Power distribution method and device of multi-machine fuel cell power generation system

Abstract

One or more embodiments of the invention provide a power distribution method and device for a multi-machine fuel cell power generation system. The method comprises the following steps: acquiring a corresponding relation between output power and efficiency of each fuel cell in a system; based on the corresponding relation of each fuel cell, constructing a target function by taking maximization of the overall efficiency of the system as a target; and performing iterative optimization on the objective function according to the load power of the system to obtain the optimal output power required to be borne by each fuel cell. According to the method, part of units can be prevented from running in a low-efficiency interval for a long time, the overall efficiency of the system is remarkably improved, dynamic distribution is achieved through iterative optimization, the system automatically adapts to load changes, it is ensured that the system always runs at the optimal efficiency point, and the running cost is effectively reduced.

Description

Technical Field

[0001] One or more embodiments of the present specification relate to the field of fuel cell technology, and in particular, to a power distribution method and device for a multi-fuel cell power generation system. Background Art

[0002] In fuel cell engineering applications, the power of a single unit is often insufficient to meet medium to high load requirements, necessitating the use of multiple fuel cells connected in parallel. Compared to a single-unit system, a multi-unit system requires a more balanced power distribution based on the characteristics of each unit and the load requirements to improve overall efficiency.

[0003] The relevant power allocation strategy mainly adopts average allocation or sequential allocation. This allocation method does not take into account the actual usage of each fuel cell, which can easily cause some units to operate in a low-efficiency range for a long time, making the overall efficiency of the system low. Summary of the Invention

[0004] In view of this, one or more embodiments of this specification provide the following technical solutions:

[0005] According to a first aspect of one or more embodiments of this specification, a power distribution method for a multi-fuel cell power generation system is provided, the method comprising:

[0006] Obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system;

[0007] Based on the corresponding relationship between each fuel cell, an objective function is constructed with the goal of maximizing the overall efficiency of the system;

[0008] According to the load power of the system, the objective function is iteratively optimized to obtain the optimal output power that each fuel cell needs to carry.

[0009] According to a second aspect of one or more embodiments of this specification, a power distribution device for a multi-fuel cell power generation system is provided, comprising:

[0010] An information processing module is used to obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system;

[0011] An optimization problem building module is used to construct an objective function based on the corresponding relationship between each fuel cell to maximize the overall efficiency of the system;

[0012] The iterative optimization module is used to iteratively optimize the objective function according to the load power of the system to obtain the optimal output power that each fuel cell needs to carry.

[0013] According to a third aspect of one or more embodiments of this specification, an electronic device is proposed, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor implements the steps of the method described in the first aspect by running the executable instructions.

[0014] According to a fourth aspect of one or more embodiments of this specification, a computer-readable storage medium is provided, on which computer instructions are stored. When the instructions are executed by a processor, the steps of the method described in the first aspect are implemented.

[0015] According to a fifth aspect of one or more embodiments of this specification, a computer program product is proposed, comprising a computer program / instruction, which implements the steps of the method described in the first aspect when executed by a processor.

[0016] As can be seen from the above examples, this specification, by obtaining the corresponding relationship between the output power and efficiency of each fuel cell and constructing a function that maximizes overall system efficiency, can specifically allocate power based on the aging and efficiency characteristics of different units. Compared to traditional average or sequential allocation strategies, this method can prevent some units from operating in low-efficiency ranges for a long time, significantly improving overall system efficiency. Furthermore, dynamic allocation through iterative optimization enables the system to automatically adapt to load changes, ensuring consistent operation at the optimal efficiency point, effectively reducing operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the architecture of a multi-fuel cell power generation system provided by an exemplary embodiment.

[0018] Figure 2 This is one of the flow charts of a power distribution method for a multi-machine fuel cell power generation system provided by an exemplary embodiment.

[0019] Figure 3 This is one of the schematic diagrams of an efficiency curve provided by an exemplary embodiment.

[0020] Figure 4 This is a second schematic diagram of an efficiency curve provided by an exemplary embodiment.

[0021] Figure 5 This is the second flowchart of a power distribution method for a multi-fuel cell power generation system provided by an exemplary embodiment.

[0022] Figure 6 It is a structural diagram of an electronic device provided by an exemplary embodiment.

[0023] Figure 7 A block diagram of a power distribution device for a multi-fuel cell power generation system provided by an exemplary embodiment. DETAILED DESCRIPTION

[0024] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this manual are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0025] In practical fuel cell applications, the power of a single fuel cell is often insufficient to meet medium to high load requirements, necessitating the parallel connection of multiple fuel cells to collectively provide the required power. Compared to a single fuel cell, a multi-unit system requires a rational allocation of power based on the characteristics of each fuel cell and current demand to achieve higher overall efficiency.

[0026] When allocating power to a multi-unit fuel cell power generation system, simple strategies such as average or sequential allocation are often used. While these methods are easy to implement, they fail to fully consider the aging of different fuel cells, differences in their efficiency curves, and the impact of each unit's output power on the overall system efficiency under various loads. Such simple allocation methods often cause some fuel cells to operate in low-efficiency ranges, resulting in lower overall efficiency. Furthermore, as fuel cells degrade over time, performance differences between units gradually increase, and fixed or single allocation strategies further degrade system efficiency.

[0027] In view of this, this specification proposes a power allocation method for a multi-fuel cell power generation system. Based on the efficiency of each fuel cell during actual use, an objective function is constructed to maximize the overall efficiency of the system and solve the optimal power allocation method.

[0028] During implementation, the corresponding relationship between the output power and efficiency of each fuel cell in the system is obtained; based on the efficiency curve of each fuel cell, an objective function is constructed with the goal of maximizing the overall efficiency of the system; according to the load power of the system, the objective function is iteratively optimized to obtain the optimal output power that each fuel cell needs to carry.

[0029] In this technical solution, by obtaining the corresponding relationship between the output power and efficiency of each fuel cell and constructing a function that maximizes the overall efficiency of the system, power allocation can be tailored to the aging and efficiency characteristics of different units. Compared with traditional average or sequential allocation strategies, this method can prevent some units from operating in low-efficiency ranges for a long time, significantly improving overall system efficiency. Furthermore, dynamic allocation through iterative optimization enables the system to automatically adapt to load changes, ensuring consistent operation at the optimal efficiency point and effectively reducing operating costs.

[0030] Figure 1 This is a schematic diagram of a multi-fuel cell power generation system architecture provided by an exemplary embodiment. Figure 1 As shown, there are n fuel cells running in parallel in the system. Each fuel cell can be connected to the DC bus through a DC converter (DC-DC converter), and the DC bus can be directly connected to the load, such as Figure 1 As shown, the grid is connected via an inverter. The power distribution control module can calculate the load power required by each fuel cell in real time and instruct the DC converter corresponding to each fuel cell to control the output power of each fuel cell.

[0031] The fuel used by the fuel cell can be hydrogen, methanol, natural gas, etc., and this application does not make any specific restrictions.

[0032] In order to enable people skilled in the art to better understand the technical solutions in this application, the technical solutions in this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this application.

[0033] See Figure 2 , Figure 2 An exemplary embodiment provides a power distribution method for a multi-fuel cell power generation system. The method can be executed by a power distribution control module of the system and specifically includes the following steps.

[0034] S201 : Obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system.

[0035] In one embodiment, the corresponding relationship between the output power and efficiency of each fuel cell in the system can be obtained, and the efficiency of the fuel cell at different output powers can be obtained based on the corresponding relationship.

[0036] For example, a specific correspondence can be pre-set for each fuel cell;

[0037] For example, the corresponding relationship between the output power and the efficiency can be determined based on the operating state of the fuel cell; wherein the operating state may include: normal, aged, severely aged, etc.; or high efficiency, medium, low efficiency, etc.;

[0038] For example, the corresponding relationship between output power and efficiency can be generated based on historical operating data of the fuel cell.

[0039] The corresponding relationship between output power and efficiency can be specifically expressed as a mapping table, an efficiency function, an efficiency curve, etc.

[0040] The efficiency of a fuel cell can specifically indicate the amount of output power that the fuel cell can generate per unit of fuel consumed, for example, Figure 1As shown, it can be the ratio of the output power of the DC converter corresponding to the fuel cell to the heat that can be generated by the fuel consumed by the fuel cell.

[0041] In one embodiment, the efficiency of the fuel cell is fc It can be composed of the following three parts: Stack efficiency η stack , fuel (such as hydrogen, methanol, natural gas, etc.) utilization rate η fuel , efficiency η of other auxiliary equipment (such as DC converter, gas control system, etc.) aux , where the specific calculation formula can be expressed as follows:

[0042] η fc =η stack ×η fuel ×η aux

[0043] Among them, the stack efficiency η stack It can be the ratio of the electrical energy generated by the fuel stack per unit time to the heat generated by the consumed fuel.

[0044] S202 : Based on the corresponding relationship between the fuel cells, an objective function is constructed with the goal of maximizing the overall efficiency of the system.

[0045] Based on the obtained correspondence between the output power and efficiency of each fuel cell, the optimization problem can be constructed with the goal of maximizing the overall efficiency of the system. The output power of the i-th fuel cell can be recorded as P i , efficiency is denoted as η i , the overall efficiency of the system is recorded as η, and the constructed optimization problem can be expressed as follows:

[0046]

[0047] The optimization problem includes the objective function to maximize the overall efficiency η of the system. And the optimization variable P i Constraints; where P is the load power; η i (P i ) represents the output power P of the i-th fuel cell i The efficiency under the condition is expressed as P imax represents the maximum output power of the i-th fuel cell.

[0048] S203 : Iteratively optimize the objective function according to the load power of the system to obtain the optimal output power that each fuel cell needs to carry.

[0049] Based on the constructed optimization problem, the objective function of the optimization problem can be iteratively optimized according to the system load power P obtained in real time, and the optimal output power that each combustion cell needs to carry while maximizing the overall efficiency of the system can be calculated. This optimal output power needs to meet the constraints of the optimization problem.

[0050] In this technical solution, by obtaining the corresponding relationship between the output power and efficiency of each fuel cell and constructing a function that maximizes the overall efficiency of the system, power allocation can be tailored to the aging and efficiency characteristics of different units. Compared with traditional average or sequential allocation strategies, this method can prevent some units from operating in low-efficiency ranges for a long time, significantly improving overall system efficiency. Furthermore, dynamic allocation through iterative optimization enables the system to automatically adapt to load changes, ensuring consistent operation at the optimal efficiency point and effectively reducing operating costs.

[0051] In one embodiment, the corresponding relationship between the output power and efficiency of each fuel cell can be represented by an efficiency curve.

[0052] The output power P of each fuel cell in the system can be obtained i The sampling data under different output power P is used to calculate the performance of each fuel cell. i The efficiency η under i The sampling data can be obtained through early experimental testing or can be historical data during actual operation. It can cover sampling data within the output power range of the fuel cell.

[0053] Then, based on the sampled data, the initial efficiency curve of each fuel cell is generated. For example, the efficiency curve η can be generated by spline interpolation, polynomial fitting, Gaussian process regression or neural network methods. i (P i ).like Figure 3 As shown, the efficiency curves of three fuel cells (fuel cell 1, 2, 3) are shown. It can be seen from the efficiency curves that Figure 3 Among the three fuel cells shown, fuel cell 1 has the best efficiency and can be considered to be in a normal working state, fuel cell 2 has a relatively poor efficiency and can be considered to be in an aging working state, and fuel cell 3 has the worst efficiency and can be considered to be in a severely aging working state.

[0054] After obtaining the initial efficiency curve of each fuel cell, the efficiency curve that exceeds the maximum output power of each fuel cell can be further smoothed and extended to make the efficiency curve drop rapidly to near zero in the part exceeding the maximum output power, thereby eliminating the constraint of the maximum output power on the efficiency curve, that is, eliminating the constraint condition P in the optimization problem expressed by the above formula (1) that the output power of each fuel cell does not exceed the maximum output power. i≤P imax .

[0055] In the above embodiment, initial efficiency curves are generated from sampled data and processed in conjunction with maximum output power constraints to accurately characterize the true efficiency performance of each fuel cell within a safe power range. By extending the curves beyond the maximum power range, efficiency values ​​drop significantly when power allocation approaches or exceeds the limit. This mitigates the risk of power overruns through the guidance of the objective function without requiring additional constraints. This simplifies the complexity of the optimization problem, ensures that the power allocation results comply with physical constraints, and enhances the flexibility and stability of the optimization process.

[0056] The method for smoothing and extending the efficiency curve beyond the maximum output power range can be set according to actual needs. For example, exponential decay extension, polynomial extension, piecewise function extension, Gaussian filtering, etc. can be used.

[0057] In one embodiment, after obtaining the initial efficiency curves of each fuel cell, Gaussian filtering can be applied to the efficiency curves that exceed the maximum output power range to perform extension processing, so that the efficiency values ​​after extension processing decay smoothly toward zero as the power increases. Figure 3 The three efficiency curves shown in the figure are extended by Gaussian filtering to obtain the following: Figure 4 The three extended efficiency curves are shown.

[0058] In the above embodiment, Gaussian filtering is used to smoothly extend the efficiency curve beyond the power range, causing the efficiency value to decay smoothly to near zero as power increases, avoiding sudden changes in the curve or gradient instability. This process not only ensures the continuity and differentiability of the objective function during the optimization process, facilitating subsequent gradient calculation and iterative solution, but also effectively prevents power allocation from exceeding the maximum allowable value. It also ensures that the extended portion seamlessly connects with the original curve, maintaining the smoothness of the overall efficiency curve, thereby improving the algorithm's convergence speed and stability and avoiding deviations in optimization results caused by curve discontinuities.

[0059] In one embodiment, the constraints P1+P2+...+P in the optimization problem represented by formula (1) can be eliminated by using the Softmax function. n =P, 0≤P i .

[0060] The power distribution ratio σ of each fuel cell can be calculated by the Softmax function. i Convert to unconstrained optimization variable x i , where the optimization variable x iIt can be used as the decision variable of the optimization problem, which can be an n-dimensional variable. The output power of each fuel cell is represented by the product of the corresponding power distribution ratio and the load power, that is, σ i P instead of P i .

[0061] The Softmax function can be expressed as follows:

[0062]

[0063] The power allocation ratio σ obtained by Softmax function transformation i Obviously, the following characteristics are met:

[0064]

[0065] Based on the characteristics of the Softmax function, it is obvious that the constraints that the sum of the output power of all fuel cells is equal to the load power and the output power is non-negative in the optimization problem expressed by formula (1) are eliminated.

[0066] After the above-mentioned smooth extension of the curve function and the Softmax function transformation, all the constraints of the above optimization problem can be eliminated, and the optimization problem can be converted into an unconstrained optimization problem, that is, the objective function can be converted into an unconstrained objective function, which can be expressed as follows:

[0067]

[0068] According to the obtained system load power, the unconstrained objective function (such as formula (4)) is iteratively optimized to obtain the optimal power distribution ratio of each fuel cell. Furthermore, based on the optimal power distribution ratio and the real-time load power, the optimal output power required by each fuel cell is obtained.

[0069] In the above embodiment, the use of a Softmax function to convert the power allocation ratio into an unconstrained optimization variable naturally satisfies the constraints that the total power equals the load power and that each power is non-negative, significantly reducing the complexity of the optimization problem. By converting the constrained objective function into an unconstrained form and combining iterative optimization to solve the power allocation ratio, the algorithm is applicable to any number of fuel cell systems, with a more efficient computation process and significantly improving the feasibility of online, real-time allocation.

[0070] In one embodiment, the efficiency curve shows that when the output power is large, the efficiency change is small. To make the difference more significant and make the objective function easier to solve, a loss function can be used to solve the optimal value, changing the optimization goal from maximizing the overall efficiency of the system to minimizing the efficiency loss. The objective function is converted as follows:

[0071]

[0072] Furthermore, the difference can be amplified by logarithmic processing, and the optimization goal can be converted to the logarithm of minimizing efficiency loss. The objective function after logarithmic processing is expressed as follows:

[0073]

[0074] Furthermore, a regularization term can be introduced into the objective function To avoid optimizing the variable x i The value of is too large or too small, which improves numerical stability and ensures that the solution vector is relatively concentrated to avoid extreme situations. The objective function can be expressed as follows:

[0075]

[0076] Among them, α is the regularization coefficient, which needs to be appropriately selected to ensure effective constraints while avoiding affecting the optimization results.

[0077] In one embodiment, in order to further improve the convergence speed during the iterative optimization of the unconstrained objective function, explicit gradient information of the unconstrained objective function with respect to the optimization variable can be derived. The specific gradient derivation process is exemplified as follows:

[0078] For ease of expression, Simply record it as Σ and perform gradient derivation.

[0079]

[0080]

[0081] δ ij Indicates that when i=j, δ ij =1; otherwise, δ ij =0.

[0082]

[0083] It is expressed as the following matrix form,

[0084]

[0085] Substitution As a result, we can get:

[0086]

[0087] Where I is an n×n dimensional identity matrix with only 1 on the diagonal, and σ is an n×1 dimensional column vector.

[0088] Based on the above results and adding the gradient of the regularization term, the gradient information of the objective function min ln loss with respect to the optimization variable x can be expressed as follows:

[0089]

[0090] Wherein, sign(x) represents the sign function.

[0091] After obtaining the unconstrained objective function and the derived gradient information, the unconstrained objective function can be iteratively optimized according to the system's load power, using the explicit gradient information as the search direction, to obtain the optimal power distribution ratio of each fuel cell; then, based on the optimal power distribution ratio and the load power, the optimal output power of each fuel cell can be calculated.

[0092] In the above embodiment, when gradient information is not provided, the solver generally uses a finite difference approximation, requiring gradient calculations at each iteration. For an n-dimensional function, each gradient calculation requires 2n+1 function calls. By directly providing explicit gradient information, however, each gradient calculation requires only one function call, significantly reducing computational overhead and providing higher precision. Each step of the direction update is more accurate, allowing for faster convergence within a smaller number of iterations.

[0093] The solution algorithm for iterative optimization of the unconstrained objective function can be called according to actual needs. For example, any first-order optimization algorithm can be used for iterative optimization.

[0094] For example, taking the quasi-Newton method as an example, after obtaining the unconstrained objective function and the derived gradient information, the unconstrained objective function can be iteratively optimized through the quasi-Newton algorithm according to the system's load power and the gradient information as the search direction to obtain the optimal power distribution ratio of each fuel cell; and then based on the optimal power distribution ratio and load power, the optimal output power of each fuel cell is calculated.

[0095] Figure 5 A schematic diagram of a power distribution method for a multi-fuel cell power generation system is shown in FIG. Figure 5 As shown, the method includes the following steps:

[0096] S501 : Input sampling data and load power P of each fuel cell.

[0097] The system can obtain sampling data related to the efficiency of each fuel cell at different output powers, which can reflect the corresponding relationship between the output power and efficiency of the fuel cell. In addition, the real-time load power P of the system can also be obtained.

[0098] S502 : Generate an efficiency curve based on the sampled data, and perform cubic spline interpolation on the efficiency curve to ensure the fitting accuracy of the efficiency curve in the local area and improve the overall smoothness.

[0099] S503 , using Gaussian filtering to perform smoothing and extension processing on the efficiency curves exceeding the maximum output power of each fuel cell, so that the efficiency value after extension decays rapidly toward zero as the power increases, thereby eliminating the limitation of the maximum output power.

[0100] S504: Establish an objective function with the goal of maximizing the overall efficiency of the system, and derive the gradient information of the objective function.

[0101] The unconstrained optimization variable x can be converted into a power allocation ratio through the Softmax function. By utilizing the characteristics of the Softmax function, an unconstrained objective function (such as formula (7)) is constructed with the goal of maximizing the overall efficiency of the system.

[0102] Based on the constructed unconstrained objective function, its gradient information about the optimization variable x is derived (such as formula (13)).

[0103] S505: Initialize the optimization variable x in the objective function.

[0104] S506 , iteratively optimize the objective function using the gradient information, calculate the gradient vector to determine the search direction and step size, and update the optimization variable x.

[0105] S507. Calculate the gradient norm corresponding to the gradient vector to measure the size of the gradient; compare the calculated gradient norm with the preset convergence threshold Tol; if the gradient norm is less than or equal to the convergence threshold, it means that the objective function has changed smoothly enough in the current iteration and the optimal solution can be obtained, and execute step S508; if the gradient norm is greater than the convergence threshold, it means that the ideal convergence state has not yet been reached, and continue iterative optimization and execute step S506.

[0106] S508. Output the optimal power distribution ratio and maximum overall system efficiency of each fuel cell, and combine it with the system load power to calculate the optimal output power of each fuel cell. The result is output and used for actual fuel cell power distribution control to maximize the overall system efficiency.

[0107] Figure 6 This is a schematic structural diagram of an electronic device provided by an exemplary embodiment. Figure 6At the hardware level, the device includes a processor 602, an internal bus 604, a network interface 606, a memory 608, and a non-volatile memory 610. Of course, it may also include hardware required for other functions. One or more embodiments of this specification can be implemented based on software, such as the processor 602 reading the corresponding computer program from the non-volatile memory 610 into the memory 608 and then running it. Of course, in addition to software implementation, one or more embodiments of this specification do not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0108] Please refer to Figure 7 The power distribution device of the multi-fuel cell power generation system can be applied to Figure 6 The device shown is used to implement the technical solution of this specification. The power distribution device of the multi-fuel cell power generation system may include: an information processing module 701, an optimization problem construction module 702, and an iterative optimization module 703. The information processing module 701 is used to obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system; the optimization problem construction module 702 is used to construct an objective function based on the corresponding relationship between each fuel cell, with the goal of maximizing the overall efficiency of the system; and the iterative optimization module 703 is used to iteratively optimize the objective function based on the load power of the system to obtain the optimal output power required by each fuel cell.

[0109] In one embodiment, the information processing module 701 is used to obtain sampling data of each fuel cell in the system at different output powers, as well as the maximum output power of each fuel cell; based on the sampling data, generate the initial efficiency curve of each fuel cell; and smooth and extend the efficiency curve that exceeds the maximum output power range to eliminate the constraint of the maximum output power on the efficiency curve.

[0110] In one embodiment, the information processing module 701 is configured to apply Gaussian filtering to extend the efficiency curve beyond the maximum output power range, so that the efficiency value after the extension process decays smoothly toward zero as the power increases.

[0111] In one embodiment, the optimization problem construction module 702 is also used to convert the power distribution ratio of each fuel cell into an unconstrained optimization variable through a Softmax function, and convert the objective function into an unconstrained objective function; the objective function is iteratively optimized according to the load power of the system to obtain the optimal output power that each fuel cell needs to carry, including: iteratively optimizing the unconstrained objective function according to the load power of the system to obtain the optimal power distribution ratio of each fuel cell; based on the optimal power distribution ratio and the real-time load power, the optimal output power that each fuel cell needs to carry is obtained.

[0112] In one embodiment, the iterative optimization module 703 is also used to derive the gradient information of the unconstrained objective function with respect to the optimization variable; based on the load power of the system, the unconstrained objective function is iteratively optimized with the gradient information as the search direction to obtain the optimal power distribution ratio of each fuel cell.

[0113] In one embodiment, the iterative optimization module 703 is used to iteratively optimize the unconstrained objective function using a quasi-Newton algorithm based on the load power of the system and with the explicit gradient information as the search direction to obtain the optimal power distribution ratio of each fuel cell.

[0114] In one embodiment, the unconstrained objective function is expressed as follows:

[0115]

[0116] Among them, σ i represents the power distribution ratio of the i-th fuel cell, P represents the load power, η i (σ i P) represents the efficiency value of the i-th fuel cell, x i Represents σ i The unconstrained optimization variables, represents the regularization term of the objective function, and α represents the regularization coefficient.

[0117] In one embodiment, the explicit gradient information is represented as follows:

[0118]

[0119] Among them, loss represents efficiency loss; ∑ represents sign(x) represents the sign function.

[0120] Based on the same concept as the above method, this specification also provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein the processor implements the steps of the method described in any of the above embodiments by running the executable instructions.

[0121] Based on the same concept as the above method, this specification also provides a computer-readable storage medium on which computer instructions are stored. When the instructions are executed by a processor, the steps of the method described in any of the above embodiments are implemented.

[0122] Based on the same concept as the above method, this specification also provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any of the above embodiments when executed by a processor.

Claims

1. A power distribution method for a multi-fuel cell power generation system, characterized in that: The method comprises: Obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system; Based on the corresponding relationship between each fuel cell, an objective function is constructed with the goal of maximizing the overall efficiency of the system; According to the load power of the system, the objective function is iteratively optimized to obtain the optimal output power that each fuel cell needs to carry.

2. The method according to claim 1, characterized in that The obtaining of the corresponding relationship between the output power and efficiency of each fuel cell in the system includes: Obtain sampling data of each fuel cell in the system at different output powers, as well as the maximum output power of each fuel cell; generating an initial efficiency curve of each fuel cell based on the sampled data; The efficiency curve exceeding the maximum output power range is smoothed and extended to eliminate the constraint of the maximum output power on the efficiency curve.

3. The method according to claim 2, characterized in that The step of smoothing and extending the efficiency curve beyond the maximum output power range to eliminate the constraint of the maximum output power on the efficiency curve includes: Gaussian filtering is applied to the efficiency curve exceeding the maximum output power range to perform extension processing, so that the efficiency value after the extension processing decays smoothly and tends to zero as the power increases.

4. The method according to claim 1, wherein After constructing the objective function, the method further includes: The power distribution ratio of each fuel cell is converted into an unconstrained optimization variable by a Softmax function, and the objective function is converted into an unconstrained objective function; The objective function is iteratively optimized according to the load power of the system to obtain the optimal output power required to be carried by each fuel cell, including: Iteratively optimizing the unconstrained objective function according to the load power of the system to obtain an optimal power distribution ratio for each fuel cell; Based on the optimal power distribution ratio and the load power, the optimal output power required to be carried by each fuel cell is obtained.

5. The method according to claim 4, characterized in that After converting to obtain the unconstrained objective function, the method further includes: Derivation of gradient information of the unconstrained objective function with respect to the optimization variable; The unconstrained objective function is iteratively optimized according to the load power of the system to obtain the optimal power distribution ratio of each fuel cell, including: According to the load power of the system and with the gradient information as the search direction, the unconstrained objective function is iteratively optimized to obtain the optimal power distribution ratio of each fuel cell.

6. The method according to claim 5, characterized in that The method of iteratively optimizing the unconstrained objective function based on the load power of the system and taking the gradient information as the search direction to obtain the optimal power distribution ratio of each fuel cell includes: According to the load power of the system and taking the gradient information as the search direction, the unconstrained objective function is iteratively optimized by a quasi-Newton algorithm to obtain the optimal power distribution ratio of each fuel cell.

7. The method according to claim 4, characterized in that The unconstrained objective function is expressed as follows: Among them, σ i represents the power distribution ratio of the i-th fuel cell, P represents the load power, η i (σ i P) represents the efficiency value of the i-th fuel cell, x i Represents σ i The unconstrained optimization variables, represents the regularization term of the objective function, and α represents the regularization coefficient.

8. The method according to claim 5, characterized in that The gradient information is expressed as follows: Among them, loss represents efficiency loss; ∑ represents sign(x) represents the sign function.

9. A power distribution device for a multi-fuel cell power generation system, characterized in that: include: An information processing module is used to obtain the corresponding relationship between the output power and efficiency of each fuel cell in the system; An optimization problem building module is used to construct an objective function based on the corresponding relationship between each fuel cell to maximize the overall efficiency of the system; The iterative optimization module is used to iteratively optimize the objective function according to the load power of the system to obtain the optimal output power that each fuel cell needs to carry.

10. An electronic device, characterized in that: include: processor; A memory for storing processor-executable instructions; wherein the processor implements the steps of the method according to any one of claims 1 to 8 by executing the executable instructions.

11. A computer-readable storage medium, characterized in that Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, characterized in that The method comprises a computer program / instruction, which implements the steps of the method according to any one of claims 1 to 8 when the computer program / instruction is executed by a processor.

Images