Water turbine runner optimization design method combining dogvessel optimization algorithm and BP neural network

Abstract

The invention discloses a water turbine runner optimization design method in combination with a doliolaria optimization algorithm and a BP neural network. The method comprises the following steps: S1, selecting runner geometric parameters influencing the performance of a water turbine; s2, obtaining a runner design scheme based on an orthogonal experiment; s3, structural accessories of rotating wheels of all the schemes are designed, and a three-dimensional overall model of the water turbine is constructed; s4, calculating the water head and efficiency of the water turbine in each scheme based on computational fluid mechanics; s5, constructing a water turbine efficiency characteristic neural network model and a water head characteristic neural network model which take the runner geometric parameters as input and take the water turbine efficiency and the water head as output respectively based on the BP neural network; and S6, obtaining optimal geometric parameters of the runner in combination with a doliolaria optimization algorithm, the water turbine efficiency characteristic neural network and the water head characteristic neural network model. The efficiency or the water head of the water turbine can be fully improved by reasonably designing the runner, and the power generation efficiency or the water head utilization capacity of the water turbine is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent design of water turbines, and in particular relates to a water turbine runner optimization design method combining a salp optimization algorithm and a BP neural network. Background Art

[0002] As a key hydraulic energy conversion device, hydraulic turbine plays a vital role in the field of renewable energy. The design of traditional hydraulic turbine runner is often affected by many complex factors, including fluid dynamics, material mechanical properties, and manufacturing process. Therefore, a more direct and effective optimization method is needed for the design of hydraulic turbine runner.

[0003] In recent years, the application of artificial intelligence technology in the engineering field has become increasingly mature. Among them, neural network, as an important machine learning method, has shown great potential in optimization problems. As one of the representatives, BP neural network has powerful nonlinear mapping and optimization capabilities and has broad application prospects in turbine runner design.

[0004] As a new optimization algorithm, the salp optimization algorithm simulates the optimization behavior of salps in the marine environment and has strong global search capabilities and robustness. Combining the salp optimization algorithm with the BP neural network can fully utilize the advantages of both and achieve the optimal design of the turbine runner.

[0005] Therefore, the turbine runner optimization design method combining the salp optimization algorithm and the BP neural network has important theoretical significance and practical application value. The technical problem to be solved by the present invention is: the limitations of the traditional turbine runner optimization design method in improving the efficiency of the turbine or utilizing the head capacity, that is, the traditional method has a large workload, is time-consuming and may not fully improve the performance of the turbine. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a turbine runner optimization design method combining the salp optimization algorithm and the BP neural network, so as to fully improve the turbine efficiency or water head by reasonably designing the runner, and effectively improve the turbine power generation efficiency or the ability to utilize the water head.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is: A method for optimizing the design of a hydraulic turbine runner by combining a salp optimization algorithm and a BP neural network, comprising the following steps: S1, select the runner geometric parameters that affect the performance of the turbine; S2, obtain the wheel design scheme based on orthogonal experiments; S3, according to the runner design scheme, designing the structural parts of the runner corresponding to the runner design scheme, and constructing a three-dimensional overall model of the turbine; S4, calculating the head and efficiency of the turbine corresponding to each runner design scheme based on computational fluid dynamics (CFD); S5, based on the BP neural network, a neural network model of turbine efficiency characteristics and a neural network model of water head characteristics are constructed, which take the runner geometric parameters as input and the turbine efficiency and water head as output respectively; S6, the optimal geometric parameters of the runner are obtained by combining the salp optimization algorithm, the turbine efficiency characteristic neural network model and the head characteristic neural network model. Preferably, in step S1, the impeller geometric parameters include blade inlet placement angle, inlet diameter, outlet diameter, inlet width and number of blades, and blade inlet placement angle, inlet diameter, outlet diameter, inlet width and number of blades are selected as basic geometric parameters of the impeller to be optimized.

[0008] Preferably, in step S2, obtaining the wheel design scheme based on the orthogonal experiment means: first selecting the number of levels corresponding to the five factors, then designing the table header, and further using the Octave tool to determine the orthogonal scheme.

[0009] Preferably, the specific process of step S3 is: according to the turbine design theory, constructing a three-dimensional model of the turbine under different orthogonal schemes; the constructed three-dimensional model of the turbine includes a volute, fixed guide vanes, movable guide vanes, a runner and a tailwater pipe.

[0010] Preferably, the specific process of step S4 is: combining the three-dimensional turbine models corresponding to each orthogonal scheme, and calculating the head and efficiency of the turbine of each scheme based on computational fluid dynamics (CFD).

[0011] Preferably, the specific process of step S5 is: based on the BP neural network, a turbine efficiency characteristic neural network model and a head characteristic neural network model are constructed with the turbine geometric parameters as input and the turbine efficiency and the water head as output respectively; wherein the number of neurons in the hidden layer of the neural network model is determined by trial and error.

[0012] Preferably, the specific process of step S6 is: Combining the salp optimization algorithm, the neural network model of turbine efficiency characteristics and the neural network model of head characteristics, the Octave tool is used to obtain the optimal geometric parameters of the runner.

[0013] To ensure the rationality of the calculation results, it is necessary to set the range of the runner geometric parameters when calculating the runner parameters corresponding to the optimal performance of the turbine based on the salp optimization algorithm. The range of the runner geometric parameters is set according to the maximum and minimum values ​​in the sample data of the efficiency characteristic neural network and the head characteristic neural network model, and rounded during calculation. After setting the initial conditions, the objective function needs to be adjusted during calculation. If the efficiency characteristic neural network model is used, the efficiency is used as the objective function; if the head characteristic neural network model is used, the head is used as the objective function.

[0014] Preferably, the three-dimensional turbine model corresponding to each orthogonal scheme is combined to calculate the head and efficiency of the turbine of each scheme based on computational fluid dynamics (CFD). The method is: Before the model starts calculating, it is necessary to select a numerical simulation method. During the simulation calculation process, the basic equations of turbulence use the continuity equation shown in equation (1), the momentum equation shown in equation (2), and the energy equation shown in equation (3): ; ; ; In the formula, is the fluid velocity in coordinates Directional weight; is the fluid density; t For time; p is the static pressure; Custom source items include gravity and the interaction between multiphase flows; is the stress tensor; is the total specific enthalpy; e is internal energy; is the effective heat transfer coefficient; T For temperature.

[0015] Standard The turbulence model can effectively simulate the internal flow field of the turbine, so the turbulence model selected in the present invention is a standard Turbulence model. Standard The turbulence model is to introduce the turbulent kinetic energy k and dissipation rate The transport equation is k and Calculating turbulent kinetic viscosity , which belongs to the two-equation model. k and dissipation rate The transport equation is as follows: ; ; In the formula, is the turbulent kinetic energy caused by the mean velocity gradient k The generated items; The turbulent kinetic energy is caused by buoyancy k The generated items; , and is an empirical constant; and is turbulent kinetic energy k and dissipation rate The corresponding Prandtl number; and It is a user-defined source item.

[0016] In this model, the turbulent kinetic energy dissipation rate is is defined as: ; The characteristic speed of the reaction is the turbulent kinetic energy k , the characteristic length scale of the response is the dissipation rate , the relationship between the two reflects the turbulent viscosity : ; Preferably, in step S5, the maximum number of neurons in the hidden layer of the neural network model is first determined to be 12 according to formula (8), and then the final value is determined by trial and error with the network error being small and the network not being overfitted as the standard: (8); In the formula, l is the number of hidden layer nodes, m is the number of nodes in the input layer; n is the number of nodes in the output layer; a is a constant between 0 and 10.

[0017] Preferably, the calculation process of the salp optimization algorithm is as follows: Step 1: Set the maximum number of iterations ( T max), spatial dimension ( Ndim ) and the number of salps ( N s), and initialize the position of the salp according to equation (9).

[0018] (9); Where: yes N s OK N dim A random matrix of columns, For the i Followers in j Dimensional position, ub j , lb j Respectively j The upper and lower bounds of the dimensional search space.

[0019] Step 2: Calculate the fitness value of each salp according to the fitness function.

[0020] Step 3: Rank the salps according to their fitness values ​​and define the best salp position as the food source position.

[0021] Step 4: Update the leader position according to equation (10).

[0022] (10); Where: For food source j The location of the dimension, For leaders in j Position in dimension; c 2 and c 3 is a random number between 0 and 1. c 2 represents the step length of salps, c 3Determine the direction in which the leader will move; c 1 is a number between 0 and 2, used to balance the algorithm's global exploration and development capabilities. c 1>1, SSA is in the exploration stage. c When 1<1, SSA is in the development stage.

[0023] Step 5: Update the follower position according to equation (11).

[0024] (11); In the formula, Indicates i -1 follower in j Position in dimension.

[0025] Step 6: Calculate the fitness value of the salp and update the location of the food source.

[0026] Step 7: Determine whether the end condition is met based on whether the maximum number of iterations has been reached. If so, output the calculation result; otherwise, continue iterating.

[0027] As for the objective function of SSA J , when the goal is to optimize efficiency, the efficiency value η is the objective function, as shown in formula (12); when the optimal water head is the goal, the water head value H is the objective function, as shown in formula (13).

[0028] (12); (13).

[0029] The present invention can achieve the following beneficial effects: 1. The invention designs a turbine runner optimization design method combining the salp optimization algorithm and the BP neural network, which fully improves the turbine efficiency or water head by rationally designing the runner, and effectively improves the turbine power generation efficiency or the ability to utilize the water head.

[0030] 2. As the core equipment for power generation in hydropower stations, the improvement of the operating performance of water turbines has always been a hot topic in the industry. Aiming at the limitations of traditional water turbine performance improvement, which is labor-intensive and time-consuming, the present invention proposes a water turbine runner optimization design method combining the salp optimization algorithm and the BP neural network.

[0031] 3. The beneficial effects of the present invention are: under the premise of short time consumption and high reliability, the optimized runner geometric parameters can fully guarantee the efficiency or water head of the turbine and effectively ensure the efficient operation of the hydropower unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 Designing steps for the method of the present invention; Figure 2 The calculation process of the salp optimization algorithm of the present invention; Figure 3 This is the neural network structure of the turbine efficiency and water head characteristics of the present invention. DETAILED DESCRIPTION

[0033] The preferred solution is Figures 1 to 3As shown, a method for optimizing the runner of a water turbine combined with a salp optimization algorithm and a BP neural network is shown. The method for optimizing the runner of a water turbine combined with a salp optimization algorithm and a BP neural network designed by the present invention includes three parts: obtaining the neural network modeling data of the water turbine efficiency characteristics and the water head characteristics based on orthogonal experiments and CFD, constructing the neural network model of the water turbine efficiency characteristics and the water head characteristics, and obtaining the optimal geometric parameters of the runner based on the salp optimization algorithm, the neural network model of the water turbine efficiency characteristics and the water head characteristics. The implementation steps mainly include: selecting the geometric parameters of the runner that affect the performance of the water turbine; obtaining the runner design scheme based on the orthogonal experiment; designing the structural accessories of the runner of each scheme and constructing a three-dimensional overall model of the water turbine; calculating the water head and efficiency of the water turbine of each scheme based on computational fluid dynamics (CFD); constructing the neural network model of the water turbine efficiency characteristics and the water head characteristics based on the BP neural network, which takes the runner geometric parameters as input and the water turbine efficiency and water head as output respectively; and obtaining the optimal geometric parameters of the runner by combining the salp optimization algorithm, the neural network model of the water turbine efficiency characteristics and the neural network model of the water head characteristics.

[0034] The turbine runner optimization design method combining the salp optimization algorithm and the BP neural network can fully improve the turbine efficiency or water head, and effectively improve the turbine power generation efficiency or the ability to utilize water head. The specific steps are: (1) Select the runner geometric parameters that affect turbine performance.

[0035] Select the blade inlet placement angle β 1. Inlet diameter D 1. Outlet diameter D 2. Import width B 1 and the number of blades Z , as the basic geometric parameters of the runner to be optimized (5 factors).

[0036] (2) Obtain the wheel design scheme based on orthogonal experiments.

[0037] First, the number of levels corresponding to the five factors is selected, then the table header is designed, and further programming using the Octave tool is used to determine the orthogonal scheme.

[0038] (3) Design the structural components of the runner for each scheme and construct a three-dimensional overall model of the turbine.

[0039] According to the turbine design theory, a three-dimensional model of the turbine under different orthogonal schemes is constructed. The constructed three-dimensional model of the turbine includes volute, fixed guide vanes, movable guide vanes, runner and draft tube.

[0040] (4) Combined with the three-dimensional turbine models corresponding to each orthogonal scheme, the head and efficiency of the turbine of each scheme are calculated based on computational fluid dynamics (CFD).

[0041] Before the model starts calculating, it is necessary to select a numerical simulation method. During the simulation calculation process, the basic equations of turbulence use the continuity equation shown in equation (1), the momentum equation shown in equation (2), and the energy equation shown in equation (3).

[0042] ; ; ; In the formula, is the fluid velocity in coordinates Directional weight; is the fluid density; t For time; p is the static pressure; Custom source items include gravity and the interaction between multiphase flows; is the stress tensor; is the total specific enthalpy; e is internal energy; is the effective heat transfer coefficient; T For temperature.

[0043] Standard The turbulence model can effectively simulate the internal flow field of the turbine, so the turbulence model selected in the present invention is a standard Turbulence model. Standard The turbulence model is to introduce the turbulent kinetic energy k and dissipation rate The transport equation is k and Calculating turbulent kinetic viscosity , which belongs to the two-equation model. k and dissipation rate The transport equation is as follows: ; ; In the formula, is the turbulent kinetic energy caused by the mean velocity gradient k The generated items; The turbulent kinetic energy is caused by buoyancy k The generated items; , and is an empirical constant; and is turbulent kinetic energy k and dissipation rate The corresponding Prandtl number; and It is a user-defined source item.

[0044] In this model, the turbulent kinetic energy dissipation rate is is defined as: ; The characteristic speed of the reaction is the turbulent kinetic energy k , the characteristic length scale of the response is the dissipation rate , the relationship between the two reflects the turbulent viscosity : ; (5) Based on the BP neural network, a neural network model of turbine efficiency characteristics and a neural network model of water head characteristics are constructed with the runner geometric parameters as input and the turbine efficiency and water head as output respectively.

[0045] According to the data in the orthogonal scheme with supplemented efficiency and head values, a turbine efficiency characteristic neural network model and a head characteristic neural network model are constructed based on the BP neural network in the feedforward neural network, which take the runner geometric parameters as input and the turbine efficiency and head as output respectively.

[0046] The maximum number of neurons in the hidden layer of the neural network model is first determined as 12 according to the empirical formula (8), and then the final value is determined by trial and error with the network error being small and the network not being overfitted as the standard.

[0047] (8); In the formula, l is the number of hidden layer nodes, m is the number of nodes in the input layer; n is the number of nodes in the output layer; a is a constant between 0 and 10.

[0048] (6) The optimal geometric parameters of the runner are obtained by combining the salp optimization algorithm, the turbine efficiency characteristic neural network model and the head characteristic neural network model.

[0049] Combining the salp optimization algorithm, the neural network model of turbine efficiency characteristics and the neural network model of head characteristics, the optimal geometric parameters of the runner corresponding to the optimal turbine efficiency and optimal head are obtained respectively by using Octave tool programming.

[0050] Among them, the salp optimization algorithm is a heuristic algorithm inspired by the behavior of marine organisms such as salps. Salps are molluscs that attach to the seabed or other objects in a fixed manner in the ocean and obtain food and survive through expansion and contraction movements. The core idea of ​​the salp optimization algorithm is to simulate the survival behavior of salps in the marine environment and find the optimal solution through cooperation and competition between individuals. In the salp optimization algorithm, each individual represents a potential solution in the solution space. Individuals search for the optimal solution by exchanging information and collaborating with each other. The parameter adjustment and individual position update process in the algorithm are inspired by the ecological behavior and environmental factors of salps, which makes the algorithm have strong global search ability and robustness.

[0051] The calculation process of the Salp Sac Optimization Algorithm (SSA) is as follows: Step 1: Set the maximum number of iterations ( T max), spatial dimension ( Ndim ) and the number of salps ( N s), and initialize the position of the salp according to equation (9).

[0052] (9); Where: yes N s OK N dim A random matrix of columns, For the i Followers in j Dimensional position, ub j , lb j Respectively j The upper and lower bounds of the dimensional search space.

[0053] Step 2: Calculate the fitness value of each salp according to the fitness function.

[0054] Step 3: Rank the salps according to their fitness values ​​and define the best salp position as the food source position.

[0055] Step 4: Update the leader position according to equation (10).

[0056] (10); Where: For food source j The location of the dimension, For leaders in j Position in dimension; c 2 and c 3 is a random number between 0 and 1.c 2 represents the step length of salps, c 3Determine the direction in which the leader will move; c 1 is a number between 0 and 2, used to balance the algorithm's global exploration and development capabilities. c 1>1, SSA is in the exploration stage. c When 1<1, SSA is in the development stage.

[0057] Step 5: Update the follower position according to equation (11).

[0058] (11); In the formula, Indicates i -1 follower in j Position in dimension.

[0059] Step 6: Calculate the fitness value of the salp and update the location of the food source.

[0060] Step 7: Determine whether the end condition is met based on whether the maximum number of iterations has been reached. If so, output the calculation result; otherwise, continue iterating.

[0061] As for the objective function of SSA J , when the goal is to optimize efficiency, the efficiency value η is the objective function, as shown in formula (12); when the optimal water head is the goal, the water head value H is the objective function, as shown in formula (13).

[0062] (12); (13); Aiming at the limitations of the traditional turbine runner optimization design method, the present invention proposes a new runner geometric parameter optimization method to improve the optimization efficiency and reliability of the results. This method will fully guarantee the turbine efficiency or water head and effectively guarantee the efficient operation of the hydropower unit.

[0063] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A method for optimizing the runner of a hydraulic turbine by combining a salp optimization algorithm and a BP neural network, characterized in that: The following steps are involved: S1, select the runner geometric parameters that affect the performance of the turbine; S2, obtain the wheel design scheme based on orthogonal experiments; S3, according to the runner design scheme, designing the structural parts of the runner corresponding to the runner design scheme, and constructing a three-dimensional overall model of the turbine; S4, calculating the head and efficiency of the turbine corresponding to each runner design scheme based on computational fluid dynamics; S5, based on the BP neural network, a neural network model of turbine efficiency characteristics and a neural network model of water head characteristics are constructed, which take the runner geometric parameters as input and the turbine efficiency and water head as output respectively; S6, the optimal geometric parameters of the runner are obtained by combining the salp optimization algorithm, the turbine efficiency characteristic neural network model and the head characteristic neural network model.

2. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: In step S1, the impeller geometric parameters include blade inlet placement angle, inlet diameter, outlet diameter, inlet width and number of blades, and blade inlet placement angle, inlet diameter, outlet diameter, inlet width and number of blades are selected as basic geometric parameters of the impeller to be optimized.

3. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: In step S2, obtaining the wheel design scheme based on the orthogonal experiment means: firstly selecting the number of levels corresponding to the five factors, then designing the table header, and further using the Octave tool to determine the orthogonal scheme.

4. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: The specific process of step S3 is: according to the turbine design theory, constructing a three-dimensional turbine model under different orthogonal schemes; the constructed three-dimensional turbine model includes a volute, fixed guide vanes, movable guide vanes, a runner and a tailwater pipe.

5. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: The specific process of step S4 is: combining the three-dimensional turbine models corresponding to each orthogonal scheme, and calculating the head and efficiency of the turbine of each scheme based on computational fluid dynamics.

6. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: The specific process of step S5 is: based on the BP neural network, a neural network model of turbine efficiency characteristics and a neural network model of water head characteristics are constructed, which take the geometric parameters of the runner as input and the turbine efficiency and water head as output respectively; wherein the number of neurons in the hidden layer of the neural network model is determined by trial and error.

7. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 1, characterized in that: The specific process of step S6 is: Combining the salp optimization algorithm, the turbine efficiency characteristic neural network and the head characteristic neural network model, the Octave tool is used to obtain the optimal geometric parameters of the runner; To ensure the rationality of the calculation results, it is necessary to set the variation range of the runner geometric parameters when calculating the runner parameters corresponding to the optimal performance of the turbine based on the salp optimization algorithm; set the variation range of the runner geometric parameters according to the maximum and minimum values ​​in the sample data of the efficiency characteristic neural network and the head characteristic neural network model, and round them during calculation; after setting the initial conditions, the objective function needs to be adjusted during calculation; if the efficiency characteristic neural network model is adopted, the efficiency is used as the objective function; if the head characteristic neural network model is adopted, the head is used as the objective function.

8. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 5, characterized in that: The three-dimensional turbine model corresponding to each orthogonal scheme is combined to calculate the head and efficiency of the turbine of each scheme based on computational fluid dynamics; the method is: Before the model starts calculating, it is necessary to select a numerical simulation method. During the simulation, the basic equations of turbulence use the continuity equation shown in equation (1), the momentum equation shown in equation (2), and the energy equation shown in equation (3): ； ； ； In the formula, is the fluid velocity in coordinates Directional weight; is the fluid density; t For time; p is the static pressure; Custom source items include gravity and the interaction between multiphase flows; is the stress tensor; is the total specific enthalpy; e is internal energy; is the effective heat transfer coefficient; T is temperature; About turbulent energy k and dissipation rate The transport equation is as follows: ； ； In the formula, is the turbulent kinetic energy caused by the mean velocity gradient k The generated items; The turbulent kinetic energy is caused by buoyancy k The generated items; , and is an empirical constant; and is turbulent kinetic energy k and dissipation rate The corresponding Prandtl number; and It is a user-defined source item; In this model, the turbulent kinetic energy dissipation rate is is defined as: ； The characteristic speed of the reaction is the turbulent kinetic energy k , the characteristic length scale of the response is the dissipation rate , the relationship between the two reflects the turbulent viscosity : 。 9. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 6, characterized in that: In step S5, the maximum number of neurons in the hidden layer of the neural network model is first set to 12 according to formula (8), and then the final value is determined by trial and error, with the network error being small and the network not being overfitted as the standard: (8)； In the formula, l is the number of hidden layer nodes, m is the number of nodes in the input layer; n is the number of nodes in the output layer; a is a constant between 0 and 10.

10. The method for optimizing the turbine runner by combining the salp optimization algorithm and the BP neural network according to claim 7, characterized in that: The calculation process of the salp optimization algorithm is as follows: Step 1: Set the maximum number of iterations T max, spatial dimension Ndim and the number of salps N s, and initialize the position of the salp according to formula (9); （9）； Where: yes N s OK N dim A random matrix of columns, For the i Followers in j Dimensional position, ub j , lb j Respectively j Dimensional search space upper and lower bounds; Step 2: Calculate the fitness value of each salp according to the fitness function; Step 3: Rank the salps according to their fitness values ​​and define the best salp position as the food source position; Step 4: Update the leader position according to formula (10); （10）； Where: For food source j The location of the dimension, For leaders in j Position in dimension; c 2 and c 3 is a random number between 0 and 1. c 2 represents the step length of salps, c 3Determine the direction in which the leader will move; c 1 is a number between 0 and 2, used to balance the algorithm's global exploration and development capabilities; when c 1>1, SSA is in the exploration stage. c When 1<1, SSA is in the development stage; Step 5: Update the follower position according to formula (11); （11）； In the formula, Indicates i -1 follower in j Position in dimension; Step 6: Calculate the fitness value of the salp and update the location of the food source; Step 7: Determine whether the end condition is met based on whether the maximum number of iterations has been reached; If yes, output the calculation result; otherwise, continue iterating; As for the objective function of SSA J , when the goal is to optimize efficiency, the efficiency value η is the objective function, as shown in formula (12); when the optimal water head is the goal, the water head value H is the objective function, as shown in formula (13); (12)； (13)。

Images