Attenuation self-adaptive optimal energy management system of intelligent fuel cell vehicle

Abstract

The invention relates to an attenuation self-adaptive intelligent fuel cell automobile optimal energy management system, an energy management strategy of the system is realized by adopting a multi-core controller, and the multi-core controller comprises a CPU0 (central processing unit 0) for estimating the SOH (state of health) of a fuel cell in real time and obtaining a fuel cell efficiency curve under the current SOH to update a fuel cell automobile model; the CPU1 is used for performing prediction based on machine learning to predict and obtain an optimal co-state variable; and the CPU 2 performs a real-time optimal energy management strategy. The invention ensures that the energy management strategy can be adjusted along with the change of the state of the fuel cell, ensures that the strategy is always consistent with the actual state of the fuel cell, improves the accuracy and adaptability of energy management, and is beneficial to prolonging the service life of the fuel cell and improving the overall performance of the vehicle. According to the invention, the execution efficiency of the energy management strategy is optimized, the real-time performance of the system is improved, and the optimal energy management of the fuel cell vehicle under the actual driving condition is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent fuel cells, and in particular to an optimal energy management system for intelligent fuel cell vehicles with adaptive attenuation. Background Art

[0002] The transportation industry is one of the main sources of energy consumption and environmental pollution. Developing new energy vehicles is an effective way to reduce pollution and energy consumption. Fuel cell vehicles, as a clean and efficient new energy vehicle, have high energy density and zero emissions, and are considered to be an important development direction for the future automotive industry.

[0003] Energy management strategies applied to fuel cell vehicles can be divided into two categories: rule-based and optimization-based. Rule-based energy management strategy is a method that relies on pre-set rules and logic to manage the energy distribution of the fuel cell system. Optimization-based energy management strategy is a method that uses mathematical models and optimization algorithms to determine the best solution for energy distribution in order to optimize system performance. Among them, it can be divided into global optimization and instantaneous optimization. Commonly used global optimization methods include dynamic programming, PMP, etc., which require all operating conditions to obtain the global optimal solution. Therefore, global optimization methods cannot be applied in real time, but are used as a standard for evaluating other energy management strategies. Instantaneous optimization is to allocate the required power in real time based on vehicle operating parameters and operating data. Commonly used instantaneous optimization methods include equivalent hydrogen consumption minimum strategy. The core idea of ​​the equivalent hydrogen consumption minimum strategy is to convert the energy of power batteries and fuel cells into equivalent hydrogen consumption and minimize the equivalent hydrogen consumption.

[0004] Fuel cells are important components of fuel cell vehicles. Due to complex operating conditions, fuel cells deteriorate significantly during their life, and the performance of fuel cells changes as the fuel cell deteriorates, thus affecting the economic efficiency of energy management strategies. Therefore, energy management control strategies that consider fuel cell attenuation adaptation are of great research significance.

[0005] At present, most of the research on fuel cell energy management strategies in the published patent literature is based on the assumption that fuel cells do not decay. It does not fully consider the impact of the decay of the fuel cell health state on the performance of the energy management strategy, nor does it study how to adjust the energy management strategy after the system decays. However, in actual applications, the efficiency of the fuel cell and the system performance change with the decrease of SOH. If the energy management strategy cannot adapt to the decayed fuel cell system, it will cause the energy management strategy to perform poorly in actual applications after the fuel cell decays, fail to achieve global optimality, reduce adaptability, and even aggravate system aging. Therefore, it is of great practical significance to conduct in-depth research on the impact of fuel cell SOH decay on the energy management strategy and develop an energy management method that can dynamically adjust to adapt to the actual SOH of the fuel cell. This will not only help improve the overall performance and energy efficiency of fuel cell vehicles, but also extend the service life of the fuel cell system.

[0006] In addition, when the existing technology combines traffic information for adaptive energy management, there are problems such as insufficient real-time data processing capabilities, low prediction accuracy, lack of dynamic adaptive mechanism, unreasonable allocation of computing resources, and insufficient use of cloud computing. As a result, the energy management strategy fails to fully consider the real-time changes of traffic information, affecting the overall performance and optimization effect of the system.

[0007] At the same time, existing technologies generally fail to perform effective hierarchical scheduling of tasks of different time scales. For example, tasks such as fuel cell SOH estimation, ELM training prediction optimal co-state factor calculation, and online energy management strategy execution have different time scales and priorities. It is difficult for existing technologies to coordinate the execution of these tasks, resulting in high-priority real-time tasks being affected by low-priority tasks, and the overall performance of the system being limited. When processing tasks of different time scales, existing technologies fail to reasonably allocate computing resources, resulting in low efficiency in computing resource utilization. For example, the fuel cell SOH estimation task may only need to be updated once a month, but in a single-core processor, these low-frequency tasks may occupy a large amount of computing resources, affecting the execution efficiency of high-frequency tasks. In addition, the high-computational-load ELM training tasks compete with the real-time energy management strategy for the same computing resources, making it difficult to ensure the priority execution of real-time tasks. Summary of the invention

[0008] In view of the above-mentioned problems, the present invention proposes an optimal energy management system for intelligent fuel cell vehicles with attenuation self-adaptation.

[0009] Terminology explanation: SOH: SOH (state of health of the fuel cell), the health status of the fuel cell, is an important indicator for evaluating battery life and performance. SOH reflects the current capacity and performance of the battery relative to the initial state of a new battery. When SOH is 100%, it means that the battery is in a brand new state, while a lower percentage means that the capacity and performance of the battery have decreased.

[0010] PMP: Pontryagin's Minimum Principle (PMP) is an important method for solving optimal control problems. The principle was proposed by Soviet mathematician Pontryagin and his colleagues in the 1950s. PMP provides a necessary condition to help determine how a control system can achieve the optimal performance indicator through the selection of control variables under given initial conditions.

[0011] ECMS: Equivalent Consumption Minimization Strategy (ECMS) is a real-time control strategy for energy management in hybrid electric vehicles (HEV) and fuel cell hybrid electric vehicles (FCHV). The main goal of this strategy is to minimize the equivalent hydrogen consumption of the vehicle by properly distributing energy between the fuel cell and the battery, thereby improving the energy efficiency of the entire system.

[0012] ELM algorithm: Extreme Learning Machine algorithm is a fast training method for single hidden layer feedforward neural networks. It achieves efficient learning and good generalization ability by randomly generating input weights and using the least squares method to analytically solve the output weights.

[0013] To achieve the above object, the present invention provides the following technical solutions: An optimal energy management system for intelligent fuel cell vehicles with adaptive decay, wherein the energy management strategy of the system is implemented by a multi-core controller, and the multi-core controller comprises: CPU0 estimates the fuel cell SOH in real time and obtains the fuel cell efficiency curve under the current SOH; CPU1, which performs prediction based on machine learning to predict the optimal co-state variables; CPU2, which performs real-time optimal energy management strategy.

[0014] As a further technical solution of the present invention, the CPU0 estimates the SOH of the fuel cell in real time, and obtains the content of the fuel cell efficiency curve under the current SOH, including: Through the real-time SOH estimation algorithm, the system can accurately estimate the SOH of the fuel cell during actual operation. Since the electrochemical characteristics of the fuel cell will change with the change of SOH, the efficiency map under different SOH states will also change accordingly; when the fuel cell SOH is high, its efficiency is high, and as the fuel cell SOH decreases, the efficiency of the fuel cell gradually decreases. This efficiency change directly affects the energy conversion efficiency of the fuel cell and the overall energy efficiency of the vehicle. In order to ensure that the energy management strategy can reflect the current actual state of the fuel cell, the fuel cell vehicle model must be dynamically updated according to the real-time SOH estimation results. The fuel cell efficiency map under different SOHs is obtained through the SOH identification algorithm and applied to the update of the fuel cell vehicle model.

[0015] As a further technical solution of the present invention, the CPU 1 performs prediction based on machine learning, and the content of predicting the optimal co-state variable includes: Obtain historical traffic data and driving data information of fuel cell hybrid vehicles to conduct power modeling; Based on the power system structure of the fuel cell hybrid vehicle, a fuel cell hybrid vehicle power system model is established; the fuel cell hybrid vehicle power system model includes: a vehicle dynamics model, a fuel cell efficiency model and a power battery model; The vehicle dynamics model is: ; In the formula is the vehicle mass, is the vehicle speed, is the required power, is the air resistance, is the rolling resistance, is the slope resistance; ; In the formula is the fuel cell stack power, is the power of the power battery; The fuel cell efficiency model is:

[0016] In the formula is the fuel cell system efficiency, is the fuel cell output power, The low calorific value of hydrogen is 120MJ / kg. is the hydrogen consumption rate; The power battery model is: ; In the formula is the open circuit voltage of the power battery, is the internal resistance of the power battery, is the rate of change of the state of charge of the power battery, The maximum charging capacity of the power battery; With the goal of minimizing hydrogen consumption, the objective function is established and the constraints are determined: ; ; Based on the Pontryagin minimum principle, the Hamiltonian function is established from the objective function; ; In the formula is the Hamiltonian function, is a co-state variable; The co-state variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ; The control variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ; The historical operating data is solved by the shooting method or Newton method to obtain the optimal co-state variables and the corresponding vehicle state information; these numerical methods can effectively solve the Hamiltonian function constructed based on the Pontryagin minimum principle, thereby determining the optimal control strategy of the system. In this process, the information obtained includes not only the optimal co-state variables, but also key data (vehicle state information) such as the vehicle's speed, driving distance, SOC trajectory and required power; Predict the optimal co-state variables at a certain moment in the future based on machine learning algorithms.

[0017] As a further technical solution of the present invention, the content of the real-time optimal energy management strategy performed by the CPU2 includes: The strategy calculates the adaptive optimal equivalent factor in real time based on the predicted optimal co-state variables and the relationship between the Pontryagin minimum principle and the equivalent consumption minimum strategy; The relationship between the co-state variables and the equivalent factors is: ; In the formula is the adaptive equivalence factor; Based on ECMS, the objective function is constructed with the goal of minimizing the total equivalent hydrogen consumption in real time: ; In the formula is the equivalent hydrogen consumption; By inputting the optimal equivalent factor obtained from the solution into the ECMS model, the optimal power distribution of the fuel cell and the power battery can be calculated. This method ensures the real-time and optimality of the energy management strategy, and effectively improves the energy efficiency of the fuel cell system and the operating performance of the vehicle.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes an optimal energy management system for intelligent fuel cell vehicles with attenuation self-adaptation. The strategy uses the SOH identification algorithm to monitor and identify the current SOH of the fuel cell in real time, and updates the fuel cell vehicle model in the energy management strategy based on the real-time identified SOH data. This dynamic update mechanism ensures that the energy management strategy can be adjusted as the fuel cell state changes, ensuring that the energy management strategy is always consistent with the actual state of the fuel cell. This adaptive update mechanism improves the accuracy, adaptability and efficiency of energy management, helps to extend the service life of the fuel cell and improve the overall performance of the vehicle.

[0019] In addition, the strategy includes a machine learning prediction algorithm combined with traffic information. Through in-depth analysis and prediction of historical operating data, the system can accurately obtain the optimal co-state variables. These data include key data such as vehicle speed, driving distance, SOC trajectory and required power, and are also trained through machine learning algorithms (such as extreme learning machine algorithms) to obtain optimal co-state variables based on machine learning. Combining the relationship between the Pontryagin minimum principle and the equivalent consumption minimum strategy, the adaptive optimal equivalent factor is calculated in real time, thereby realizing the real-time optimal energy management strategy and ensuring the optimal power distribution of fuel cells and power batteries.

[0020] In addition, in order to achieve efficient energy management, the present invention adopts the technology of multi-core controller parallel computing to distribute different computing tasks to multiple CPUs to run simultaneously. Through this multi-core parallel processing method, the system's operating time is significantly reduced, the computing efficiency and response speed are improved, and it can adapt to dynamically changing working conditions more quickly. This method not only optimizes the execution efficiency of the energy management strategy, but also improves the real-time performance of the overall system, ensuring that fuel cell vehicles can achieve optimal energy management under various working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flowchart of a multi-core controller provided in an embodiment of the present invention.

[0022] Figure 2 This is an efficiency curve diagram under different SOH provided by an embodiment of the present invention.

[0023] Figure 3 A control block diagram of fuel cell degradation adaptive energy management provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The technical solution of this patent is further described in detail below in conjunction with specific implementation methods.

[0025] The present invention aims to propose an optimal energy management system for intelligent fuel cell vehicles with adaptive decay. The energy management strategy is implemented using a multi-core controller, and its workflow is shown in the following figure. Figure 1 As shown. In this system, the first core CPU0 of the multi-core processor estimates the SOH of the fuel cell in real time and obtains the fuel cell efficiency curve under the current SOH state to update the fuel cell vehicle model in CPU1 and CPU2. The second core CPU1 of the multi-core processor performs a machine learning prediction algorithm combined with traffic information. By solving the information of the fuel cell hybrid vehicle in the driving process of the historical working conditions, the optimal co-state variable is predicted. The third core CPU2 of the multi-core processor performs a real-time energy management strategy, that is, the equivalent fuel consumption minimum strategy, and obtains the optimal allocated power by solving the equivalent factor. Since the decay rate of the fuel cell is relatively slow, its health state (SOH) usually decreases after a long time. Therefore, in the case of differences in the time domain of the multi-core CPU, the fuel cell state update may only need to be performed once every few months. This design ensures that the system can maintain efficient energy management performance in long-term operation.

[0026] The specific steps include: The first core CPU0 of the multi-core processor estimates the fuel cell SOH in real time and obtains the fuel cell efficiency curve under the current SOH, including: Through the real-time SOH estimation algorithm, the system can accurately estimate the SOH of the fuel cell during actual operation. Since the electrochemical characteristics of the fuel cell will change with the change of SOH, the efficiency map under different SOH states will also change accordingly; when the fuel cell SOH is high, its efficiency is high, and as the fuel cell SOH decreases, the efficiency of the fuel cell gradually decreases. This efficiency change directly affects the energy conversion efficiency of the fuel cell and the overall energy efficiency of the vehicle. In order to ensure that the energy management strategy can reflect the current actual state of the fuel cell, the fuel cell vehicle model must be dynamically updated according to the real-time SOH estimation results. Figure 2 As shown, the fuel cell efficiency map under different SOHs is obtained through the SOH identification algorithm and applied to the update of the fuel cell vehicle model.

[0027] The second core CPU1 of the multi-core processor performs prediction based on machine learning, and the contents of the prediction to obtain the optimal co-state variables include: Obtain historical traffic data and driving data information of fuel cell hybrid vehicles to conduct power modeling; Based on the power system structure of the fuel cell hybrid vehicle, a fuel cell hybrid vehicle power system model is established; the fuel cell hybrid vehicle power system model includes: a vehicle dynamics model, a fuel cell efficiency model and a power battery model; The vehicle dynamics model is: ; In the formula is the vehicle mass, is the vehicle speed, is the required power, is the air resistance, is the rolling resistance, is the slope resistance; ; In the formula is the fuel cell stack power, is the power of the power battery; The fuel cell efficiency model is:

[0028] In the formula is the fuel cell system efficiency, is the fuel cell output power, The low calorific value of hydrogen is 120MJ / kg. is the hydrogen consumption rate; The power battery model is: ; In the formula is the open circuit voltage of the power battery, is the internal resistance of the power battery, is the rate of change of the state of charge of the power battery, The maximum charging capacity of the power battery; With the goal of minimizing hydrogen consumption, the objective function is established and the constraints are determined: ; ; Based on the Pontryagin minimum principle, the Hamiltonian function is established from the objective function; ; In the formula is the Hamiltonian function, is a co-state variable; The co-state variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ; The control variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ; The historical operating data is solved by the shooting method or Newton method to obtain the optimal co-state variables and the corresponding vehicle state information; these numerical methods can effectively solve the Hamiltonian function constructed based on the Pontryagin minimum principle, thereby determining the optimal control strategy of the system. In this process, the information obtained includes not only the optimal co-state variables, but also key data (vehicle state information) such as the vehicle's speed, driving distance, SOC trajectory, and required power.

[0029] Furthermore, these optimization and prediction results are input into machine learning algorithms, such as extreme learning machine algorithms, for training and obtaining optimal co-state variables based on machine learning. The obtained vehicle speed, driving distance, SOC trajectory, and required power are used as input factors of the machine learning algorithm. The entire model processes information through the input layer, feature mapping layer, and output layer to predict the optimal co-state variables at a certain moment in the future. In the specific implementation process, the input layer is responsible for receiving and preprocessing various data from the vehicle, including the vehicle's speed, SOC trajectory, current and historical power demand, and cumulative driving distance. The feature mapping layer maps these input data into a high-dimensional feature space through a series of complex nonlinear transformations to capture the potential complex relationships between the data. Finally, the output layer predicts based on the high-dimensional features generated by the feature mapping layer and generates the optimal co-state variables at a certain moment in the future to guide the energy management of the fuel cell system.

[0030] The third core CPU2 of the multi-core processor performs the real-time optimal energy management strategy including: The strategy calculates the adaptive optimal equivalent factor in real time based on the predicted optimal co-state variables and the relationship between the Pontryagin minimum principle and the equivalent consumption minimum strategy; The relationship between the co-state variables and the equivalent factors is: ; In the formula is the adaptive equivalence factor; Based on ECMS, the objective function is constructed with the goal of minimizing the total equivalent hydrogen consumption in real time: ; In the formula is the equivalent hydrogen consumption; By inputting the optimal equivalent factor obtained from the solution into the ECMS model, the optimal power distribution of the fuel cell and the power battery can be calculated. This method ensures the real-time and optimality of the energy management strategy, and effectively improves the energy efficiency of the fuel cell system and the operating performance of the vehicle.

[0031] In summary, the present invention proposes an optimal energy management system for intelligent fuel cell vehicles with adaptive decay. Figure 3 As shown in the figure, the efficiency curve is obtained by using the fuel cell degradation model in the fuel cell degradation adaptive energy management strategy. Then, the efficiency curve under the current SOH is selected, and the fuel cell vehicle model is updated using the curve. The optimal co-state variable is obtained by combining the prediction of traffic information and the prediction of the machine learning algorithm. Then, according to the relationship between the PMP co-state variable and the ECMS equivalent factor, the predicted optimal co-state variable is used to calculate the optimal equivalent factor of real-time change, thereby realizing the real-time optimal energy management strategy and solving the optimal power distribution of the fuel cell vehicle.

[0032] As another preferred embodiment of the present invention, the second core (CPU1) of the multi-core processor is responsible for combining the machine learning prediction algorithm with traffic information to obtain the optimal co-state variable. In view of the large workload of this part, in order to improve computing efficiency and resource utilization, this computing task can be migrated to the cloud for execution to achieve vehicle-cloud collaborative computing. This method can not only effectively share the computing load of the on-board processor, but also utilize the powerful computing power and storage resources of the cloud to improve the accuracy and real-time performance of the prediction results.

[0033] Specifically, by uploading historical operating condition data to the cloud. After estimating the fuel cell SOH in real time, the on-board system uploads it to the cloud, updates the fuel cell vehicle model on the cloud, and performs complex machine learning model training and prediction tasks. The optimal co-state variable prediction results completed in the cloud are then transmitted back to the on-board system through the network, allowing the vehicle to apply these optimization results in real time during actual driving. Due to the slow attenuation rate of fuel cells, their SOH usually decreases after a long time. Therefore, the update of the fuel cell vehicle model may only need to be performed every few months. Through regular retraining, the model can maintain its predictive ability and accuracy, ensuring that it can still provide valid results in a changing environment.

[0034] The advantage of cloud computing lies in its powerful processing power and scalability. By using cloud resources, large-scale data can be processed in parallel to quickly complete the model training and optimization process. By adopting advanced machine learning algorithms, such as the extreme learning machine algorithm, the vehicle's driving state can be predicted more accurately, thereby optimizing the energy management strategy of the fuel cell. Through vehicle-cloud collaborative computing, not only can real-time and efficient energy management be achieved, but also the model can be continuously optimized and iterated to improve the intelligence level of the system.

[0035] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0036] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A decay-adaptive intelligent fuel cell vehicle optimal energy management system, characterized in that: The energy management strategy of the system is implemented by a multi-core controller, and the multi-core controller includes: CPU0 estimates the fuel cell SOH in real time and obtains the fuel cell efficiency curve under the current SOH; CPU1, which performs prediction based on machine learning to predict the optimal co-state variables; CPU2, which performs real-time optimal energy management strategy.

2. The attenuation adaptive intelligent fuel cell vehicle optimal energy management system according to claim 1 is characterized in that: The CPU0 estimates the fuel cell SOH in real time and obtains the fuel cell efficiency curve under the current SOH, including: Through the real-time SOH estimation algorithm, the system accurately estimates the SOH of the fuel cell during actual operation. In order to ensure that the energy management strategy reflects the current actual state of the fuel cell, the fuel cell vehicle model is dynamically updated according to the real-time SOH estimation results. The fuel cell efficiency map under different SOHs is obtained through the SOH identification algorithm and applied to the update of the fuel cell vehicle model.

3. The attenuation adaptive intelligent fuel cell vehicle optimal energy management system according to claim 1 is characterized in that: The CPU 1 performs prediction based on machine learning, and the content of the prediction to obtain the optimal co-state variable includes: Obtain historical traffic data and driving data information of fuel cell hybrid vehicles to conduct power modeling; Based on the power system structure of the fuel cell hybrid vehicle, a fuel cell hybrid vehicle power system model is established; the fuel cell hybrid vehicle power system model includes: a vehicle dynamics model, a fuel cell efficiency model and a power battery model; The vehicle dynamics model is: ； In the formula is the vehicle mass, is the vehicle speed, is the required power, is the air resistance, is the rolling resistance, is the slope resistance; ； In the formula is the fuel cell stack power, is the power of the power battery; The fuel cell efficiency model is: ； In the formula is the fuel cell system efficiency, is the fuel cell output power, The low calorific value of hydrogen is 120MJ / kg. is the hydrogen consumption rate; The power battery model is: ； In the formula is the open circuit voltage of the power battery, is the internal resistance of the power battery, is the rate of change of the state of charge of the power battery, The maximum charging capacity of the power battery; With the goal of minimizing hydrogen consumption, the objective function is established and the constraints are determined: ； ； Based on the Pontryagin minimum principle, the Hamiltonian function is established from the objective function; ； In the formula is the Hamiltonian function, is a co-state variable; The co-state variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ； The control variable is The dynamic equations of Find the partial differential and the calculation formula is as follows: ； Solve the historical operating data by shooting method or Newton method to obtain the optimal co-state variables and corresponding vehicle state information; Predict the optimal co-state variables at a certain moment in the future based on machine learning algorithms.

4. The attenuation adaptive intelligent fuel cell vehicle optimal energy management system according to claim 3 is characterized in that: The content of the real-time optimal energy management strategy performed by the CPU2 includes: The strategy calculates the adaptive optimal equivalent factor in real time based on the predicted optimal co-state variables and the relationship between the Pontryagin minimum principle and the equivalent consumption minimum strategy; The relationship between the co-state variables and the equivalent factors is: ； In the formula is the adaptive equivalence factor; Based on ECMS, the objective function is constructed with the goal of minimizing the total equivalent hydrogen consumption in real time: ； In the formula is the equivalent hydrogen consumption; By inputting the optimal equivalent factor obtained by solving the problem into the ECMS model, the optimal power distribution of the fuel cell and the power battery is calculated.

5. The attenuation adaptive intelligent fuel cell vehicle optimal energy management system according to claim 3 is characterized in that: The second core CPU1 of the multi-core processor performs prediction based on machine learning, predicts the optimal co-state variables, and migrates its computing tasks to the cloud for execution, thus realizing vehicle-cloud collaborative computing: By uploading historical operating condition data to the cloud; the on-board system uploads the real-time estimation of the fuel cell SOH to the cloud, updates the fuel cell vehicle model on the cloud, and performs machine learning model training and prediction tasks. The optimal co-state variable prediction results completed on the cloud are then transmitted back to the on-board system through the network, allowing the vehicle to apply the optimization results in real time during actual driving.

Images