Fuel cell commercial vehicle control method and system based on offline capacity matching and online energy management optimization strategy

Abstract

The invention discloses a fuel cell commercial vehicle control method and system based on offline capacity matching and an online energy management optimization strategy. Comprising the following steps: acquiring power demand data, and determining an optimal energy source component size according to an energy source component size algorithm and a size solution space constructed by an offline layer; and according to the size of the optimal energy source assembly and a multi-target cost function constructed by an online layer, determining the minimum cost according to a motor power demand, and obtaining the fuel cell output power corresponding to the minimum cost. A decoupling framework based on off-line-on collaborative optimization: an off-line layer adopts a double-layer design, an upper layer realizes efficient parameter space exploration by using a genetic algorithm, and a lower layer guarantees global optimality of a strategy through a linear programming method to obtain an optimal component size; according to the online layer, an electric power steering system comprising four penalty factors is designed according to the actual working condition of the fuel cell commercial vehicle, and instantaneous multi-target cost minimization is achieved.

Description

Technical Field

[0001] This invention relates to the field of fuel cell hybrid power technology, and in particular to a control method and system for fuel cell commercial vehicles based on offline capacity matching and online energy management optimization strategies. Background Technology

[0002] As an important branch of new energy vehicles, fuel cell commercial vehicles, with their advantages of zero emissions, long driving range, and rapid hydrogen refueling, have shown broad application prospects in medium and heavy-duty transportation sectors such as port logistics and intercity freight. However, the unidirectional energy conversion characteristics of fuel cells necessitate the introduction of auxiliary energy storage systems (such as lithium-ion batteries) to meet peak power demands and achieve regenerative braking. Due to the significant differences in dynamic response characteristics and degradation mechanisms between fuel cell systems and lithium-ion batteries, it is essential to optimize component size and coordinate energy management strategies (EMS) to effectively protect the hybrid energy storage system (HESS) and thus meet vehicle power requirements.

[0003] The current power configuration of fuel cells and power batteries needs to take into account both the economic efficiency throughout the entire life cycle and the adaptability to dynamic operating conditions. However, traditional methods often adopt single-objective optimization, which makes it difficult to balance the complex relationship between hydrogen consumption cost, component degradation and initial investment. In addition, existing real-time energy management strategies (including rule-based strategies, equivalent hydrogen consumption strategies and model predictive control) still have significant defects in energy source co-optimization: (1) the level of intelligence of rule-based strategies is low and it is difficult to adapt to complex operating conditions; (2) the equivalent hydrogen consumption strategy (ECMS) has high parameter sensitivity and insufficient robustness; (3) model predictive control (MPC) relies too much on the accuracy of future vehicle speed prediction, resulting in deviations in actual application. Therefore, decoupling and optimizing energy source capacity matching and energy management strategies has become a key breakthrough direction.

[0004] In the process of optimizing the objective function, current research methods mainly have two limitations: most studies use parameter traversal methods or DIRECT algorithms, which require the calculation of a large number of energy source parameter combinations, resulting in a surge in computation time; although a few studies based on metaheuristic algorithms can reduce the computational dimension, their underlying framework still needs to embed traditional energy management strategies, which means that the algorithm still needs to perform strategy adaptive adjustments, and the overall optimization efficiency still cannot meet engineering requirements. Summary of the Invention

[0005] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a control method and system for fuel cell commercial vehicles based on an offline capacity matching and online energy management optimization strategy. This method integrates an optimization framework for offline energy source capacity matching and online energy management strategies, solving the problems of low computational efficiency and insufficient multi-objective collaborative optimization in existing capacity matching and energy management strategies.

[0006] To achieve the above objectives, in a first aspect, the present invention proposes a control method for fuel cell commercial vehicles based on an offline capacity matching and online energy management optimization strategy, the method comprising: S200: Obtain power demand data, and determine the optimal energy source component size based on the energy source component size algorithm and size solution space constructed in the offline layer; the upper layer of the energy source component size algorithm uses a genetic algorithm to achieve efficient parameter space exploration, and the lower layer uses linear programming to achieve global optimality evaluation of energy management strategies; S300: Based on the optimal energy source component size and the multi-objective cost function constructed in the online layer, determine the minimum cost according to the motor power requirement, and obtain the fuel cell output power corresponding to the minimum cost.

[0007] In another embodiment, before constructing the energy source component size algorithm and size solution space, the method further includes: S110, Construct an energy source model; wherein, the energy source model includes: a hydrogen consumption model for a fuel cell system, a lifespan degradation model for a fuel cell system, an equivalent hydrogen consumption model for a power battery system, and a lifespan degradation model for a power battery system.

[0008] S120, Based on the energy source model, construct an energy source component size algorithm and a multi-objective cost function.

[0009] In another embodiment, the energy source component size algorithm includes: the upper layer of the algorithm uses a genetic algorithm to explore component size parameters in the size solution space, and the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter.

[0010] In another embodiment, the upper layer of the algorithm uses a genetic algorithm to explore component size parameters within the size solution space, and the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter, including: S210, Determine the initialization parameters, which include the number of groups, the number of iterations, the crossover probability, the mutation probability, and the bitwise operation mutation probability; S220, Based on the size solution space, a random function is used to determine the initial position of an individual in the population, where the initial position is the component size parameter; S230, Based on the power demand data, linear programming is used to evaluate the global optimality of the energy management strategy corresponding to each initial position; S240: Based on the initialization value, each initial position is selected, crossed, and mutated to become a new initial position, and the process returns to step S230 until the iteration stop condition is met, thereby determining the optimal energy component size.

[0011] In another embodiment, evaluating the global optimality of the energy management strategy corresponding to each initial position using linear programming based on power demand data includes: S231, Determine the average power based on the power demand data; S232, Based on the average power, the minimum hydrogen consumption of the energy management strategy corresponding to each initial position is calculated using a linear programming method.

[0012] In another embodiment, the energy component size algorithm is as follows:

[0013] In another embodiment, constructing the multi-objective cost function includes: S310, Construct a standard cost function with multiple objectives; S320, optimize the standard cost function according to the penalty coefficient to obtain the target cost function.

[0014] In another embodiment, optimizing the standard cost function according to the penalty function to obtain the target cost function includes: optimizing the target cost function according to four penalty coefficients; the four penalty coefficients include two penalty coefficients that constrain the upper and lower limits of battery SOC, and two penalty coefficients that restrict frequent start-stop and frequent load changes of fuel cells.

[0015] In another embodiment, the target cost function is:

[0016] Secondly, this application proposes an energy management system for a fuel cell vehicle, characterized in that the system includes: a fuel cell, a power battery, a power control unit, and a data receiving device. The power control unit includes a memory and a processor. The memory stores a computer program. The data receiving device is used to receive motor power demand information and transmit it to the processor. When the processor executes the computer program according to the motor power demand information, it implements the steps of any of the control methods described above.

[0017] Compared to existing technologies, the advantages of this invention are as follows: A decoupled framework based on offline-online collaborative optimization: The offline layer employs a two-layer design method, with the upper layer utilizing a genetic algorithm for efficient parameter space exploration, and the lower layer using linear programming to ensure the global optimality of the strategy, thereby obtaining the optimal component size for a given driving cycle; the online layer, designed for the actual operating conditions of fuel cell commercial vehicles, incorporates an electric power steering system with four penalty factors to minimize instantaneous multi-objective costs. This also allows for the selection of fuel cell system capacity and reduces the overall cost throughout the fuel cell system's lifecycle. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating an embodiment of the fuel cell commercial vehicle based on an offline-online control method according to the present invention; Figure 2 for Figure 1 The flowchart of the model construction method in the control method shown is shown. Figure 3 for Figure 1 The flowchart shown illustrates the offline layer of the control method based on genetic algorithm and linear programming. Figure 4 for Figure 3 A flowchart of a sub-step of the control method shown; Figure 5 This is a schematic diagram of the entire process algorithm of the offline layer based on genetic algorithm and linear programming according to an embodiment of the present invention; Figure 6 for Figure 1 The flowchart shown illustrates the online layer of the control method based on a multi-objective cost minimization control method. Figure 7 This is a schematic diagram of the entire algorithm of an online layer based on a multi-objective cost minimization control method according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the energy management system structure of a fuel cell commercial vehicle according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the composition of a fuel cell commercial vehicle according to an embodiment of the present invention; Figure 10 for Figure 1 The fitness curve obtained during the driving cycle of a fuel cell commercial vehicle control method according to one embodiment is shown. Figure 11 The simulation results of the total cost curve for 20 repeated CHTC-TT driving cycles are shown in the figure. Figure 12 The simulation results show the cost of the three components over 20 repeated CHTC-TT driving cycles. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0021] Firstly, this invention proposes a control method for fuel cell commercial vehicles based on an offline capacity matching and online energy management optimization strategy, such as... Figure 2 As shown, the method includes: S110, Construct an energy source model; wherein, the energy source model includes: a hydrogen consumption model for a fuel cell system, a lifespan degradation model for a fuel cell system, an equivalent hydrogen consumption model for a power battery system, and a lifespan degradation model for a power battery system.

[0022] S120, Based on the energy source model, construct an energy source component size algorithm and a multi-objective cost function.

[0023] Among them, the energy source component size algorithm and the multi-objective cost function both need to be calculated in conjunction with the energy source model.

[0024] Step 1: Establish the longitudinal dynamics model of the whole vehicle based on the vehicle parameters, as shown in the equation:

[0025] Step 2: Establish the motor model and braking energy recovery strategy.

[0026] Since the drive motor is the sole power source for the FCHCV, it acts as a traction motor during vehicle propulsion and transforms into a generator during regenerative braking. To achieve brake energy recovery, this invention employs a speed-dependent regenerative braking strategy: regenerative braking is disabled when the vehicle speed is below 10 km / h; however, when the vehicle speed exceeds this threshold, the drive motor recovers some energy, and the remaining braking force is provided by the mechanical brakes. The equations for torque and speed are shown below:

[0027] Step 3: Model the energy source described in this invention, including the hydrogen consumption model of the fuel cell system, the life degradation model of the fuel cell system, the equivalent hydrogen consumption model of the power battery system, and the life degradation model of the power battery system.

[0028] Among them, the empirical model of fuel cells shows a direct correlation between hydrogen consumption and efficiency. Based on this relationship, the hydrogen consumption model is shown in the equation:

[0029] Empirical models of fuel cell degradation have been categorized into four main factors: prolonged idle time, frequent start-stop cycles, high load, and rapid dynamic load. Specifically, large variable loads and constant loads lead to a significant reduction in fuel cell lifespan.

[0030] Therefore, the life degradation model of the fuel cell system is shown in the equation:

[0031] During vehicle operation, fluctuations in battery operating status lead to capacity loss, thus shortening its lifespan. This invention, based on the classic Arrhenius degradation model, establishes a battery lifespan degradation model as shown in the equation:

[0032] In another embodiment, such as Figure 1 As shown, the method also includes: S200: Obtain power demand data, and determine the optimal energy source component size based on the energy source component size algorithm and size solution space constructed in the offline layer; the upper layer of the energy source component size algorithm uses a genetic algorithm to achieve efficient parameter space exploration, and the lower layer uses linear programming to achieve global optimality evaluation of energy management strategies; S300: Based on the optimal energy source component size and the multi-objective cost function constructed in the online layer, determine the minimum cost according to the motor power requirement, and obtain the fuel cell output power corresponding to the minimum cost.

[0033] The power requirement includes a given driving cycle. In practical applications, the long-distance transportation conditions for HCVs (High-Speed ​​Transport Vehicles) are very complex, making it difficult to accurately optimize the actual cost of long-distance vehicles using a single driving cycle. In one embodiment of the present invention, a comprehensive driving cycle covering highway, urban, and suburban driving conditions is generated by combining two conditions: CHTC-TT and C-WTVC. The driving cycle includes various parameters such as road gradient angle, vehicle speed, acceleration, and vehicle resistance.

[0034] In another embodiment, the energy source component size algorithm includes: the upper layer of the algorithm uses a genetic algorithm to explore component size parameters in the size solution space, and the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter.

[0035] The algorithm for calculating the size of the energy source component is as follows:

[0036] in:

[0037] Hydrogen fuel cell commercial vehicles exhibit significant power demand characteristics, reaching a peak of 460 kW and an average of 90 kW in the CHTC-TT and C-WTVC drive cycles, respectively. Since the power battery provides some energy, configuring excessively high fuel cell output power leads to inefficiency. Therefore, in a specific embodiment of this invention, the optimal maximum output power range of the fuel cell system is determined to be 100-300 kW. In the CHTC-TT and C-WTVC test cycles, the HCV requires nearly 200 kWh of energy to achieve a pure electric driving range of 100 km. However, increasing the number of series and parallel power batteries significantly increases battery volume. Therefore, to balance energy capacity and space constraints, the power battery capacity is set to not exceed 200 kWh, within a configurable range of 50-200 kWh. That is, the dimensional solution space is an optimal maximum output power range of 100-300 kW for the fuel cell system, with a battery capacity of 50-200 kW. This involves exploring component size parameters within the size solution space, specifically selecting a maximum output power within the 100-300 kW range and a battery capacity within the 50-200 kWh range, thus forming a set of size solutions. In this application, a genetic algorithm is used to optimize the combination of sizes within the size solution space.

[0038] After determining the energy source component sizing algorithm and sizing solution space, selecting a suitable energy management strategy (EMS) is crucial for evaluating the target value. This invention proposes an EMS framework based on linear programming (LP) to achieve efficient computation of the energy source component sizing algorithm. LP, as an efficient offline optimization method, utilizes the convexity of the fuel cell efficiency curve. When the charging and discharging power loss of the power battery is ignored, and the final SOC of the power battery equals the initial SOC in a given driving cycle, the total energy used to drive the vehicle is provided by the fuel cell, which can be expressed as a formula.

[0039]

[0040] Therefore, when evaluating the global optimality of the dimensional solution, the final SOC of the power battery can be kept equal to the initial SOC, and the minimum hydrogen consumption of the fuel cell can be calculated. The obtained minimum hydrogen consumption value is used to evaluate the quality of the dimensional solution.

[0041] In another embodiment, such as Figure 3 As shown, the upper layer of the algorithm uses a genetic algorithm to explore component size parameters within the size solution space, and the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter, including: S210, Determine the initialization parameters, which include the number of groups, the number of iterations, the crossover probability, the mutation probability, and the bitwise operation mutation probability; S220, Based on the size solution space, a random function is used to determine the initial position of an individual in the population, where the initial position is the component size parameter; S230, Based on the power demand data, linear programming is used to evaluate the global optimality of the energy management strategy corresponding to each initial position; S240: Based on the initialization value, each initial position is selected, crossed, and mutated to become a new initial position, and the process returns to step S230 until the iteration stop condition is met, thereby determining the optimal energy component size.

[0042] Among them, such as Figure 5 As shown, the energy source component size algorithm uses a genetic algorithm in the upper layer to explore combinations of size solutions in the size solution space, and a linear normalization method in the lower layer to evaluate the global optimality. The methods include: Step 1: Initialize the genetic algorithm according to the initialization parameters, such as setting the population size to 20 and the number of iterations to 30.

[0043] Step 2: After initialization, a random function is used to determine the initial position of individuals in the population in the size solution space, where the position is the size solution; Step 3: Use linear programming to evaluate the global optimality of the energy management strategy corresponding to each initial position, thereby achieving the evaluation of each individual; record the current optimal individual (the minimum solution corresponding to the global optimality evaluation index). The global optimality evaluation index is the minimum operating cost of the fuel cell system calculated using a linear programming algorithm to maintain the final SOC of the power battery equal to the initial SOC.

[0044] Step 4: Selection operation. Individuals are selected based on their fitness, for example, using roulette wheel selection: the probability of an individual being selected = its own fitness / the total fitness of the population; or tournament selection: randomly select several individuals, and choose the one with the highest fitness to enter the next generation. Fitness can be directly calculated as the reciprocal of the linear programming value; thus, a higher fitness reflects a lower calculated result.

[0045] Step 5: Crossover operation. Selected individuals (parents) are paired up according to the crossover probability, and some chromosome segments are exchanged to produce offspring, increasing population diversity.

[0046] Step 6: Mutation operation. Based on the mutation probability and bitwise operation mutation probability, certain genes on the offspring chromosome are randomly modified to avoid the algorithm getting trapped in local optima (avoiding premature convergence). The mutation probability is usually set between 0.01 and 0.1 (too low a probability makes mutation difficult, too high a probability destroys high-quality genes).

[0047] Step 7: Population Update and Iteration. After selection, crossover, and mutation, a new generation of population (with the same size as the initial population) is generated. The value of the energy source component size algorithm and fitness of each individual in the new population are calculated, and the current best individual (the energy source size corresponding to the minimum fitness value) is recorded.

[0048] Repeat steps 4-7 until the number of iterations is reached, thereby comparing the best individuals in each iteration and determining the smallest solution as the final optimal solution.

[0049] In another embodiment, such as Figure 4 As shown, the method of evaluating the global optimality of the energy management strategy corresponding to each initial position using linear programming includes: S231, Determine the average power based on the power demand data; S232, Based on the average power, the minimum hydrogen consumption of the energy management strategy corresponding to each initial position is calculated using a linear programming method.

[0050] As described above, hydrogen consumption is lowest when the vehicle operates at its average power during a given driving cycle. For a given driving cycle, the average power is calculated, and this average power is used as the power requirement for the energy management strategy. The hydrogen efficiency of the energy management strategy for each individual's location is then calculated. In some embodiments, the hydrogen consumption at average power output for each location can be used as an evaluation metric for the energy management strategy to assess global optimality; that is, in the genetic algorithm, the reciprocal of the minimum hydrogen consumption for the energy management strategy corresponding to each location is used as the fitness.

[0051] The offline layer adopts the above-mentioned two-layer optimization method, which combines parameter space exploration based on genetic algorithm (GA) and global EMS based on linear programming (LP), effectively solving the problem of large-load computation, while ensuring the accuracy of computation. Linear programming (LP) has higher accuracy and faster convergence speed.

[0052] In another embodiment, such as Figure 6 As shown, constructing the multi-objective cost function includes: S310, Construct a standard cost function with multiple objectives; S320, optimize the standard cost function according to the penalty coefficient to obtain the target cost function.

[0053] In the above analysis, a genetic algorithm-linear programming-based energy source component sizing algorithm was used to determine the energy source parameters that minimize the total cost. However, certain physical constraints of the fuel cell system were neglected in the above analysis, including the upper and lower limits of the battery's state of charge (SOC) and the battery's charge and discharge losses. Meanwhile, the goal of the online optimization strategy is to minimize the total cost of the vehicle at each time point. Therefore, these physical constraints are incorporated into the energy source component sizing algorithm by introducing a penalty coefficient. Based on the energy source component sizing algorithm model constructed above, and considering hydrogen consumption and system degradation, a normalized multi-objective standard cost function is constructed, as shown in the equation:

[0054] In another embodiment, optimizing the standard cost function according to the penalty function to obtain the target cost function includes: optimizing the target cost function according to four penalty coefficients; the four penalty coefficients include two penalty coefficients that constrain the upper and lower limits of battery SOC, and two penalty coefficients that restrict frequent start-stop and frequent load changes of fuel cells.

[0055] In another embodiment, the optimized objective cost function after introducing four penalty coefficients is:

[0056] When the battery's state of charge (SOC) approaches its lower limit, the weight of the battery's equivalent hydrogen consumption cost should be increased, discharging should be reduced, and charging should be prioritized. Conversely, when the SOC approaches its upper limit, the weight of the battery's equivalent hydrogen consumption cost should be reduced, allowing discharging to provide energy in conjunction with the fuel cell. Through empirical calibration, the control parameters were set to b=2.5 and c=0.5; these coefficients remain unchanged when the SOC exceeds a preset threshold.

[0057] Introducing a SOC penalty factor leads to frequent start-up and shutdown of fuel cell output power, as well as rapid load changes, resulting in increased start-up and shutdown costs and variable load costs, and significantly reducing the lifespan of the fuel cell.

[0058] like Figure 7 As shown, in order to mitigate these effects while maintaining system efficiency, two penalty coefficients are introduced to limit excessive degradation of the fuel cell due to start-up, shutdown, and variable load, as shown in the equation:

[0059] Cost minimization is achieved through a multi-objective cost function, with dynamic adjustments to energy allocation instantaneously at each time point. Simulation results show that the proposed multi-objective cost minimization EMS not only achieves solution quality comparable to the linear programming (LP) benchmark.

[0060] The above method will be verified through experimental simulation.

[0061] In the offline layer, the initialization parameters are first determined as shown in Table 1 below.

[0062] In practical applications, the long-distance transportation conditions of HCVs (High-Speed ​​Transport Vehicles) are very complex, making it difficult to accurately optimize the actual cost of long-distance vehicles using a single driving cycle. Therefore, it is necessary to combine different driving cycles to optimize HCV parameters to improve vehicle adaptability. In this invention, a comprehensive driving cycle covering highway, urban, and suburban driving conditions is generated by combining CHTC-TT and C-WTVC conditions. Table 1 provides the initialization parameters of the GA (Genetic Algorithm). Figure 10 The fitness curves obtained during the optimization process are shown. As can be seen from the figure, the total running cost decreases with increasing iterations, indicating that the parameter vectors gradually improve during the GA optimization process. The total cost decreased from RMB 170.10 to RMB 164.91, a reduction of 3.15%. The synergy between GA and LP (linear programming) makes efficient navigation possible in a wide search space, fully utilizing the global exploration capabilities of GA and the computational efficiency of LP. Thanks to the efficiency of LP, this optimization framework requires only 3799 seconds of computation time while maintaining the required accuracy. Such high efficiency cannot be achieved if DP (dynamic programming) is used as the EMS within the component-size framework.

[0063] According to the performance curves, the total cost is optimal when the maximum output power of the fuel cell is 178 kW and the battery capacity is 200 kW. This optimized parameter combination will be used to verify the effectiveness of the proposed instantaneous multi-objective EMS (MOEMS).

[0064] In the online layer, such as Figure 11 As shown, a comparative analysis of four energy management systems (EMS) over 20 driving cycles reveals that LP exhibits higher cost-effectiveness (RMB 1,526.50), while the instantaneous multi-objective EMS (MOEMS) costs RMB 1,537.37, only 0.71% higher than LP, demonstrating considerable economic feasibility over a 36,000-second operating period. REMS has a moderately high cost of RMB 1,552.24. In contrast, PFEMS has the highest cost (RMB 1,662.04), primarily because its output power only tracks the average motor power, a strategy extremely detrimental to fuel cell lifespan.

[0065] Figure 12The report presents three components of the total cost: equivalent hydrogen consumption cost, fuel cell degradation cost, and battery degradation cost. LP exhibits the lowest equivalent hydrogen consumption cost, followed by MOEMS, then PFEMS, while REMS has the highest cost. This indicates that MOEMS achieves comparable hydrogen economy to LP over a preset driving cycle. MOEMS effectively manages degradation costs, second only to LP in terms of degradation costs for both energy sources. The slight difference between MOEMS and LP is primarily attributed to minor fluctuations in fuel cell output power, but these fluctuations remain within acceptable limits. Similarly, REMS performs well in balancing degradation costs for both energy sources, but its performance is slightly inferior to MOEMS due to its wider load variation range. Conversely, PFEMS excels in managing battery degradation but performs very poorly in managing fuel cell degradation.

[0066] In terms of computational efficiency, the average single-step computation time for PFEMS and REMS is... s, while MOEMS is slightly higher Both online EMS systems can meet real-time requirements, while LP, although having a longer computational latency, remains cost-effective globally.

[0067] In summary, the four coefficients introduced in MOEMS effectively limit the range of SOC variation, reduce the load transfer frequency of the fuel cell, and keep the output power within the average range calculated by LP. Considering that the marginal cost difference between LP and MOEMS is less than 1%, coupled with the real-time response capability of MOEMS (less than 0.2 milliseconds per step), this strategy demonstrates its feasibility in practical vehicle applications.

[0068] In summary, the four coefficients introduced in MOEMS effectively limit the range of SOC variation, reduce the load transfer frequency of the fuel cell, and keep the output power within the average range calculated by LP. Considering that the marginal cost difference between LP and MOEMS is less than 1%, coupled with the real-time response capability of MOEMS (less than 0.2 milliseconds per step), this strategy demonstrates its feasibility in practical vehicle applications.

[0069] Secondly, this application proposes an energy management system 1000 for a fuel cell vehicle, characterized in that the system includes: a fuel cell, a power battery, a power control unit, and a data receiving device. The power control unit includes a memory and a processor. The memory stores a computer program. The data receiving device is used to receive motor power demand information and transmit it to the processor. When the processor executes the computer program according to the motor power demand information, it implements the steps of the control method described above.

[0070] Figure 9 The diagram shows the structural structure of a fuel cell hybrid vehicle system provided in an embodiment of the present invention. The system includes a hydrogen tank, a proton exchange membrane fuel cell (PEMFC) system, a DC-DC converter, a DC-AC inverter, a drive motor, and a transmission system. The drive motor is the sole power transmission device of the system. The fuel cell, Li battery, and DC / DC converter constitute a composite power system, wherein the fuel cell and Li battery serve as energy sources, and the required power is provided jointly by the fuel cell and Li battery. Furthermore, the battery can be charged by the fuel cell or through regenerative braking.

[0071] Among them, such as Figure 8 As shown, the data receiving device 1100 receives vehicle operating condition information from vehicle data acquisition devices and / or navigation systems and / or vehicle remote monitoring terminals, and transmits the operating condition information to the power control unit 1200. The power control unit 1200 is the vehicle's own data processing and computing center, including a processor and a memory. The memory stores programs such as the methods described above. The processor determines the maximum power and battery capacity of the fuel cell according to the offline control method described above, based on a given driving cycle. Then, based on the power demand sent by the driver to the motor, it determines the minimum cost according to the online control method described above, thereby determining the fuel cell output power corresponding to the minimum cost. In addition, the power control unit generates a hydrogen release control command based on the determined fuel cell output power and sends it to the hydrogen release control device 1300. The hydrogen control device 1300 releases hydrogen stored in the hydrogen storage device 1400 according to the hydrogen release control command, thereby controlling the power output of the fuel cell. The hydrogen storage device 1400 can be a compressed hydrogen storage tank or a metal hydrogen storage tank. When the power output of the fuel cell exceeds the motor demand, the excess power is used to charge the Li battery 1600.

[0072] Thirdly, this application proposes a fuel cell vehicle, which includes a body, a chassis, and an energy management system 1000 as described above.

[0073] The basic components of a fuel cell vehicle include: body, chassis, fuel cell booster (such as a DC / DC converter), fuel cell, power control unit 1200, drive motor, hydrogen storage device 1400, Li battery 1600, hydrogen release control device 1300, data receiving device 1100, etc.

[0074] Fuel cell boost converter: Used to increase voltage to meet the maximum output requirements of the drive motor. Since the overall voltage of the fuel cell does not exceed 300V, a boost converter is needed to further increase the voltage for better power supply to the drive motor. This is typically a DC-DC converter located at the fuel cell output. The DC-DC power from the DC-DC converter can charge the 1600mAh battery and / or, after passing through a DC / AC converter, be used to drive the motor. This allows the drive motor to drive the wheels through mechanical transmission devices such as a gearbox, thus propelling the vehicle.

[0075] Fuel cell: Fuel cells are a core component of automobiles. In this embodiment, the fuel cell is a hydrogen fuel cell 1500, which typically consists of hundreds of fuel cells and utilizes charge transfer during the chemical reaction of hydrogen and oxygen to generate an electric current.

[0076] Power Control Unit 1200: The power control unit 1200 of a fuel cell vehicle can control the charging and discharging of the Li battery 1600 and the power output control of the fuel cell under different driving conditions, thereby optimizing strategies and improving energy utilization economy. For example, Figure 9 As shown, the power control unit includes a processor and memory, as well as a DC / DC converter connected to the output of the fuel cell, a DC / DA converter connected between the drive motor and the DC / DC converter, and a DC bus connected between the DC / DA converter and the DC / DC converter. Drive motor: Converts the electrical energy output from the fuel cell and / or power battery 1600 into mechanical energy to drive the vehicle. Motor types include DC motors, AC motors, permanent magnet synchronous motors, and switched reluctance motors, etc. For example, Figure 9 As shown, the drive motor is connected to a DA / DC converter to receive electrical energy from the fuel cell and / or power battery 1600. The drive motor then drives the wheels to rotate via a transmission, thereby propelling the vehicle.

[0077] Hydrogen storage device 1400: The hydrogen storage device 1400 can be a high-pressure hydrogen storage tank, a hydrogen storage device connected to the fuel cell to supply hydrogen to the fuel cell. It is typically located in the trunk or under the chassis of the vehicle. These tanks can store up to 5 kg of hydrogen at a pressure of up to 70 MPa. These carbon fiber Kevlar composite hydrogen storage tanks are not only lightweight but also resistant to light firearms attacks, offering excellent safety performance. In some other embodiments, the hydrogen storage device 1400 can also use a metal hydrogen storage tank, which offers better hydrogen storage capacity and safety than high-pressure hydrogen storage.

[0078] Li-ion 1600 battery: Used to store the excess electrical energy output by the fuel cell and the electrical energy recovered during vehicle operation, as well as to assist the fuel cell in powering acceleration. In hydrogen fuel cell vehicles, the 1600 power battery is typically located behind the rear seats and can be used independently to power the vehicle when the overall vehicle load is low. For example... Figure 9 As shown, the power battery 1600 is connected to the DC / DC converter 0 to receive excess electrical energy from the fuel cell 0. Additionally, the power battery 1600 is also connected to the DC bus to provide power to the drive motor, or to reverse charge the Li battery via the DC / AC converter 122 when the drive motor is reverse-driven.

[0079] The hydrogen storage device 1400, hydrogen release control device 1300, power control unit 1200, and data receiving device 1100 constitute a capacity management system for managing the power output of the fuel cell.

[0080] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A control method for fuel cell commercial vehicles based on offline capacity matching and online energy management optimization strategies, characterized in that, include: S200: Obtain power demand data, and determine the optimal energy source component size based on the energy source component size algorithm and size solution space constructed in the offline layer; the upper layer of the energy source component size algorithm uses a genetic algorithm to achieve efficient parameter space exploration, and the lower layer uses linear programming to achieve global optimality evaluation of energy management strategies; S300: Based on the optimal energy source component size and the multi-objective cost function constructed in the online layer, determine the minimum cost according to the motor power requirement, and obtain the fuel cell output power corresponding to the minimum cost.

2. The control method according to claim 2, characterized in that, Before constructing the energy source component size algorithm and size solution space, the method also includes: S110, Construct an energy source model; wherein, the energy source model includes: a hydrogen consumption model for a fuel cell system, a lifespan degradation model for a fuel cell system, an equivalent hydrogen consumption model for a power battery system, and a lifespan degradation model for a power battery system. S120, Based on the energy source model, construct an energy source component size algorithm and a multi-objective cost function.

3. The control method according to claim 1, characterized in that, The energy source component size algorithm includes: the upper layer of the algorithm uses a genetic algorithm to explore component size parameters within the size solution space, and the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter.

4. The control method according to claim 3, characterized in that, The upper layer of the algorithm uses a genetic algorithm to explore component size parameters within the size solution space, while the lower layer uses linear programming to evaluate the global optimality of the energy management strategy corresponding to any component size parameter, including: S210, Determine the initialization parameters, which include the number of groups, the number of iterations, the crossover probability, the mutation probability, and the bitwise operation mutation probability; S220, Based on the size solution space, a random function is used to determine the initial position of an individual in the population, where the initial position is the component size parameter; S230, Based on the power demand data, linear programming is used to evaluate the global optimality of the energy management strategy corresponding to each initial position; S240: Based on the initialization value, each initial position is selected, crossed, and mutated to become a new initial position, and the process returns to step S230 until the iteration stop condition is met, thereby determining the optimal energy component size.

5. The control method according to claim 4, characterized in that, Based on the power demand data, the evaluation of the global optimality of the energy management strategy corresponding to each initial position using linear programming includes: S231, Determine the average power based on the power demand data; S232, Based on the average power, the minimum hydrogen consumption of the energy management strategy corresponding to each initial position is calculated using a linear programming method.

6. The control method according to claim 1, characterized in that, The energy component size algorithm is as follows: The dimension solution is: in, Represents the total cost. This indicates the hydrogen consumption cost of fuel cells. This represents the equivalent hydrogen consumption cost of the battery. Indicates the degradation cost of fuel cells, This represents the cost of battery degradation, where T represents the duration of a given driving cycle. This indicates the maximum output power of the fuel cell. U represents the battery capacity; U represents the dimensional solution space, where u represents the dimensional solution for any given driving cycle.

7. The control method according to claim 1, characterized in that, Constructing the multi-objective cost function includes: S310, Construct a standard cost function with multiple objectives; S320, optimize the standard cost function according to the penalty coefficient to obtain the target cost function.

8. The control method according to claim 6, characterized in that, The step of optimizing the standard cost function according to the penalty function to obtain the target cost function includes: optimizing the target cost function according to four penalty coefficients; the four penalty coefficients include two penalty coefficients that constrain the upper and lower limits of battery SOC, and two penalty coefficients that restrict frequent start-stop and frequent load changes of fuel cells.

9. The control method according to claim 7, characterized in that, The objective cost function is: in, The state of charge of the power battery. and These are the upper and lower limits of the state of charge of the power battery, respectively. For the output power of the fuel cell, and These are the minimum and maximum output power of the fuel cell. The output power of the power battery, This represents the maximum charging power of the power battery. This represents the maximum discharge power of the power battery. and These are the two penalty coefficients for the power battery. and For fuel cells, there are two penalty factors. These represent the lifespan degradation of fuel cells caused by constant load degradation, variable load degradation, and start-stop degradation, respectively. This represents the minimum total cost.

10. An energy management system for a fuel cell vehicle, characterized in that, The system includes: a fuel cell, a power battery, a power control unit, and a data receiving device. The power control unit includes a memory and a processor. The memory stores a computer program. The data receiving device is used to receive motor power demand information and transmit it to the processor. When the processor executes the computer program according to the motor power demand information, it implements the steps of the control method as described in any one of claims 1 to 9.

Images