved by a proximal strategy optimization algorithm to realize the energy management of the hybrid vehicle.

[0050] The energy management method proposed in this embodiment, as one or more implementation methods, includes four steps: hybrid vehicle modeling, construction of an energy management model based on a proximal policy optimization algorithm, determination of weight coefficients, and algorithm solution. The algorithm's operating environment is established by modeling the vehicle, followed by modeling and solving the temperature estimation equation for the power battery. Based on this, the principles of the reinforcement learning algorithm are described, and an energy management framework is established. A new evaluation function is constructed using the sum-of-squares ideal point method, and the weight coefficients are analytically derived. The reward function is determined, and the algorithm is finally solved.

[0051] (1) Hybrid Electric Vehicle Modeling

[0052] Considering the practical versatility, this embodiment models a single-axle parallel hybrid vehicle. Figure 2 There are four operating modes for the vehicle: pure electric mode, engine-only mode, motor-engine hybrid mode, and regenerative braking mode.

[0053] The dynamic equation of the vehicle can be described as:

[0054]

[0055] Where, F t is the driving force of the vehicle, m is the vehicle mass, g is the acceleration due to gravity, f is the rolling resistance coefficient, C d is the air resistance coefficient, A is the frontal area of the vehicle, δ is the correction factor, and v is the vehicle speed.

[0056] 1) Engine model

[0057] The fuel consumption rate of an engine at a fixed operating point is only related to its speed and torque. Based on experimental data, the fuel consumption rate can be obtained through two-dimensional interpolation. The fuel consumption map of the engine is as follows: Figure 3 The fuel consumption of the engine at each moment can be calculated using the following formula:

[0058]

[0059] Where b e is the fuel consumption rate per unit time, T e is the engine torque, n e is the engine speed.

[0060] 2) Motor model

[0061] The motors of the hybrid electric vehicles involved in this embodiment are all permanent magnet synchronous motors. According to formula (3), the operating efficiency of the motor can also be obtained by interpolation. The operating efficiency map of the motor is as follows: Figure 4 As shown. The output power of the motor can be calculated:

[0062] η m (n m ,T m )=f(n m ,T m ) (3)

[0063]

[0064] Where η m is the motor’s operating efficiency, T m is the motor torque, n m is the motor speed.

[0065] 3) Battery electrical model

[0066] The power balance equation of the battery system is as follows:

[0067] P bat =P b +P l =P b+I 2 R b (5)

[0068] Where, P bat is the total power consumption of the battery, P b is the power flowing into and out of the battery, P l is the power loss caused by the internal resistance of the battery, R b It is the internal resistance.

[0069] The calculation formula of SOC is as follows:

[0070]

[0071] Where U oc is the open circuit voltage, Q b is the rated capacity.

[0072] 4) Battery thermal model

[0073] The temperature rise of power batteries is usually caused by two reasons: one is the heat generated by ohmic internal resistance, and the other is the heat generated by the chemical reaction inside the battery. Among them, the heat generated by ohmic internal resistance is irreversible, while the heat generated by the chemical reaction inside the battery is reversible. The heat generation formula is as follows:

[0074]

[0075] The battery thermal model is established according to the law of conservation of energy. The heat balance equation can be expressed as:

[0076]

[0077] Where m b is the battery mass, c b is the average specific heat capacity of the battery, A b is the heat exchange coefficient of the battery, Q h is the heat generation rate of the battery, T en is the ambient temperature, T bat is the battery temperature, I bat is the charge and discharge current, U t is the terminal voltage, and h is the heat transfer coefficient.

[0078] After equivalent transformation, the temperature estimation equation can be expressed as:

[0079]

[0080] (2) Proximal Strategy Optimization Algorithm

[0081] The process from the agent starting from a specific state to the end of the task is called a complete scene. In a scene with T time, the agent continuously interacts with the environment, forming the following sequence:

[0082] τ={s1,a1,s2,a2,…s T ,a T} (10)

[0083] Since the actions taken by the agent in different states may be different, the sequence is uncertain. If the current strategy is π θ When , the probability of the sequence occurring is:

[0084]

[0085] The reward of sequence τ is the sum of the rewards in each stage, which is R(τ).

[0086] Therefore, the expected reward can be obtained as follows:

[0087]

[0088] To maximize the expected reward, the policy function π θ Continuously adjust through the policy gradient method. The policy gradient solution process is as follows:

[0089]

[0090] To avoid using the same reward for all data in the same scenario, an advantage function is introduced to convert rewards into values related to state and action. Furthermore, the proximal policy optimization algorithm also incorporates importance sampling, which allows for the reuse of sampled data and improves training speed.

[0091] The idea of importance sampling is as follows:

[0092]

[0093] A θ (s t ,a t )=Q θ (s t ,a t )-V θ (s t ) (15)

[0094]

[0095]

[0096] The application of the importance sampling method makes the algorithm contain two strategies, which usually requires that the parameters of the two strategies cannot be too far apart, so p θ (s t ) and p θ' (st ) are considered equal. The objective function of the algorithm becomes:

[0097]

[0098] When performing importance sampling, we expect the parameters of the two networks to be close. When the same state is input, the action probability distributions obtained by the networks should be highly similar. Therefore, the proximal policy optimization algorithm introduces a Clip function to clip the importance weights to an acceptable range. The updated objective function of the algorithm becomes:

[0099]

[0100] (3) Energy management framework based on proximal strategy optimization algorithm

[0101] Reinforcement learning is used to solve the problem of maximizing rewards by continuously learning and updating strategies during the interaction between an agent and its environment. Reinforcement learning algorithms treat learning as a trial-and-error process. The goal of the system is to dynamically adjust parameters to maximize rewards. First, the agent randomly selects an action and applies it to the environment. After the environment receives the action, its state changes, and a reward is generated and fed back to the agent. Based on the current state of the environment and the size of the reward, the agent selects its next action. The principle of selection is to increase the probability of receiving a larger reward in the future.

[0102] like Figure 5 The energy management model shown first calculates the required torque at the current vehicle speed using the required torque formula. The proximal strategy optimization algorithm then randomly outputs an action value (engine torque). The torque relationship and the upper and lower limits of the torque output are used to solve for the motor torque. The fuel consumption map, SOC calculation formula, and temperature estimation formula are then used to calculate the vehicle's equivalent fuel consumption, SOC, and temperature at the next moment, given the engine and motor torques. Based on the vehicle's fuel consumption, SOC, and temperature at the next moment, the reward value corresponding to the current action value (i.e., engine torque) is calculated. Following the principle that the greater the reward, the better, new action values are continuously explored until the reward value converges to the maximum value, outputting the optimal result.

[0103] In this embodiment, the agent is the energy management controller and the environment is the power system of a hybrid bus. The agent's task is to reasonably allocate energy flow to achieve the optimal operating result while meeting the constraints. The interaction process between the agent and the environment is as follows: Figure 6 shown.

[0104] Aiming at the energy management problem of hybrid electric vehicles, the three elements of reinforcement learning algorithm are defined: state, action and reward.

[0105] State s: In this embodiment, considering that the temperature change of the battery has a certain impact on driving safety, the vehicle speed, acceleration, SOC and battery temperature are set as the state inputs of the hybrid drive system.

[0106] s=[v,acc,SOC,T bat ] (20)

[0107] Where acc is the acceleration of the vehicle, T bat is the battery temperature.

[0108] Action a: To reduce the amount of calculation, the present invention defines only one action, namely the engine torque. Then, the motor torque is obtained according to formula (22).

[0109] a=[T e ] (twenty one)

[0110]

[0111] Where, T req is the required torque, i g is the transmission ratio, i o is the final reduction ratio, r wheel is the tire radius and η is the transmission efficiency.

[0112] Reward r: In this example, the reward function is not only focused on achieving optimal fuel economy and maintaining SOC stability, but also strives to slow down the temperature rise of the battery. When the agent is learning, it tends to explore action values with greater rewards. Therefore, in this example, the reward function needs to be negative, that is:

[0113]

[0114] Where α, β, and γ are weight coefficients.

[0115]

[0116] Where, SOC fin It is the lower limit of SOC, and is set to 0.3 in the present invention.

[0117]

[0118] Where, T high The upper limit of the battery temperature is set to 40°C in the present invention. The operating temperature range of the power battery is generally 0-40°C.

[0119] The weight coefficient is determined by the square sum optimal solution method. The process is as follows: first, the optimal solution (ideal point) of the single-objective optimization problem is solved, and the evaluation function is constructed by the square sum ideal point method; then, the optimal solution of the evaluation function corresponding to the ideal point or close to the ideal point is found; finally, the optimal solution is analyzed and the weight coefficient is derived. For the convenience of description, let f1 = m fuel , f2=f SOC , The specific steps are as follows:

[0120] (1) Solve the optimal solutions of the three single-objective optimization problems and record them as ideal points

[0121] (2) Constructing the evaluation function by the method of square sum ideal point Find the optimal solution of the evaluation function remember

[0122] (3) According to the formula Analyze the optimal solution of the evaluation function, determine the proportion of each single target value close to the ideal point, and thus determine the weight coefficient.

[0123] The Proximal Policy Optimization algorithm is based on an Actor-Critic architecture, an online learning network with relatively slow parameter updates. To improve the efficiency of this network, the Proximal Policy Optimization algorithm introduces an Actor network, separating the online training agent from the agent interacting with the environment, thus speeding up the algorithm's training.

[0124] The Actor-Critic network is a fully connected neural network with 100 neurons and 1 hidden layer. For the Actor network, the output function is the Sigmoid function to limit the output action value to the range [0, 1]. The Sigmoid function is:

[0125] Figure 7 This paper demonstrates how the proximal policy optimization algorithm can be applied to the energy management problem of hybrid electric vehicles. At the beginning of a journey, the agent interacts with the environment to obtain the current vehicle speed, acceleration, battery temperature, and state of charge (SOC). To eliminate the influence of numerical variables, the obtained state values are normalized. The action network then makes decisions based on the current state and calculates the reward value for each decision. Through continuous iteration, a series of trajectories are obtained. The critic network and another actor network use these trajectories to update their parameters. Parameter updates are generally based on backpropagation and the advantage function. After several iterations, the optimal energy allocation result is finally achieved. The pseudo code is shown in Table 1.

[0126] Table 1 Pseudocode of proximal strategy optimization algorithm

[0127]

[0128]

[0129] (4) Example simulation

[0130] The method proposed in this embodiment is verified by using Python software simulation. The simulation results are shown in Table 2.

[0131] Table 2 Simulation comparison of energy management methods

[0132]

[0133] From the comparison of simulation results in Table 2, it can be seen that compared with other reinforcement learning-based methods, the method proposed in this invention shows great advantages in improving fuel economy, slowing down battery temperature rise, and reducing calculation time, which proves the superiority and rationality of the proposed strategy.

[0134] like Figure 8 As shown, compared with the energy management strategy based on the proximal strategy optimization algorithm that does not consider the battery temperature, the final temperature of the battery of the method proposed in this embodiment dropped by 1.509°C, which proves that the method proposed in this embodiment has a good effect in managing the battery temperature.

[0135] Example 2

[0136] The second embodiment of the present disclosure introduces a hybrid electric vehicle energy management system based on multi-objective optimization.

[0137] like Figure 9 A hybrid electric vehicle energy management system based on multi-objective optimization is shown, comprising:

[0138] an acquisition module configured to acquire operating data of the hybrid vehicle and construct a hybrid vehicle model;

[0139] a modeling module configured to establish a multi-objective optimized energy management model for the hybrid vehicle based on the constructed hybrid vehicle model, and to solve weight coefficients among the multiple objectives by using a sum-of-squares optimal method;

[0140] The energy management module is configured to solve the energy management model of the hybrid vehicle through a proximal strategy optimization algorithm to achieve energy management of the hybrid vehicle.

[0141] The detailed steps are the same as those of the hybrid electric vehicle energy management method based on multi-objective optimization provided in the first embodiment, and are not described again here.

[0142] Example 3

[0143] A third embodiment of the present disclosure provides a computer-readable storage medium.

[0144] A computer-readable storage medium stores a program, which, when executed by a processor, implements the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in the first embodiment of the present disclosure.

[0145] The detailed steps are the same as those of the hybrid electric vehicle energy management method based on multi-objective optimization provided in the first embodiment, and are not described again here.

[0146] Example 4

[0147] A fourth embodiment of the present disclosure provides an electronic device.

[0148] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in the first embodiment of the present disclosure are implemented.

[0149] The detailed steps are the same as those of the hybrid electric vehicle energy management method based on multi-objective optimization provided in the first embodiment, and are not described again here.

[0150] The foregoing description is merely a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art will readily appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present disclosure shall be included within the scope of protection of the present disclosure.

Claims

1. A hybrid electric vehicle energy management method based on multi-objective optimization, characterized in that: The following steps are involved: Obtain hybrid vehicle operating data and build a hybrid vehicle model; Based on the constructed hybrid vehicle model, a multi-objective optimized energy management model of the hybrid vehicle is established, and weight coefficients between the multiple objectives are solved by a sum-of-squares optimal method; The weight coefficient is determined by using the square sum optimal solution method. The process is as follows: first, the optimal solution of the single-objective optimization problem is solved, and the evaluation function is constructed by the square sum ideal point method; then, the optimal solution of the evaluation function corresponding to the ideal point or close to the ideal point is found; finally, the optimal solution is analyzed and the weight coefficient is derived; Solving the energy management model of the hybrid vehicle by using a proximal strategy optimization algorithm to achieve energy management of the hybrid vehicle; The objective function of energy management includes fuel economy, stability of the power battery state of charge, and power battery temperature. The optimal weight coefficient is determined based on the square sum optimal method. That is, a single objective is set as the objective function, and the optimal torque distribution result and the corresponding first target value under the current objective function are calculated; then a new evaluation function is constructed based on the optimal target value solved by the single objective, and the corresponding optimal torque distribution result and the corresponding second target value under the constructed new evaluation function are solved; according to the weight coefficient calculation formula, the weight coefficient values corresponding to the three objective functions are calculated.

2. A hybrid electric vehicle energy management method based on multi-objective optimization as claimed in claim 1, characterized in that: The operation modes of the hybrid electric vehicle include a pure electric mode, an engine-only operation mode, a motor and engine mixed operation mode, and a regenerative braking mode.

3. The hybrid electric vehicle energy management method based on multi-objective optimization as claimed in claim 1, characterized in that: The operating data of the hybrid vehicle includes vehicle speed, initial temperature of the power battery and state of charge of the power battery.

4. A hybrid electric vehicle energy management method based on multi-objective optimization as claimed in claim 3, characterized in that: Based on the acquired vehicle speed, the vehicle's acceleration and required torque are obtained; the vehicle speed, acceleration, power battery state of charge, and power battery initial temperature are used as inputs to the proximal strategy optimization algorithm to obtain the current engine torque and then solve the motor torque; based on the vehicle model, the power battery state of charge and power battery temperature after adopting the torque are calculated.

5. A hybrid electric vehicle energy management method based on multi-objective optimization as claimed in claim 4, characterized in that: Based on the vehicle's fuel consumption, the battery state of charge at the next moment, and the battery temperature, the reward value corresponding to the current engine torque is calculated. Based on the principle that the greater the reward, the better, the optimal result is output to achieve energy management.

6. A hybrid electric vehicle energy management method based on multi-objective optimization as claimed in claim 1, characterized in that: The reward function is determined according to the objective function of the energy management model. By setting the reward function, the reward value corresponding to each action value is calculated. According to the principle of finding actions with increasingly larger reward values and the parameter update formula, the parameters of the neural network are continuously updated, and the output reward value converges to the maximum action value, that is, the optimal result that satisfies the objective function.

7. A hybrid electric vehicle energy management system based on multi-objective optimization, characterized in that: include: an acquisition module configured to acquire operating data of the hybrid vehicle and construct a hybrid vehicle model; a modeling module configured to establish a multi-objective optimized energy management model for the hybrid vehicle based on the constructed hybrid vehicle model, and to solve weight coefficients among the multiple objectives by using a sum-of-squares optimal method; The weight coefficient is determined by using the square sum optimal solution method. The process is as follows: first, the optimal solution of the single-objective optimization problem is solved, and the evaluation function is constructed by the square sum ideal point method; then, the optimal solution of the evaluation function corresponding to the ideal point or close to the ideal point is found; finally, the optimal solution is analyzed and the weight coefficient is derived; an energy management module configured to solve the energy management model of the hybrid vehicle by using a proximal strategy optimization algorithm to implement energy management of the hybrid vehicle; The objective function of energy management includes fuel economy, stability of the power battery state of charge, and power battery temperature. The optimal weight coefficient is determined based on the square sum optimal method. That is, a single objective is set as the objective function, and the optimal torque distribution result and the corresponding first target value under the current objective function are calculated; then a new evaluation function is constructed based on the optimal target value solved by the single objective, and the corresponding optimal torque distribution result and the corresponding second target value under the constructed new evaluation function are solved; according to the weight coefficient calculation formula, the weight coefficient values corresponding to the three objective functions are calculated.

8. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the hybrid electric vehicle energy management method based on multi-objective optimization as described in any one of claims 1 to 6 are implemented.

Images