Optimized component type selection and power distribution method for fuel cell hybrid power system

Abstract

The invention discloses a fuel cell hybrid power system optimization part type selection and power distribution method. The method comprises the following steps: establishing a simulation model comprising a fuel cell stack, a power cell, fuel cell voltage attenuation and power cell capacity attenuation; a double-layer optimization architecture is established, the upper layer carries out model selection optimization on the fuel cell power and the power cell capacity based on a genetic algorithm, and the lower layer carries out hybrid power system power distribution based on an improved dynamic planning method. An improved dynamic planning algorithm adopts a method of combining parallel operation and coarse and fine grids; and a double-layer architecture optimization method is adopted to find a part type selection and power distribution result which enables the whole life cycle cost to be lowest. According to the method, collaborative optimization of component type selection and power distribution of the fuel cell hybrid power system with the lowest life cycle cost is achieved, the method is suitable for various fuel cell systems such as mobile fuel cell systems and fixed fuel cell systems, operability is high, system performance is effectively improved, and the life cycle cost of the hybrid power system is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fuel cells, and particularly relates to a fuel cell hybrid power system component selection and power distribution optimization method. BACKGROUND

[0002] Fuel cell hybrid power systems have attracted widespread attention in the field of new energy vehicles and stationary energy due to their high efficiency and environmental protection. However, the key challenge of this system is the reasonable selection of fuel cell stacks and power batteries and the power distribution strategy, which directly affects the energy consumption, life and economy of the system.

[0003] The coordinated optimization of component selection and power distribution is the key to reducing the life cycle cost. Component selection determines the rated power and capacity of fuel cell stacks and power batteries, while power distribution determines the energy flow and component load level during system operation. Traditional methods usually optimize the two separately, resulting in mismatch between selection results and operation strategy, affecting the overall economy. The present application adopts a double-layer optimization architecture, the upper layer uses genetic algorithm to optimize component selection, and the lower layer uses an improved dynamic programming method to optimize power distribution, considering the actual operation during the selection stage, so as to ensure that the final scheme considers energy consumption, life and economy, and realizes the global optimal design of fuel cell hybrid power system.

[0004] Dynamic programming (DP) is a method commonly used to solve optimal control problems, suitable for optimization problems with stage decision characteristics. In fuel cell hybrid power system power distribution, DP can solve the optimal strategy by recursion according to system state, control variables and cost function. However, the standard DP method has high computational complexity, which is difficult to meet the actual engineering application requirements.

[0005] Traditional power distribution methods are mainly based on empirical rules or short-term economic optimization, and do not fully consider the life cycle cost, resulting in high hydrogen consumption and premature component degradation during system operation, increasing long-term operating costs. Therefore, there is an urgent need for an optimization method that considers the life cycle cost to improve the overall performance and economic benefits of fuel cell hybrid power systems. SUMMARY

[0006] In view of the deficiencies in the prior art, the purpose of the present application is to provide a fuel cell hybrid power system optimization component selection and power distribution method, which realizes reasonable component selection and energy optimization distribution by establishing a double-layer optimization architecture, so that the system meets the operation requirements while reducing the life cycle cost. This method improves the economy and durability of the fuel cell system on the basis of optimizing the calculation efficiency and accuracy, and is suitable for different types of fuel cell application scenarios. In order to achieve the above purposes and other advantages according to the present application, a fuel cell hybrid power system optimization component selection and power distribution method is provided, comprising:

[0007] A plurality of simulation models are established, including a fuel cell stack simulation model, a power battery simulation model, a fuel cell voltage decay simulation model, and a power battery capacity decay simulation model;

[0008] A double-layer optimization architecture is constructed, including an upper-layer architecture and a lower-layer architecture, the upper-layer architecture is based on genetic algorithm to select and optimize the fuel cell power and the power battery capacity;

[0009] The lower-layer architecture is based on an improved dynamic programming method to perform power distribution of the hybrid system.

[0010] Preferably, the fuel cell stack simulation model is calculated based on a polarization curve equation, as shown below:

[0011] V cell =E-V act -V ohm -V con , wherein Vcell represents the fuel cell voltage, E, Vact, Vohm, and Vcon represent the Nernst voltage, the activation overpotential, the ohmic overpotential, and the concentration overpotential, respectively.

[0012] Preferably, the power battery simulation model is established by using the R-int model.

[0013] Preferably, the fuel cell voltage decay rate of the fuel cell voltage decay simulation model is obtained according to the following formula: ΔP loss =k p (k1×t1+k2×n2+k3×t3+k4×t4), wherein ΔPloss(%) represents the decay rate of the fuel cell within a time interval Δt, kp represents a correction coefficient considering the difference between experiments and actual conditions, k1-k4 represent the experimentally measured fuel cell power decay rates under various harsh working conditions, t1, n2, t3, and t4 represent the fuel cell idle time, the number of start-stop times, the output power mutation time, and the high load operation time, respectively, with units of h and times.

[0014] Compared with the prior art, the present application has the beneficial effects of: life cycle optimization: considering hydrogen consumption, battery attenuation cost and system durability, long-term economic optimization is realized. Double-layer optimization architecture: combining genetic algorithm and improved dynamic programming, collaborative optimization of component selection and power distribution is realized. High calculation efficiency: using parallel operation and dynamic programming method combining coarse and fine grids, the calculation efficiency is improved while the accuracy is ensured. Wide application range: applicable to mobile and fixed fuel cell systems, with strong adaptability and operability. Reducing operating cost: optimizing power distribution strategy, prolonging the life of fuel cell and power battery, reducing maintenance cost and energy consumption. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 An algorithm flowchart of the fuel cell hybrid power system component selection and power distribution optimization method according to the present application;

[0016] Figure 2 A double-layer optimization architecture diagram of the fuel cell hybrid power system component selection and power distribution optimization method according to the present application;

[0017] Figure 3 A simulation result diagram based on CLTC-B working condition of the fuel cell hybrid power system component selection and power distribution optimization method according to the present application;

[0018] Figure 4 A simulation result diagram based on NEDC working condition of the fuel cell hybrid power system component selection and power distribution optimization method according to the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0020] REFERENCE Figures 1-4 A fuel cell hybrid power system component selection and power distribution optimization method, comprising:

[0021] A plurality of simulation models are established, including a fuel cell stack simulation model, a power battery simulation model, a fuel cell voltage attenuation simulation model, and a power battery capacity attenuation simulation model;

[0022] A fuel cell power system model is established, including a fuel cell system simulation module, a power battery simulation module, a fuel cell voltage attenuation simulation module, and a power battery capacity attenuation simulation module. The fuel cell adopts a quasi-static model, including a stack net power, a system hydrogen consumption rate, and a system efficiency.

[0023] The stack net power calculation formula is as follows: P fc = P st -P aux The system hydrogen consumption rate can be calculated by the following formula: The system efficiency is calculated according to the following formula: The power battery establishes a Rint model, wherein the battery current is calculated according to the following formula: The battery voltage calculation formula is as follows: U bat = E bat -I bat R bat When there is current passing through, the power battery SOC change amount and the corresponding time SOC value SOCbat calculation formula is as follows: The power battery efficiency value in different charging and discharging processes can be calculated according to the following formula:

[0024] The fuel cell voltage attenuation rate of the fuel cell voltage attenuation simulation model is obtained according to the following formula: ΔP loss = k p (k1×t1+k2×n2+k3×t3+k4×t4), wherein ΔP loss (%) represents the attenuation rate of the fuel cell within a time interval Δt, kp represents a correction coefficient considering the difference between experiments and actual conditions, k1-k4 represent the fuel cell power attenuation rates measured in various harsh working conditions, t1, n2, t3, and t4 represent the fuel cell idle time, the number of start-stop times, the output power mutation time, and the high load running time, respectively, with units of h and times.

[0025] The power battery capacity attenuation establishes an Arrhenius model, as shown in the following formula: The Arrhenius model is discretized to obtain the result shown in the following formula

[0026]

[0027] A double-layer optimization architecture is constructed, including an upper layer architecture and a lower layer architecture. The upper layer architecture is based on a genetic algorithm for selecting and optimizing the fuel cell power and the power battery capacity. The optimization objective function is: The specific steps of the genetic algorithm are as follows: (1) initialize the population: generate multiple individuals, each individual representing a [P fc ,Wbat ] parameter combination; (2) fitness evaluation: call the lower layer architecture (dynamic programming) to calculate the objective function value under the corresponding control strategy; (3) selection: select individuals with high fitness by methods such as roulette; (4) crossover: cross the selected individuals to generate new individuals; (5) mutation: disturb part of the gene sites to introduce diversity; (6) iterative update: repeat steps (2)-(5) until the termination condition (such as fitness convergence or reaching the maximum number of iterations) is met; the lower layer architecture is based on an improved dynamic programming method for hybrid system power distribution, and the improvement points include, grid design: adopt a two-stage coarse and fine grid search strategy, first lock the optimal control area in the global range with large step (△SOC=5%, △P fc =5kW), and then search in the area with small step (△SOC=1%, △P fc =1kW) to improve the accuracy of the solution; parallel optimization calculation: divide the entire driving condition into urban, suburban and highway sub-conditions, and perform parallel optimization respectively.

[0028] The dynamic programming optimization process includes, (1) definition and discretization of variable space: taking the SOC of the power battery as the state variable, and taking the output power P fc of the fuel cell as the control variable, setting the feasible region and upper and lower limits of the state variable and the control variable, and initially using large step discretization (△SOC=5%, △P fc =5kW) to construct a coarse grid to form an initial search space; (2) stage cost function construction: construct a cost function that comprehensively considers hydrogen consumption, equivalent hydrogen consumption, fuel cell aging and battery aging cost, and the expression is as follows: (3) coarse grid search stage: use coarse grid to recursively calculate the cumulative cost function for each time step in the entire control space, and record the optimal control amount and its neighborhood at each state; (4) fine grid refinement stage: within the neighborhood of the optimal solution at each time step determined by coarse search, a finer grid is reconstructed with smaller step (△SOC=1%, △P fc =1kW) to further accurately search for the optimal control solution and improve the optimization accuracy; (5) parallel computing strategy: to improve the calculation efficiency, the driving condition is divided into multiple typical road sections (urban, suburban and highway) for task-level parallel computing, which significantly speeds up the execution speed of the entire dynamic programming process; (6) optimal path backtracking: after the dynamic programming completes the optimal value calculation of all time steps, the optimal fuel cell power output P fc (t) corresponding to each time step is recovered step by step from the target endpoint to the initial state to obtain the complete power distribution path. Further, the fuel cell stack simulation model is calculated based on the polarization curve equation as follows:

[0029] Vcell =EV act -V ohm -V con , where Vcell represents the fuel cell voltage, E, Vact, Vohm, and Vcon represent the Nernst voltage, activation overpotential, ohmic overpotential, and concentration overpotential, respectively.

[0030] Furthermore, the power battery simulation model is established using the R-int model.

[0031] Furthermore, the fuel cell voltage decay rate of the fuel cell voltage decay simulation model is obtained according to the following formula: ΔP loss =k p (k1×t1+k2×n2+k3×t3+k4×t4), where ΔPloss (%) represents the attenuation rate of the fuel cell within the time interval Δt, kp represents the correction coefficient considering the difference between the experiment and the actual situation, k1-k4 represent the fuel cell power attenuation rate measured experimentally under various harsh working conditions, t1, n2, t3, and t4 represent the fuel cell idling time, number of starts and stops, output power mutation time, and high-load operation time, respectively, with units of h and times.

[0032] Furthermore, the power battery capacity decay rate of the power battery capacity decay simulation model is determined according to the Arrhenius model.

[0033] Furthermore, the lower-level architecture performs hybrid system power allocation based on an improved dynamic programming method, wherein the objective function of the power allocation includes four parts: stack hydrogen consumption cost L1, battery equivalent hydrogen consumption cost L2, fuel cell life attenuation cost L3, and power battery life attenuation cost L4; wherein: Where m H2 [x(k),u(k),k] represents the instantaneous hydrogen consumption of the fuel cell system, in kg; m H2,equ [x(k),u(k),k] represents the instantaneous equivalent hydrogen consumption of the power battery, in kg; P fc_max Indicates the maximum power of the fuel cell, unit KW; C fc_init Indicates the initial purchase cost of the fuel cell, unit KW / yuan; ΔP loss is the instantaneous voltage decay rate of the fuel cell, unit is %; P loss_end =10, unit %, is the voltage decay rate of the fuel cell when it reaches the end of its vehicle life; W batt_norm Indicates the rated capacity of the power battery, in KWh; C bat_init Indicates the initial purchase cost of the power battery, unit KWh / yuan, ΔW loss is the instantaneous capacity attenuation rate of the power battery, unit: %; W loss_end= 20, unit %, the voltage attenuation rate of the power battery when reaching the end of the vehicle service life.

[0034] Further, the control variable of the power distribution strategy based on dynamic programming is fuel cell power, the state variable is power battery SOC value, and the constraint conditions include P fc_min ≤ P fc (k) ≤ P fc_max , P bat_min ≤ P bat (k) ≤ P bat_max , ΔP fc_min ≤ ΔP fc (k) ≤ ΔP fc_max , SOC min ≤ SOC(k) ≤ SOC max , SOC(0) = SOC(end), respectively limiting the fuel cell power range, the power battery power range, the fuel cell power change rate range, the power battery SOC fluctuation range and the SOC value at the end time.

[0035] Further, in the power distribution of the hybrid system based on the improved dynamic programming method, a method combining parallel computing and coarse and fine grids is adopted, and the cycle working condition is divided into several segments according to its operating characteristics, the segments including city, suburb and highway; and dynamic programming optimization calculation is performed in parallel for each segment, and at each calculation time step, the control variable and the state variable are first divided in the feasible range according to a large step (△SOC = 5%, △P fc = 5kW) to quickly find the optimal solution range of each time step that minimizes the life cycle cost, and then the optimal solution coarse grid of each time step is divided according to a smaller step (△SOC = 1%, △P fc = 1kW) to more accurately determine the optimal point in the optimal range. The above method improves the calculation efficiency and ensures the calculation accuracy.

[0036] Embodiment:

[0037] This embodiment is based on the double-layer collaborative optimization method of the fuel cell hybrid power system proposed in the application, and simulation tests are performed for two typical working conditions of CLTC-B and NEDC.

[0038] In this embodiment, a double-layer optimization architecture including an upper layer and a lower layer is constructed:

[0039] (I) Upper layer optimization architecture - component selection based on genetic algorithm

[0040] The upper layer optimization target is to determine the key configuration parameters of the fuel cell system, including the rated power of the fuel cell stack and the capacity of the power battery. The optimization objective function is as follows:

[0041] The standard genetic algorithm (GA) is used for solving the problem. The specific process is as follows:

[0042] 1. Encoding method: P fc and W bat Use real number coding;

[0043] 2. Fitness function: Calls the underlying dynamic programming module, calculates and returns based on the current configuration;

[0044] 3. Selection, crossover, and mutation operations: using roulette wheel selection, single-point crossover, and Gaussian mutation strategies;

[0045] 4. Termination criteria: After iterative optimization, the optimal configuration combination suitable for the current working conditions is finally output [P fc ,W bat ].

[0046] (2) Lower-layer optimization architecture: power allocation based on improved dynamic programming

[0047] Based on the component configuration parameters output by the upper layer, a lower-level dynamic programming model is constructed to optimize the energy distribution of the entire vehicle at each time step during the entire operating process. That is, while ensuring that the driving power requirements are met, a reasonable power distribution relationship between the fuel cell and the battery is determined.

[0048] The dynamic programming design is as follows:

[0049] 1. State variables and control variables: State variable: battery state of charge SOC, discrete range [0.3, 0.8]; Control variable: fuel cell output power P fc .

[0050] 2. Objective function: Construct a function that comprehensively considers hydrogen consumption, equivalent hydrogen consumption, fuel cell aging, and battery aging costs. The expression is as follows:

[0051] 3. Coarse and fine grid joint search strategy: In the initial stage, the coarse grid is searched in the entire feasible solution space (the step size is set as: △SOC=5%, △P fc =5kW,

[0052] Quickly obtain the optimal control area for each time step; based on the coarse search, perform a fine grid search for the optimal interval (the step size is set to: △SOC = 1%,

[0053] △P fc =1kW), improving optimization accuracy.

[0054] 4. State transfer and cost accumulation: Perform state transfer and calculate and store the optimal cost, and finally obtain the complete optimal control trajectory through path backtracking.

[0055] 5. Parallel acceleration mechanism: Perform task-level parallel division for different time periods (urban section, suburban section, highway section) to significantly improve overall operating efficiency.

[0056] (3) Simulation results and analysis

[0057] The above optimization process was implemented using the MATLAB / Simulink platform, and simulations were performed under CLTC-B and NEDC conditions. The simulation results are shown in the attached figure. Figure 3 With attached Figure 4 shown.

[0058] The number of devices and processing scales described herein are intended to simplify the description of the present invention, and the application, modification, and variation of the present invention will be apparent to those skilled in the art. Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiment. They can be applied to various fields suitable for the present invention. For those skilled in the art, additional modifications can be easily implemented. Therefore, the present invention is not limited to the specific details and illustrations shown and described herein without departing from the general concept defined by the claims and their equivalents.

Claims

1. A method for optimizing component selection and power distribution of a fuel cell hybrid system, characterized in that: include: Establishing multiple simulation models, including a fuel cell stack simulation model, a power battery simulation model, a fuel cell voltage decay simulation model, and a power battery capacity decay simulation model; Constructing a two-layer optimization architecture, which includes an upper layer architecture and a lower layer architecture. The upper layer architecture optimizes the fuel cell power and power battery capacity based on a genetic algorithm. The lower-level architecture performs hybrid system power distribution based on an improved dynamic programming method.

2. A method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, characterized in that: The fuel cell stack simulation model is calculated based on the polarization curve equation, as shown below: V cell =EV act -V ohm -V con , where Vcell represents the fuel cell voltage, E, Vact, Vohm, and Vcon represent the Nernst voltage, activation overpotential, ohmic overpotential, and concentration overpotential, respectively.

3. The method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, characterized in that: The power battery simulation model is established using the R-int model.

4. A method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, characterized in that: The fuel cell voltage decay rate of the fuel cell voltage decay simulation model is obtained according to the following formula: ΔP loss =k p (k1×t1+k2×n2+k3×t3+k4×t4), where ΔPloss (%) represents the attenuation rate of the fuel cell within the time interval Δt, kp represents the correction coefficient considering the difference between the experiment and the actual situation, k1-k4 represent the fuel cell power attenuation rate measured experimentally under various harsh working conditions, t1, n2, t3, and t4 represent the fuel cell idling time, number of starts and stops, output power mutation time, and high-load operation time, respectively, with units of h and times.

5. The method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, wherein: The power battery capacity decay rate of the power battery capacity decay simulation model is determined according to the Arrhenius model.

6. The method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, wherein: The lower-level architecture performs hybrid system power allocation based on an improved dynamic programming method, wherein the objective function of the power allocation includes four parts: the stack hydrogen consumption cost L1, the battery equivalent hydrogen consumption cost L2, the fuel cell life degradation cost L3, and the power battery life degradation cost L4; wherein: Where m H2 [x(k),u(k),k] represents the instantaneous hydrogen consumption of the fuel cell system, in kg; m H2,equ [x(k),u(k),k] represents the instantaneous equivalent hydrogen consumption of the power battery, in kg; P fc_max Indicates the maximum power of the fuel cell, unit KW; C fc_init Indicates the initial purchase cost of the fuel cell, unit KW / yuan; ΔP loss is the instantaneous voltage decay rate of the fuel cell, unit is %; P loss_end =10, unit %, is the voltage decay rate of the fuel cell when it reaches the end of its vehicle life; W batt_norm Indicates the rated capacity of the power battery, in KWh; C bat_init Indicates the initial purchase cost of the power battery, unit KWh / yuan, ΔW loss is the instantaneous capacity attenuation rate of the power battery, unit: %; W loss_end =20, unit %, is the voltage decay rate of the power battery when it reaches the end of its vehicle life.

7. The method for optimizing component selection and power distribution of a fuel cell hybrid power system according to claim 1, characterized in that: In the hybrid system power distribution based on the improved dynamic programming method, a method combining parallel computing and coarse and fine grids is used to divide the cycle operating conditions into several segments according to their operating characteristics, wherein the segments include urban, suburban, and highway segments. Dynamic programming optimization calculations are performed in parallel in each segment. At the same time, within each segment, at each calculation time step, the control variables and state variables are first divided according to a larger step size within their feasible domain to quickly find the optimal solution range that minimizes the full life cycle cost at each time step. Then, the coarse grid of the optimal solution at each time step is divided according to a smaller step size to more accurately determine the optimal point within the optimal range.

Images