A multi-power source vehicle energy management strategy optimization method, system and device
By optimizing the energy management strategy for multi-power source vehicles using Genetic Topological Neural Network (NEAT), the problem of gradient methods getting stuck in local optima is solved, achieving stronger adaptability and real-time performance, and improving the efficiency and economy of energy management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING UNIV OF TECH
- Filing Date
- 2023-08-15
- Publication Date
- 2026-07-24
AI Technical Summary
In existing energy management strategies for multi-power source vehicles, gradient-based neural network methods are prone to getting stuck in local optima, are complex to debug and lack real-time performance, and have cumbersome neural network structure debugging, making it difficult to find the global optimal solution.
Genetic Topological Neural Network (NEAT) is used to optimize the energy management strategy for multi-power source vehicles. By adjusting the network topology and weights through a genetic algorithm and combining standard operating conditions and vehicle models, the power allocation of motors is dynamically adjusted to avoid local optima and improve adaptability.
It achieves greater adaptability and real-time performance, reduces debugging cycles, avoids data storage waste, and can better allocate energy flow to achieve optimal economy and power.
Smart Images
Figure CN117021971B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of new energy vehicle technology, and in particular to a method, system and equipment for optimizing energy management strategies for multi-power source vehicles. Background Technology
[0002] To promote a comprehensive green transformation of economic and social development and achieve carbon peaking and carbon neutrality, traditional fuel vehicles with severe pollution emissions will gradually transition to new energy vehicles. Whether it's hybrid electric vehicles, hydrogen fuel cell vehicles, or distributed drive pure electric vehicles with multiple motors, all require an Energy Management Strategy (EMS) to efficiently allocate energy / power. The performance of the EMS is of great significance for optimizing the energy efficiency of these various types of new energy vehicles.
[0003] Currently, EMS (Electrical Management Systems) for multi-powered vehicles can be broadly categorized into three types: rule-based EMS, optimization-based EMS, and learning-based EMS. Rule-based EMS requires extensive experience and knowledge, considering all possible scenarios and developing corresponding rules for each, resulting in a cumbersome design process and rigid execution. While optimization-based management strategies can find optimal solutions, they require prior knowledge of the complete operating conditions before applying optimization algorithms, making them unsuitable for unknown conditions, computationally intensive, and lacking real-time performance. Learning-based EMS offers significant advantages over the former two. Its principle involves training the agent using a large amount of existing data. After exploring the action space, the agent generates a set of internal logical rules, enabling it to execute actions based on various situations.
[0004] Most current learning-based EMS methods, such as Deep Deterministic Policy Gradient (DDPG), Twin Delayed Deep Deterministic Policy Gradient (TD3), and Distributed Proximal Policy Optimization (DPPO), update network parameters based on gradients. Gradient-based optimization methods share a common drawback—they are prone to getting trapped in local optima. This is because these methods rely on noise for exploration. Once the network parameters reach the loss extreme point with gradient updates, the gradients of all parameters become zero. If the noise level is insufficient to escape the influence of this extreme point, the global optimum cannot be found, leading to a local optimum. In addition, these methods have other limitations—the neural network connection method, the number of network layers, and the number of hidden layer neurons need to be manually tuned. Although expert experience can quickly lead to the most suitable topology, it still adds a lot of repetitive mechanical work, and it is almost impossible to manually find the optimal topology. If the neural network structure being debugged is too simple, it is easy to cause underfitting, which increases the difficulty of exploring logical rules and makes it difficult to converge. Conversely, an overly complex structure will cause waste of redundant layers and neurons, and may also lead to overfitting, thereby reducing the generalization ability of the agent. Moreover, during the training process, too many network layers may lead to gradient vanishing or gradient exploding. Summary of the Invention
[0005] The purpose of this invention is to provide a method, system, and device for optimizing energy management strategies for multi-power-source vehicles, which optimizes energy management strategies for multi-power-source vehicles through a genetic topological neural network.
[0006] To achieve the above objectives, the present invention provides the following solution:
[0007] A method for optimizing energy management strategies for multi-power source vehicles includes:
[0008] The current vehicle state is simulated based on standard operating conditions and a vehicle model; the standard operating conditions are global light vehicle test cycle operating conditions, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle state includes wheel speed, motor speed, power demand, state of charge, and battery health throughout the entire operating condition period.
[0009] Based on a genetic topological neural network, the proportion of the required power allocated to the motor is determined;
[0010] The motor torque is calculated based on the ratio, and the power consumption of the motor is calculated based on the motor torque.
[0011] The reward value of all individuals in the genetic topological neural network is calculated based on the power consumed.
[0012] Based on the reward value, the optimal individual is determined, and the optimal energy management strategy is obtained.
[0013] Optionally, based on a genetic topological neural network, the proportion of the required power allocated to the motor is determined, specifically including:
[0014] Initialize the gene population for the genetic topological neural network;
[0015] The state of charge and battery health of the vehicle battery are input into the genetic topological neural network. After initialization, all individuals in the gene population multiply the two input values by the connection weights and add the bias of the neurons to perform data transmission between neurons, resulting in different output values.
[0016] Based on the required power, the output values of all individuals in the genetic topological neural network are restricted to the range of 0 to 1 using the sigmoid function as an action to adjust the proportion of the required power allocated by the motor.
[0017] Alternatively, the formula for calculating motor torque is as follows:
[0018]
[0019]
[0020] Where T1 and T2 are the torques that allocate the required power to the first motor and the second motor, respectively; action is the action with a value ranging from 0 to 1; P is the required power; and n1 and n2 are the speeds of the first motor and the second motor, respectively.
[0021] Optionally, the formula for calculating the power consumption is as follows:
[0022]
[0023]
[0024] Where P1 and P2 are the power consumption of the first motor and the second motor, respectively; Ratio1 and Ratio2 are the transmission ratios of the reducers connected to the first motor and the second motor, respectively; V is the vehicle speed; R is the wheel radius; and E1 and E2 are the efficiencies of the first motor and the second motor obtained by interpolation in the motor efficiency diagram, respectively.
[0025] Optionally, the formula for calculating the reward value is as follows:
[0026]
[0027] Where reward is the reward value at the current moment, P is the required power, and P1 and P2 are the power consumed by the first motor and the second motor, respectively.
[0028] Optionally, the optimal individual is determined based on the reward value to obtain the optimal energy management strategy, specifically including:
[0029] The individual with the highest fitness is determined based on the reward value;
[0030] The individual with the highest fitness is selected as the optimal individual, thus obtaining the optimal energy management strategy.
[0031] The present invention also provides a multi-power source vehicle energy management strategy optimization system, comprising:
[0032] The vehicle state simulation module is used to simulate the current vehicle state based on standard operating conditions and a vehicle model. The standard operating conditions are global light vehicle test cycle operating conditions, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle state includes wheel speed, motor speed, power demand, state of charge, and battery health throughout the entire operating period.
[0033] The allocation ratio determination module is used to determine the proportion of the required power allocated to the motor based on a genetic topological neural network.
[0034] The torque and power consumption calculation module is used to calculate the motor torque based on the ratio and to calculate the motor power consumption based on the motor torque.
[0035] The reward value calculation module is used to calculate the reward value of all individuals in the genetic topological neural network based on the power consumption.
[0036] The optimal energy management strategy determination module is used to determine the optimal individual based on the reward value, thereby obtaining the optimal energy management strategy.
[0037] Optionally, the allocation ratio determination module specifically includes:
[0038] An initialization unit is used to initialize the gene population of a genetic topological neural network.
[0039] The output value determination unit is used to input the state of charge and battery health of the vehicle battery into the genetic topological neural network. After initialization, all individuals in the gene population multiply the two input values by the connection weights and add the bias of the neurons to perform data transmission between neurons and obtain different output values.
[0040] The allocation ratio determination unit is used to adjust the proportion of the motor allocating the required power based on the required power by using the sigmoid function to limit the output values of all individuals in the genetic topological neural network to the range of 0 to 1.
[0041] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described multi-power source vehicle energy management strategy optimization method.
[0042] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for optimizing energy management strategies for multi-power source vehicles.
[0043] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0044] This invention replaces the gradient-based weight-updating neural network commonly used in EMS for multi-power-source vehicles with a genetic topological neural network (NEAT) based on evolutionary updating of topology and weights. The NEAT network has a minimal optimization structure, which can avoid unnecessary data storage caused by the EMS. Moreover, the NEAT network can avoid local optima, has stronger adaptability, and can reduce the EMS debugging cycle. By determining the optimal EMS, the energy flow between vehicle power sources is allocated, generally by power allocation, so that the vehicle can achieve the best economy while meeting the power requirements. Attached Figure Description
[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0046] Figure 1 This is a flowchart of the multi-power source vehicle energy management strategy optimization method provided in Embodiment 1 of the present invention;
[0047] Figure 2 A detailed flowchart of the multi-power source vehicle energy management strategy optimization method provided in Embodiment 1 of the present invention;
[0048] Figure 3 A schematic diagram of a dual-motor fixed reduction ratio model;
[0049] Figure 4This is a schematic diagram of the NEAT network structure;
[0050] Figure 5 This is a diagram showing the operating points of a vehicle motor under a specific working condition. Detailed Implementation
[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0052] The purpose of this invention is to provide a method, system, and device for optimizing energy management strategies in multi-power source vehicles. The neural network based on gradient-updated weights, which is currently widely used in the EMS of multi-power source vehicles, can be replaced with a neural network based on evolution-updated topology and weights to optimize the energy management strategy of multi-power source vehicles.
[0053] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0054] Example 1
[0055] like Figures 1-2 As shown, a method for optimizing energy management strategies for multi-power source vehicles includes the following steps:
[0056] S1: Simulate the current vehicle state based on standard operating conditions and vehicle model; the standard operating conditions are global light vehicle test cycle operating conditions, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle state includes wheel speed, motor speed, required power, state of charge and battery health throughout the entire operating period.
[0057] This invention uses the World Light Vehicle Test Cycle (WLTC) as the standard operating condition. The vehicle model is a dual-motor fixed reduction ratio model, such as... Figure 3 As shown, this model has a simpler structure compared to other multi-power-source pure electric vehicle models, making it well-suited for testing the performance of various algorithms. Based on the WLTC operating conditions, it calculates information such as wheel speed, motor speed, power demand, stage of charge (SOC), and battery health (SOH) throughout the entire operating period. The calculation formula is as follows:
[0058]
[0059] n1 = nw ·Ratio1 (2)
[0060] n2=n w ·Ratio2 (3)
[0061]
[0062]
[0063]
[0064] Where, n w V is the wheel rotational speed, R is the vehicle speed, and R is the wheel radius; n1 and n2 are the rotational speeds of the first and second motors, respectively; Ratio1 and Ratio2 are the transmission ratios of the reducers connected to the first and second motors, respectively; P is the required power, F f For rolling resistance, F w For air resistance, F j For acceleration resistance, V is the vehicle speed, G is the total weight of the vehicle, f is the rolling resistance coefficient, and C is the rolling resistance coefficient. D Let be the air resistance coefficient, A be the vehicle's frontal area, δ be the vehicle's rotational mass coefficient, and g be the acceleration due to gravity. It is the derivative of velocity with respect to time. This formula assumes the road is flat, so gradient resistance is not calculated; SOC(t0) is the initial SOC, C rated Where τ is the vehicle battery's rated capacity, I is the vehicle's driving time. b I is the battery current. loss The current consumed by the loss reaction; Q rated Q is the rated capacity of the battery. discharge This represents the actual capacity of the battery.
[0065] S2: Based on the genetic topological neural network (NEAT), determine the proportion of the required power allocated to the motor.
[0066] The role of the NEAT network is to train the most suitable agent as the decision-maker for the EMS (Electronic Management System) based on external operating conditions and vehicle models. It starts with a unified network population without hidden nodes; that is, its initial population is a simple structure with only input and output neurons, such as... Figure 4As shown. Each individual in a neural network can be expressed through a genome, which contains a list of node genes and a list of connection genes. Node genes contain all nodes in the neural network and their corresponding attributes, while connection genes contain the nodes of two connections, the weight of the connections, whether the connection genes are expressed, and innovative encodings. The first step of NEAT reinforcement learning in this invention is to perform cross-pairing of the neural network population according to the list of connection genes, such as... Figure 4 As shown, the connection genes are sorted according to their innovation codes, distinguishing between matching genes (e.g., 2, 4, and 7), disconnected genes (e.g., 1, 3, and 6), and redundant genes (e.g., 8 and 9). Disconnected and redundant genes are directly inherited by the next generation, while matching genes are randomly inherited from both sides. The second step generates variations in connection weights and network structure. These variations are categorized into two types: adding connections and adding nodes. Figure 3 As shown. In the connection mutation, a new connection gene is added to connect two previously unconnected nodes. In the node mutation, an existing connection is split, the new node is placed in the position where the old connection was, the old connection is not activated, and two new connections are added to the genome. The third step uses the previously calculated power demand and the neural network population after the first two steps. First, the State of Charge (SOC) and State of Health (SOH) of the vehicle battery are used as observation variables to assign values to the two input neurons of all individuals. Then, all neural network individuals in the population multiply the two input values by the connection weights and add the neuron bias to perform data transmission between neurons. Since each neural network individual has its own unique topology and weight parameters, it will have different network outputs. Then, the sigmoid function is used to restrict the output values of all neural networks to the range of 0 to 1 as an action to adjust the ratio of the power demand allocated between the two motors, and then the motor torque is calculated based on the action.
[0067]
[0068]
[0069] In the formula, T1 and T2 are the torques of motor 1 and motor 2, respectively, action is the action with a value range of 0 to 1, i.e. the power distribution ratio, P is the required power calculated in the previous formula (4), and n1 and n2 are the speeds of motor 1 and motor 2, respectively.
[0070] S3: Calculate the motor torque based on the ratio, and calculate the power consumption of the motor based on the motor torque.
[0071] If the power allocated to the motor exceeds its maximum output, termination will be triggered, and a large negative reward will be given to incentivize the agent to avoid this choice. Different power allocations between the two motors will cause their operating points to fall within different operating ranges, resulting in varying drive efficiencies that affect the vehicle's fuel economy. Figure 5 This is a diagram showing the operating points of the vehicle motor under a given operating condition. The horizontal and vertical axes represent speed and torque, respectively. Based on the motor's speed and torque, the specific operating point can be determined, and the efficiency of the current operating point can be interpolated. The darker the area, the higher the motor efficiency. Then, the power consumed by the two motors under the action control of the intelligent agent can be calculated using formulas (9) and (10).
[0072]
[0073]
[0074] In the formula, P1 and P2 are the power consumed by the two motors, Ratio1 and Ratio2 are the transmission ratios of the reducers connected to motor 1 and motor 2, V is the vehicle speed, R is the wheel radius, and E1 and E2 are the efficiencies of motor 1 and motor 2 obtained by interpolation in the motor efficiency diagram.
[0075] S4: Calculate the reward value of all individuals in the genetic topological neural network based on the power consumption.
[0076] The closer the work points are distributed to the high-efficiency zone, the less energy the vehicle consumes, and the greater the reward the agent receives. The specific calculation is as follows:
[0077]
[0078] In the formula, reward is the reward at the current moment, P is the required power, and -100 will be used as the reward if termination is triggered.
[0079] S5: Determine the optimal individual based on the reward value to obtain the optimal energy management strategy.
[0080] The vehicle battery's State of Charge (SOC) and State of Hypothesis (SOH) change based on the current power consumption of the two motors, and these changes serve as input for the next iteration of the neural network. Higher rewards result in higher calculated fitness. To maintain a constant population size, individuals in the neural network are ranked and eliminated based on their fitness before entering the next iteration cycle.
[0081] It is worth noting that new individuals generated through crossover and mutation typically have low fitness. To protect topological innovation and individual diversity, the population was speciesized based on topological similarity. The number of redundant and disjoint genes between a pair of genomes is a measure of their compatibility. The compatibility distance is calculated as shown in Equation (12):
[0082]
[0083] In the formula, δ represents the compatibility distance, E represents the number of excess genes, and D represents the number of disconnected genes. To match the average weight differences of genes, coefficients c1, c2, and c3 are used to adjust the importance of these three factors, where N is the number of genes in the larger-connected genome. Species classification requires setting a compatibility threshold δ. t If the calculated distance δ is greater than the compatibility threshold δ t If the number of genomes is small, then the two genomes are considered as the same species. Species expansion or contraction depends on whether their average fitness is above or below the population average. Variations in species numbers are shown in equation (13):
[0084]
[0085] In the formula N j and N' j f represents the number of individuals of species j before and after the change. ij The fitness of individual i in species j. Let be the average fitness of the total population. As can be seen from equation (13), as long as there are individuals with high fitness in a species, they can protect newly generated individuals with low fitness from being eliminated under their influence, thus protecting topological innovation and individual diversity. The ideal situation is that a species dominates the population, thereby stopping the iteration, but in practice, the number of iterations can be set to terminate the algorithm.
[0086] like Figure 2 As shown, the actual driving data of the vehicle is finally used to verify the performance of the neural network individual with the highest fitness. If the requirements are met, the optimal multi-power source vehicle EMS is obtained; otherwise, the initial population is reset and iterated again.
[0087] This invention replaces the gradient-based weight-updating neural network commonly used in EMS of multi-power source vehicles with an evolution-based neural network that updates the topology and weights, achieving the following effects:
[0088] 1. Minimal optimization structure avoids the waste of extra layers caused by using complex networks to solve relatively simple problems, and avoids unnecessary data storage caused by EMS.
[0089] 2. Avoid local optima: Gradient-based neural networks can only explore through noise, while evolution-based neural network populations rely on crossover and mutation for large-scale search, which has a stronger exploration capability and makes it easier for EMS to find the optimal way to allocate energy flow.
[0090] 3. It is more adaptable because the structure can be changed, so it can generate the most suitable structure for specific problems. It does not require manually adjusting some hyperparameters (such as the number of neural network layers and the number of neurons in the hidden layer) in order to try to find the most suitable structure for the current problem, thus reducing the EMS debugging cycle.
[0091] Example 2
[0092] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, a multi-power source vehicle energy management strategy optimization system is provided below.
[0093] The system includes:
[0094] The vehicle state simulation module is used to simulate the current vehicle state based on standard operating conditions and a vehicle model. The standard operating conditions are global light vehicle test cycle operating conditions, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle state includes wheel speed, motor speed, required power, state of charge, and battery health throughout the entire operating period.
[0095] The allocation ratio determination module is used to determine the proportion of the required power allocated to the motor based on a genetic topological neural network.
[0096] The torque and power consumption calculation module is used to calculate the motor torque based on the ratio and to calculate the motor power consumption based on the motor torque.
[0097] The reward value calculation module is used to calculate the reward value of all individuals in the genetic topological neural network based on the power consumption.
[0098] The optimal energy management strategy determination module is used to determine the optimal individual based on the reward value, thereby obtaining the optimal energy management strategy.
[0099] Furthermore, the allocation ratio determination module specifically includes:
[0100] An initialization unit is used to initialize the gene population of a genetic topological neural network.
[0101] The output value determination unit is used to input the state of charge and battery health of the vehicle battery into the genetic topological neural network. After initialization, all individuals in the gene population multiply the two input values by the connection weights and add the bias of the neurons to perform data transmission between neurons and obtain different output values.
[0102] The allocation ratio determination unit is used to adjust the proportion of the motor allocating the required power based on the required power by using the sigmoid function to limit the output values of all individuals in the genetic topological neural network to the range of 0 to 1.
[0103] Example 3
[0104] Embodiment 3 of the present invention provides an electronic device, including a memory and a processor. The memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the multi-power source vehicle energy management strategy optimization method provided in Embodiment 1.
[0105] In practical applications, the aforementioned electronic devices can be servers.
[0106] In practical applications, electronic devices include: at least one processor, memory, bus, and communication interface.
[0107] The processor, communication interface, and memory communicate with each other via a communication bus.
[0108] A communication interface is used to communicate with other devices.
[0109] The processor is used to execute programs, specifically the methods described in the above embodiments.
[0110] Specifically, the program may include program code, which includes computer operation instructions.
[0111] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The electronic device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.
[0112] Memory is used to store programs. Memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.
[0113] Example 4
[0114] Based on the description of Embodiment 3, Embodiment 4 of the present invention provides a storage medium storing a computer program thereon, which can be executed by a processor to implement the multi-power source vehicle energy management strategy optimization method of Embodiment 1.
[0115] The multi-power-source vehicle energy management strategy optimization system provided in Embodiment 2 of this invention exists in various forms, including but not limited to:
[0116] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include: smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones, etc.
[0117] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access capabilities. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.
[0118] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes: audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.
[0119] (4) Other electronic devices with data interaction functions.
[0120] Specific embodiments of the subject matter have now been described. Other embodiments are within the scope of the appended claims. In some cases, the actions described in the claims can be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing can be advantageous.
[0121] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.
[0122] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, in implementing this invention, the functions of each unit can be implemented in one or more software and / or hardware components. Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0123] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0125] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0126] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0127] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0128] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined in this invention, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0129] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0130] This invention can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules.
[0131] Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific transactions or implement specific abstract data types. This invention can also be practiced in distributed computing environments where transactions are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.
[0132] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.
[0133] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A method for optimizing energy management strategies for multi-power source vehicles, characterized in that, include: Simulate the current vehicle state based on standard operating conditions and a vehicle model; The standard operating condition is the global light vehicle test cycle, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle status includes wheel speed, motor speed, power demand, state of charge and battery health throughout the entire operating period. Based on a genetic topological neural network, the proportion of the required power allocated to the motor is determined. Specifically, this includes: initializing the gene population of the genetic topological neural network; inputting the state of charge and battery health of the vehicle battery into the genetic topological neural network; multiplying the two input values by the connection weights and adding the neuron biases to perform data transmission between neurons in all individuals of the initialized gene population, resulting in different output values; and using the sigmoid function to limit the output values of all individuals in the genetic topological neural network to the range of 0 to 1 as an action to adjust the proportion of the required power allocated to the motor. The motor torque is calculated based on the ratio, and the power consumption of the motor is calculated based on the motor torque. The reward value of all individuals in the genetic topological neural network is calculated based on the power consumed. The optimal individual is determined based on the reward value, and the optimal energy management strategy is obtained; specifically, this includes: determining the individual with the highest fitness based on the reward value; and taking the individual with the highest fitness as the optimal individual to obtain the optimal energy management strategy.
2. The method for optimizing energy management strategies for multi-power source vehicles according to claim 1, characterized in that, The formula for calculating the motor torque is as follows: in, , These are the torques of the first motor and the second motor, respectively. It is an action with a value ranging from 0 to 1, where P is the power demand. , These are the rotational speeds of the first motor and the second motor, respectively.
3. The method for optimizing energy management strategies for multi-power source vehicles according to claim 1, characterized in that, The formula for calculating the power consumption is as follows: in, , These are the power consumption of the first motor and the second motor, respectively. and These are the transmission ratios of the reducers connected to the first motor and the second motor, respectively. For vehicle speed, For the wheel radius, , These are the efficiencies of the first and second motors obtained by interpolation in the motor efficiency graph, respectively.
4. The method for optimizing energy management strategies for multi-power source vehicles according to claim 1, characterized in that, The formula for calculating the reward value is as follows: in, This is the reward value at the current moment. For the required power, , These represent the power consumption of the first motor and the second motor, respectively.
5. A multi-power source vehicle energy management strategy optimization system, characterized in that, include: The vehicle state simulation module is used to simulate the current vehicle state based on standard operating conditions and a vehicle model. The standard operating condition is the global light vehicle test cycle, the vehicle model is a dual-motor fixed reduction ratio model, and the vehicle status includes wheel speed, motor speed, power demand, state of charge and battery health throughout the entire operating period. The allocation ratio determination module is used to determine the proportion of the required power allocated to the motor based on a genetic topological neural network. The torque and power consumption calculation module is used to calculate the motor torque based on the ratio and to calculate the motor power consumption based on the motor torque. The reward value calculation module is used to calculate the reward value of all individuals in the genetic topological neural network based on the power consumption. The optimal energy management strategy determination module is used to determine the optimal individual based on the reward value to obtain the optimal energy management strategy; specifically, it includes: determining the individual with the highest fitness based on the reward value; and taking the individual with the highest fitness as the optimal individual to obtain the optimal energy management strategy. The allocation ratio determination module specifically includes: An initialization unit is used to initialize the gene population of a genetic topological neural network. The output value determination unit is used to input the state of charge and battery health of the vehicle battery into the genetic topological neural network. After initialization, all individuals in the gene population multiply the two input values by the connection weights and add the bias of the neurons to perform data transmission between neurons and obtain different output values. The allocation ratio determination unit is used to adjust the proportion of the motor allocating the required power based on the required power by using the sigmoid function to limit the output values of all individuals in the genetic topological neural network to the range of 0 to 1.
6. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform the multi-power source vehicle energy management strategy optimization method according to any one of claims 1-4.
7. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the multi-power source vehicle energy management strategy optimization method as described in any one of claims 1-4.