A PEM-ORC flexible combined intelligent power generation system and method

By combining proton exchange membrane fuel cell with ORC system, using linear regression model and whale intelligent optimization algorithm, the problems of waste heat recovery and parameter regulation in fuel cell power generation system are solved, the power generation efficiency and energy utilization rate are improved, and the system is realized efficient operation and stable control.

CN118472331BActive Publication Date: 2024-11-08MAOMING ELECTRIC POWER ENGINEERING SUPERVISION CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410587258.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-13
Publication Date
2024-11-08
Estimated Expiration
2044-05-13

AI Technical Summary

Technical Problem

The existing distributed power generation system combining proton exchange membrane fuel cells with traditional generators has problems such as inability to recover waste heat, low energy utilization, low power generation efficiency and incompatible regulation, resulting in complex operation.

Method used

Using PEM-ORC flexible combined with intelligent power generation system, waste heat is recovered through the low-grade waste heat generated by the proton exchange membrane fuel cell for the ORC system, and the linear regression model and gradient descent whale intelligent optimization algorithm are used to accurately control the working parameters of the fuel cell to improve the Power generation efficiency.

Benefits of technology

It effectively improves the energy utilization rate and power generation efficiency of the power generation system, reduces unnecessary energy losses, and through real-time data monitoring and intelligent control, it avoids fluctuations caused by human factors, and improves the system's adaptability and control accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118472331B_ABST
    Figure CN118472331B_ABST
Patent Text Reader

Abstract

The present invention discloses a PEM-ORC flexible combined intelligent power generation system and method, which is composed of a proton exchange membrane fuel cell and an organic Rankine cycle system. The system uses a proton exchange membrane fuel cell to generate electricity using hydrogen and oxygen, and inputs the electricity of the proton exchange membrane fuel cell into the generator of the ORC cycle system. The waste heat in the process of power generation of the fuel cell can be recovered through the ORC cycle system, thereby improving the power generation efficiency of the entire system. The SGD-WOA gradient descent whale intelligent optimization algorithm and linear regression model predict and adjust the battery environmental parameters to ensure that the working environment of the fuel cell remains optimal, thereby further maximizing the final power generation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to an intelligent power generation system, and in particular to a PEM-ORC flexible combined intelligent power generation system and method. Background Art

[0002] There are many application cases of fuel cells in new energy distributed power generation systems. Among them, proton exchange membrane fuel cells (PEMFC) have gained more application markets due to their low investment costs in the early stage. Proton exchange membrane fuel cells (PEMFC) are combined with traditional generators to form a distributed power generation system, which directly meets user needs and jointly provides emergency power supply for nearby users. However, the existing power generation system has the problem that waste heat cannot be recovered, energy utilization is not high, and power generation efficiency is not high. The regulation of the two is incompatible, resulting in overly complicated operation. Summary of the invention

[0003] Purpose of the invention: In view of the above problems, the present invention proposes a PEM-ORC flexible combined intelligent power generation system and method, which can improve the energy utilization rate and power generation efficiency of the power generation system.

[0004] Technical solution: The technical solution adopted by the present invention is a PEM-ORC flexible combined intelligent power generation system, including:

[0005] Proton exchange membrane fuel cells, which use hydrogen and oxygen to generate electricity and power the ORC system;

[0006] The ORC system uses low-grade waste heat to generate electricity. The low-grade waste heat includes hot waste water and hot water vapor generated during the power generation process of the proton exchange membrane fuel cell, and the waste heat is recovered through the heat exchanger in the ORC system;

[0007] A parameter feedback device for collecting internal environmental data of the proton exchange membrane fuel cell and sending it to a central controller;

[0008] The central controller calculates the expected system power generation efficiency through the system power generation efficiency prediction model according to the internal environment data of the proton exchange membrane fuel cell, and then calculates the optimal values ​​of the working parameters of the proton exchange membrane fuel cell with the goal of achieving the expected system power generation efficiency, and sends the optimal values ​​of the working parameters to the environment controller;

[0009] The environmental controller controls the operation of the heater and humidifier according to the optimal values ​​of the working parameters to adjust the working environment temperature and cathode humidity.

[0010] The system power generation efficiency prediction model adopts a linear regression model and is modeled based on historical internal environmental data and system power generation efficiency.

[0011] The optimal values ​​of the operating parameters of the proton exchange membrane fuel cell are calculated, and the gradient descent whale intelligent optimization algorithm is used to search for the optimal values.

[0012] The internal environment data of the proton exchange membrane fuel cell includes the working environment temperature and cathode humidity of the proton exchange membrane fuel cell.

[0013] The present invention proposes a power generation control method applied to the PEM-ORC flexible combined intelligent power generation system, which is characterized by: according to the internal environmental data of the proton exchange membrane fuel cell, the expected system power generation efficiency is calculated through a system power generation efficiency prediction model, and then the optimal values ​​of the working parameters of the proton exchange membrane fuel cell are calculated with the goal of achieving the expected system power generation efficiency, and the power generation of the proton exchange membrane fuel cell is controlled according to the optimal values ​​of the working parameters.

[0014] The system power generation efficiency prediction model adopts a linear regression model and is modeled based on historical internal environmental data and system power generation efficiency, including the following steps:

[0015] Step 1, obtaining internal environmental history data of the proton exchange membrane fuel cell and corresponding system power generation efficiency data;

[0016] Step 2, data preprocessing;

[0017] Step 3, select characteristic variables related to system power generation efficiency as independent variables of the model;

[0018] Step 4: Establish a linear regression model with the formula:

[0019] y=a+b1x1+b2x2+b n x n

[0020] Where y represents the system power generation efficiency, x1, x2, ... x n Represents the selected characteristic variables, a and b i (i=1,2,...,n) represents the coefficient in the linear regression model;

[0021] Step 5: According to the internal environment historical data and the corresponding system power generation efficiency data, the coefficients in the linear regression model are fitted and solved to obtain the system power generation efficiency prediction model.

[0022] The optimal value search by the whale intelligent optimization algorithm of gradient descent includes the following steps: Solving the coefficients in the linear regression model by the least squares method.

[0023] Calculate the optimal values ​​of the operating parameters of the proton exchange membrane fuel cell and use the gradient descent whale intelligent optimization algorithm to search for the optimal value;

[0024] In the whale intelligent optimization algorithm of gradient descent, the position iteration formula is:

[0025]

[0026] Among them, γ is the learning rate, x k+1 is the whale position at the k+1th moment, x k is the whale position at time K, represents the device parameter value at the next moment and the current moment respectively; f(·) is the iterative function of position update.

[0027] In the gradient descent whale intelligent optimization algorithm, the objective function is:

[0028]

[0029]

[0030]

[0031]

[0032] η is the expected value of the system power generation efficiency output by the system power generation efficiency prediction model, h0 and S0 are the initial temperature and humidity values ​​when the fuel cell is working, S and h are the humidity and temperature at a certain moment respectively; t and T are the working time of the heater and humidifier respectively, f r and f s are the total power consumption of the heater and humidifier, respectively, r and p s is the rated power of the heater and humidifier, α and β are the working coefficients of the heater and humidifier respectively; f d is the total power generated by the battery itself, τ is the working time of the fuel cell; x0 and x are the starting time and the working end time; p f The power generated by the battery.

[0033] The optimal value search through the whale intelligent optimization algorithm of gradient descent includes the following steps:

[0034] (2.1) Initialization parameters;

[0035] (2.2) Calculating the fitness of each whale corresponds to calculating the expected value of the battery operating parameters that the system can currently achieve at a certain moment, which corresponds to the expected value of the system's power generation efficiency. Find the position of the current optimal whale and keep it, and then enter the prey search phase;

[0036] (2.3) In the target encirclement phase, the calculation formula for updating the whale position is:

[0037]

[0038]

[0039] Among them, represents the position of the $i$-th individual at the $(t + 1)$-th iteration, is the position of the whale individual that obtains the global optimal solution within the population when searching to the $t$-th generation, represents the position of the $i$-th individual at the $t$-th iteration; $D_1$ represents the surrounding step size, and $A$ and $C$ are coefficient vectors;

[0040] (2.4) Bubble-net attack, the calculation formula for updating the whale position is:

[0041]

[0042] Among them, $D_2$ represents the distance between the whale and the prey, $b$ is a constant controlling the logarithmic spiral shape, and $l$ is a random number on $[-1, 1]$;

[0043] (2.5) Enter the stage of searching for prey again. The whale randomly selects an individual from the current population as the target for position update, and the position in step (2.4) is preferentially retained, aiming to increase the search range; the calculation formula for updating the whale position is:

[0044]

[0045]

[0046] Among them, is the position of a randomly selected individual from the current population;

[0047] (2.6) In the predation stage, the calculation formula for updating the whale position depends on $p$, where $p$ is a random number between $[0, 1]$. When $p\geq W$, the calculation formula in step (2.4) is used for position update. When $p < W$, it is updated according to the value of $A$. When $A\leq1$, the calculation formula in step (2.3) is used for position update; otherwise, the calculation formula in step (2.5) is used for update; $W$ is a constant less than 1;

[0048] (2.7) Repeat steps (2.3)-(2.6). When the preset termination condition is reached, the optimal solution is output.

[0049] Beneficial effects: Compared with the prior art, the present invention has the following advantages: The present invention combines a proton exchange membrane fuel cell and an organic Rankine cycle system, uses a proton exchange membrane fuel cell, utilizes hydrogen and oxygen to generate electricity, and inputs the electricity of the proton exchange membrane fuel cell into the generator of the ORC cycle system. The waste heat in the process of power generation of the fuel cell is recovered through the ORC cycle system, thereby improving the overall power generation efficiency of the power generation system. The environmental parameters are precisely controlled by the gradient descent whale intelligent optimization algorithm, so that the power generation efficiency of the entire system is optimized, the equipment efficiency is effectively improved, and the loss of unnecessary energy is reduced. Combined with the gradient descent whale intelligent optimization algorithm and the linear regression model, the real-time data monitoring of the sensor module and the intelligent control of the central control module ensure the accuracy of environmental monitoring and the consistency of the control manager's actions, avoiding fluctuations caused by human factors. At the same time, the system can act in a tendency according to the current different environmental conditions, and has strong adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural schematic diagram of the PEM-ORC flexible combined intelligent power generation system of the present invention;

[0051] Figure 2 It is a schematic diagram of the working process of the intelligent power generation system of the present invention;

[0052] Figure 3 It is a flow chart of the gradient descent whale intelligent optimization algorithm described in the present invention. DETAILED DESCRIPTION

[0053] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0054] Example 1

[0055] The PEM-ORC flexible combined intelligent power generation system of the present invention has a structural schematic diagram as shown in FIG. Figure 1 The system includes a proton exchange membrane fuel cell, an ORC power generation system, a central controller, an environmental controller, and a parameter feedback device. The work flow diagram of the system is shown in Figure 2 shown.

[0056] Proton exchange membrane fuel cells use hydrogen and oxygen to generate electricity, and the electricity generated is used to power the ORC cycle power generation system, thereby improving the power generation efficiency of the entire system. The ORC power generation system recovers the low-grade waste heat from the fuel cell power generation process. A large amount of waste water and water vapor is generated during the operation of the fuel cell. These hot waste water and water vapor are collected and recycled through the waste heat recovery heat exchanger.

[0057] The central controller calculates the optimal values ​​of the fuel cell operating parameters through the linear regression model and the whale intelligent optimization algorithm of gradient descent. The central controller sends the calculated optimal values ​​of the fuel cell operating parameters to the environmental controller, which adjusts the fuel cell operating parameters through the environmental controller to ensure that the working environment of the fuel cell remains optimal, further maximizing the final power generation.

[0058] In this embodiment, the proton exchange membrane of the fuel cell is mainly affected by two main factors that affect the power generation efficiency: temperature and cathode humidity. The parameters of the fuel cell operation are set to the working environment temperature and cathode humidity. The parameter feedback device collects data on the temperature and cathode humidity parameter values ​​of the real-time working environment of the battery and transmits them to the central control unit. The central control unit analyzes the data by establishing a mathematical model of the system power generation efficiency and a gradient descent whale intelligent optimization algorithm to calculate the specific parameters of the device with the goal of achieving the expected system power generation efficiency, and adjusts the environmental control unit accordingly for reasonable control. The environmental controller controls the actions of the heater and humidifier respectively to adjust the working environment temperature and cathode humidity. When the environmental controller adjusts the equipment, it can also provide parameter feedback to improve the precision and accuracy of the equipment, thereby reducing the errors and delays caused by the equipment itself, data delays, and robustness, so as to meet the expected requirements and optimize the working efficiency.

[0059] Example 2

[0060] The PEM-ORC flexible combined intelligent power generation method of the present invention is applied to the above-mentioned PEM-ORC flexible combined intelligent power generation system. The method comprises: firstly analyzing data through a linear regression model and predicting the expected working efficiency that can be achieved, that is, the system power generation efficiency; then calculating the parameters corresponding to the heater and humidifier in the environmental controller with the best efficiency through the SGD-WOA gradient descent whale intelligent optimization algorithm, and the environmental controller further controls the actions of the heater and humidifier respectively, so that the system power generation efficiency can be optimized.

[0061] Taking the establishment of a linear regression model as an example, the specific formula and steps are as follows:

[0062] Step 1, collect historical data: collect historical environmental variable data related to the prediction target, such as temperature, cathode humidity, and corresponding system power generation efficiency data.

[0063] Step 2, data preprocessing: preprocess the collected data, including processing missing values, detecting and processing outliers, normalizing or standardizing the data, etc., for better use.

[0064] Step 3, feature variable selection: select the feature variables most relevant to the prediction target based on domain knowledge or statistical analysis methods. In this embodiment, temperature and cathode humidity are used as independent variables in the model.

[0065] Step 4: Establish a linear regression model: In the linear regression model, the target variable is predicted by linearly combining the feature variables. The model formula is:

[0066] y=a+b1x1+b2x2+b n x n (1)

[0067] Where y represents the predicted target variable, i.e., the system power generation efficiency, x1, x2, ...x n represents each selected characteristic variable, namely temperature and cathode humidity, and a and bi (i=1, 2, ..., n) represent the coefficients in the linear regression model.

[0068] Step 5, model training: Use the collected historical data to solve the coefficients in the model through the least squares method, that is, determine the values ​​of a and bi so that the model can fit the historical data as closely as possible.

[0069] Step 6, model evaluation and optimization: Use methods such as cross-validation to evaluate the model and see how well it predicts new data. Based on the evaluation results, you can adjust the model parameters or feature variable selection to further optimize the model and improve prediction accuracy.

[0070] Step 7, model application: The evaluated and optimized model can be used to predict unknown data, process new environmental variable data using the determined coefficients, and predict the corresponding system power generation efficiency. When establishing and using a linear regression model, it is necessary to make appropriate adjustments and improvements based on specific problems and data conditions.

[0071] The gradient descent whale intelligent optimization algorithm uses the gradient descent algorithm to minimize the objective function f(x). The gradient descent algorithm iterates in the opposite direction of the gradient vector to obtain the extreme point of the function. Parameter iteration formula:

[0072]

[0073] Among them, γ is the learning rate, and the initial value of the parameter is x0, which is the starting position. Set the iteration stop condition: the gradient is less than the set value or the maximum number of iterations is reached.

[0074] Among them, the objective function is:

[0075]

[0076] η is the expected value of the system power generation efficiency output by the linear regression model, h0 and S0 are the initial temperature and humidity values ​​of the heater and humidifier when the battery is initially working, S and h are the parameter values ​​at a certain moment, t and T are the working time of the heater and humidifier, respectively, and f r and f s are the total power consumption of the corresponding work, p r and p s is the rated power of the heater and humidifier, α and β are the working coefficients of the equipment respectively; f d is the total power generated by the battery itself, τ is the working time of the fuel cell battery; x0 and x are the starting time and the working end time; p f The power generated by the battery.

[0077] The flowchart of the gradient descent whale intelligent optimization algorithm is as follows Figure 3 As shown, the specific steps are:

[0078] (2.1) Initialize parameters. Initialize the whale position according to the acquired environmental data. Initialize the initial parameter values ​​of the equipment operation, including the initial temperature and humidity values. Set the number of whales N and the maximum number of iterations of the algorithm T. max , initialize the location information.

[0079] (2.2) Calculating the fitness of each whale corresponds to calculating the expected value of the equipment parameters corresponding to the expected value of the system power generation efficiency that the system can currently achieve at a certain moment, finding the position of the current optimal whale and retaining it, and then entering the prey search phase. WOA has three phases: searching for prey, surrounding targets, and spiral bubble net predation strategy. The performance of the algorithm depends on the balance between the global exploration phase and the local development phase. The central control system optimizes data through modeling and intelligent whale algorithms, so that it can search globally and obtain the optimal spatial solution vector. In WOA, these three strategies are used to achieve a dynamic balance between the exploration and development processes.

[0080] (2.3) Surrounding the target stage. The whales identify the position of the prey and surround them. The central control device starts to take corresponding actions according to the current situation. Through the data collection of the sensors and the data stored in the ROM, the dynamic data parameters of the optimal spatial solution are obtained through continuous calculation, and then the heater and humidifier can be reasonably adjusted. Since the optimal solution is an unknown spatial solution vector, WOA assumes that the optimal whale position in the current population is the target prey position, and approaches the prey in different ways, that is, adjusting the temperature and humidity parameter values ​​up and down, and then enters the stage of surrounding the prey. The other whale individuals in the group use formula (4) to update their positions and approach the optimal individual, corresponding to the current state of the fuel cell, and approaching the prey setting parameters, so as to achieve the state of optimal power generation efficiency.

[0081]

[0082]

[0083] in, is the position of the individual whale that obtains the global optimal solution in the population when searching to the tth generation, It represents the position of the i-th individual at the t-th iteration, D1 represents the enclosing step length, A and C represent the coefficient vectors, and the calculation formula is:

[0084]

[0085] Where T is the maximum number of iterations and r is a random number between [0,1].

[0086] (2.4) Bubble net attack, the whale spits out a bubble net along the spiral line to approach the prey for hunting, corresponding to the reasonable adjustment of the temperature and cathode humidity of the power generation conditions through the central control device and the environmental regulator, and accurately adjusts the data up and down, so as to maximize the power generation efficiency. The following formula is used for positioning update:

[0087]

[0088] in, represents the distance between the whale and its prey, b is a constant that controls the shape of the logarithmic spiral, and l is a random number on [-1,1].

[0089] (2.5) Entering the prey search phase again, the whale no longer selects the current optimal individual as the target for position update, but randomly selects an individual from the current population as the target for position update. The purpose is to increase the search range, find the optimal solution while maintaining population diversity, and continuously adjust the parameter value again in accordance with the optimal environment that the device can achieve in the actual working environment, so as to get closer to the expected value. The secondary search range ensures the accuracy and optimality of the data. The best of the two is retained, so only a small change in (1) and (2) is needed to obtain the mathematical model of this stage:

[0090]

[0091]

[0092] in, is the position of an individual randomly selected from the current population. When 0<A<1, (X, Y) is equal to the current random position (X * , Y * ) between possible positions.

[0093] (2.6) Predation phase. The whale swims toward the spiral line. The update or optimization of the whale's position depends on p, which is a random number between [0,1]. When p≥0.5, the position is updated using formula (7). When p<0.5, the position is updated according to the value of A. When A≤1, the position is updated using formula (4); otherwise, it is updated using (8). That is, the environmental regulator, including the heater and humidifier, acts according to the optimal path determined by the model prediction of the central control module and the regulation of the intelligent algorithm, continuously updates the current state, and then feeds back and updates, greatly improving the accuracy of the device action.

[0094] (2.7) Repeat steps (2.3)-(2.6) and output the optimal solution when the preset termination condition is reached.

Claims

1. A PEM-ORC flexible combined intelligent power generation system, characterized in that: include: Proton exchange membrane fuel cells, which use hydrogen and oxygen to generate electricity and power the ORC system; The ORC system uses low-grade waste heat to generate electricity. The low-grade waste heat includes hot waste water and hot water vapor generated during the power generation process of the proton exchange membrane fuel cell, and the waste heat is recovered through the heat exchanger in the ORC system; A parameter feedback device for collecting internal environmental data of the proton exchange membrane fuel cell and sending it to a central controller; The central controller calculates the expected system power generation efficiency through the system power generation efficiency prediction model according to the internal environment data of the proton exchange membrane fuel cell, and then calculates the optimal values ​​of the working parameters of the proton exchange membrane fuel cell with the goal of achieving the expected system power generation efficiency, and sends the optimal values ​​of the working parameters to the environment controller; The environmental controller controls the operation of the heater and humidifier according to the optimal values ​​of the working parameters to adjust the working environment temperature and cathode humidity; The system power generation efficiency prediction model adopts a linear regression model and is modeled based on historical internal environmental data and system power generation efficiency, and includes the following steps: Step 1, obtaining internal environmental historical data of the proton exchange membrane fuel cell and corresponding system power generation efficiency data; Step 2, data preprocessing; Step 3, select characteristic variables related to system power generation efficiency as independent variables of the model; Step 4: Establish a linear regression model with the formula: y=a+b1x1+b2x2+b n x n Where y represents the system power generation efficiency, x1, x2, ... x n Represents the selected characteristic variables, a and b i (i=1,2,...,n) represents the coefficient in the linear regression model; Step 5: According to the internal environment historical data and the corresponding system power generation efficiency data, the coefficients in the linear regression model are fitted and solved to obtain the system power generation efficiency prediction model.

2. The PEM-ORC flexible combined intelligent power generation system according to claim 1 is characterized in that: The optimal values ​​of the operating parameters of the proton exchange membrane fuel cell are calculated, and the gradient descent whale intelligent optimization algorithm is used to search for the optimal values.

3. The PEM-ORC flexible combined intelligent power generation system according to claim 1 is characterized in that: The internal environment data of the proton exchange membrane fuel cell includes the working environment temperature and cathode humidity of the proton exchange membrane fuel cell.

4. A power generation control method applied to the PEM-ORC flexible combined intelligent power generation system according to claim 1, characterized in that: Based on the internal environmental data of the proton exchange membrane fuel cell, the expected system power generation efficiency is calculated through the system power generation efficiency prediction model, and then the optimal values ​​of the working parameters of the proton exchange membrane fuel cell are calculated with the goal of achieving the expected system power generation efficiency. The power generation of the proton exchange membrane fuel cell is controlled according to the optimal values ​​of the working parameters.

5. The power generation control method according to claim 4, characterized in that: The system power generation efficiency prediction model adopts a linear regression model and is modeled based on historical internal environmental data and system power generation efficiency, including the following steps: Step 1, obtaining internal environmental history data of the proton exchange membrane fuel cell and corresponding system power generation efficiency data; Step 2, data preprocessing; Step 3, select characteristic variables related to system power generation efficiency as independent variables of the model; Step 4: Establish a linear regression model with the formula: y=a+b1x1+b2x2+b n x n Where y represents the system power generation efficiency, x1, x2, ... x n Represents the selected characteristic variables, a and b i (i=1,2,...,n) represents the coefficient in the linear regression model; Step 5: According to the internal environment historical data and the corresponding system power generation efficiency data, the coefficients in the linear regression model are fitted and solved to obtain the system power generation efficiency prediction model.

6. The power generation control method according to claim 5, characterized in that: Solve for the coefficients in the linear regression model using the least squares method.

7. The power generation control method according to claim 4, characterized in that: Calculate the optimal values ​​of the operating parameters of the proton exchange membrane fuel cell and use the gradient descent whale intelligent optimization algorithm to search for the optimal value; In the whale intelligent optimization algorithm of gradient descent, the position iteration formula is: Among them, γ is the learning rate, x k+1 is the whale position at the k+1th moment, x k is the whale position at time K, represents the device parameter value at the next moment and the current moment respectively; f(·) is the iterative function of position update.

8. The power generation control method according to claim 7, characterized in that: In the gradient descent whale intelligent optimization algorithm, the objective function is: η is the expected value of the system power generation efficiency output by the system power generation efficiency prediction model, h0 and S0 are the initial temperature and humidity values ​​when the fuel cell is working, S and h are the humidity and temperature at a certain moment respectively; t and T are the working time of the heater and humidifier respectively, f r and f s are the total power consumption of the heater and humidifier, respectively, r and p s is the rated power of the heater and humidifier, α and β are the working coefficients of the heater and humidifier respectively; f d is the total power generated by the battery itself, τ is the working time of the fuel cell; x0 and x are the starting time and the working end time; p f The power generated by the battery.

9. The power generation control method according to claim 7, characterized in that: The optimal value search by the whale intelligent optimization algorithm through gradient descent includes the following steps: (2.1) Initialize parameters; (2.2) Calculate the fitness of each whale, which corresponds to calculating the predicted battery operating parameters corresponding to the closest value to the expected power generation efficiency of the system at a certain moment. Find the position of the current optimal whale and retain it, and then enter the stage of searching for prey; (2.3) Encircling the target stage, the calculation formula for updating the whale position is: in, represents the position of the i-th individual at the t+1th iteration, is the position of the individual whale that obtains the global optimal solution in the population when searching to the tth generation, represents the position of the i-th individual at the t-th iteration; D1 represents the encircling step length, A and C are coefficient vectors; (2.4) Bubble net attack, the calculation formula for updating the whale position is: Among them, D2 represents the distance between the whale and the prey, b is a constant controlling the logarithmic spiral shape, and l is a random number on [-1, 1]; (2.5) Enter the stage of searching for prey again. The whale randomly selects an individual from the current population as the target for position update, and the position is preferentially retained with that in step (2.4) to increase the search range. The calculation formula for updating the whale position is: in, is the position of an individual randomly selected from the current population; (2.6) Predation stage, the calculation formula for updating the whale position depends on p. p is a random number between [0, 1]. When p≥W, the calculation formula in step (2.4) is used for position update. When p<W, it is updated according to the value of A. When |A|≤1, the calculation formula in step (2.3) is used for position update; otherwise, the calculation formula in step (2.5) is used for update; W is a constant less than 1; (2.7) Repeat steps (2.3)-(2.6). When the preset termination condition is reached, output the optimal solution.