Variable-frequency single-unit variable-speed optimization scheduling method based on multi-strategy fusion improved particle swarm optimization

Abstract

The invention discloses a variable-frequency single-unit variable-speed optimization scheduling method based on a multi-strategy fusion improved particle swarm optimization algorithm, and the method comprises the steps: building a model according to the single-unit data of a pump station, building a variable-frequency single-unit variable-speed optimization model with the minimum power consumption cost as a target, the total water lifting amount in one day and the unit operation power are used as constraint conditions; aiming at the established variable-frequency single-unit variable-speed optimization model, solving by adopting a multi-strategy fusion improved particle swarm algorithm to obtain a decision-making process and operating parameters of a pump station single unit; and realizing single-unit operation of the pump station based on the decision-making process and the operation parameters. Sine and cosine factors are introduced to replace learning factors, and the nonlinear decreasing inertia weight is combined, so that the collaborative improvement of the search capability of the particle swarm algorithm in different stages is realized, the calculation precision of the algorithm is improved, and the feasibility of the optimization scheduling result is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of pump station optimization scheduling, and in particular to a variable speed optimization scheduling method for a variable frequency single unit based on a multi-strategy fusion improved particle swarm algorithm. Background Art

[0002] With the continuous expansion of the application field of pump station engineering and the increasing complexity of operation and management, the optimization and dispatching of pump station units to achieve economic operation has become an important research field. Many scholars at home and abroad have explored and studied the construction and solution methods of pump station optimization models: in terms of model construction, the main goal is to minimize the power consumption cost of pump station operation. The target value is greatly affected by electricity price and head. Therefore, it is hoped that less or no water will be pumped when the electricity price is high and the head is low, and more water will be pumped when the electricity price is low and the head is low, so as to achieve the lowest single unit operation cost under the condition of meeting the water lifting capacity constraint. In terms of model solution, traditional methods mainly include dynamic programming, nonlinear programming, system test optimization method, etc. Modern intelligent algorithms mainly include genetic algorithm, particle swarm algorithm, simulated annealing algorithm and wolf pack algorithm.

[0003] With the continuous expansion of the application field of pump station engineering and the increasing complexity of operation and management, the optimization and dispatching of pump station units to achieve economic operation has become an important research field. Many scholars at home and abroad have explored and studied the construction and solution methods of pump station optimization models: in terms of model construction, the main goal is to minimize the power consumption cost of pump station operation. The target value is greatly affected by electricity price and head. Therefore, it is hoped that less or no water will be pumped when the electricity price is high and the head is low, and more water will be pumped when the electricity price is low and the head is low, so as to achieve the lowest operating cost of a single unit under the condition of meeting the water lifting capacity constraint. In terms of model solving, traditional methods mainly include dynamic programming, nonlinear programming, system test optimization method, etc., but traditional methods have more or less problems such as the curse of dimensionality and complex calculation process. With the development of computer technology, modern intelligent algorithms (such as genetic algorithms, particle swarm algorithms, simulated annealing algorithms and wolf pack algorithms) have provided a new way to solve the optimization and scheduling of pump station projects. However, when the objective function is a complex multi-peak function, the intelligent algorithm may fall into a local optimal solution; for complex multi-dimensional problems, its spatial search ability, convergence speed and accuracy will be affected by parameter settings. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a variable speed optimization scheduling method for a single frequency unit based on a multi-strategy fusion improved particle swarm algorithm.

[0005] Technical solution: The present invention comprises the following steps:

[0006] (1) A model is established based on the data of a single unit in a pumping station. A variable frequency single unit speed optimization model is established with the goal of minimizing power consumption. The decision variable is the speed of the unit in each period, and the total amount of water pumped in a day and the operating power of the unit are used as constraints.

[0007] (2) Based on the established variable frequency single unit speed optimization model, the particle swarm algorithm improved by multi-strategy fusion is used to solve the decision-making process and operating parameters of the pump station single unit; the operating parameters include the unit speed, flow, efficiency, water lifting volume, operating power consumption and average head per hour under different time periods;

[0008] (3) Based on the decision-making process and operating parameters of step (2), the single unit operation of the pump station is realized.

[0009] Furthermore, the variable speed optimization model of the variable frequency single unit in step (1) includes:

[0010] Objective function:

[0011]

[0012] Constraint 1: The total amount of water pumped in one day is

[0013] Constraint 2: Unit rated power constraint, i.e. N i (n i (l) )≤N e ;

[0014] Wherein, F represents the power consumption cost of a single unit; i represents the time period, i=1, 2, ..., SN, SN represents the total number of time periods divided in a day; They represent the lth speed value and the corresponding water lifting flow rate in the i-th period respectively; H i , ΔT i and P i They represent the average lift, start-up time and time-of-use electricity price in the i-th period respectively; η z，i , η mot , η int , η f Respectively represent the efficiency of single unit, motor efficiency, transmission efficiency and frequency conversion device efficiency; W d is the total daily water extraction volume of a single unit; N e Indicates rated power; N i (n i ( l)) represents the actual operating power of the unit in the i-th period.

[0015] Furthermore, the step (2) includes parameter setting, optimizing initial population improvement, solving fitness and updating optimal position, improving inertia weight, improving learning factor, updating position and judging iteration conditions.

[0016] Furthermore, the parameter setting includes:

[0017] Set the population number and movement speed range V max and V min Unit speed range n max and n min , the current number of iterations t and the maximum number of iterations T.

[0018] Furthermore, the optimization of the initial population improvement includes:

[0019] The Sobol sequence is introduced to initialize the population. The population position formula generated by the Sobol sequence initialization is:

[0020] X i =(b u -b l )*r+b l

[0021] Among them, b u and b l are the upper and lower bounds of the search space respectively; r is a random number in the range of [0, 1] generated by the Sobol sequence.

[0022] Furthermore, the fitness solution and optimal position update include:

[0023] The fitness value is calculated according to the objective function; and according to the fitness value, the current position is compared with the previous optimal position pbest. If it is better, the current position is updated to the local optimal position pbest; the current position is compared with the historical optimal position gbest. If it is better, the current position is updated to the global optimal position gbest.

[0024] Furthermore, the inertia weight improvement includes:

[0025] An inertia weight that decreases nonlinearly with the number of iterations is used; the change law of ω with the number of iterations is as follows:

[0026]

[0027] Among them, ω max ,ω min is the maximum and minimum value of the inertia weight; t is the current number of iterations, T is the maximum number of iterations; ω max Take 0.9, ω min Take 0.4.

[0028] Furthermore, the learning factor improvement includes:

[0029] The sine and cosine factors in the sine and cosine algorithm position update formula are introduced to replace the learning factor. At this time, the particle velocity update formula is as follows:

[0030]

[0031] r 1 =0.4(1-t / T)

[0032] Where t represents the number of iterations; v li t+1 、x li t+1 They represent the velocity and position of particle l in dimension i after iteration t+1 times; ω represents the inertia weight; pbest li t represents the optimal position of particle l in dimension i after t iterations; gbest i t represents the optimal position of the entire population on dimension i after t iterations; c 1 、c 2 Represent the learning factors of particles and populations respectively; r 1 、r 2 is a random number in the range [0, 1]; r 1 The control algorithm is transformed from global search to local development; 2 describes the direction in which the individual moves, r 2 ∈[0, 2π], r 3 is the probability between sine update and cosine update, r 3 ∈[0, 1].

[0033] Furthermore, the particle position of the position update is expressed as:

[0034]

[0035] Among them, x li t 、x li t+1 They represent the position of particle l in dimension i after iterations t and t+1 respectively.

[0036] Furthermore, the iterative judgment condition includes:

[0037] When t≤T, repeat the steps of optimizing initial population improvement, inertia weight improvement, learning factor improvement and position update until exiting the iteration process and outputting the results, including the time-averaged lift H i , speed n at different time periods i 、Period flow Qi 、Equipment efficiency η i , water extraction volume in each period and total operating electricity consumption minF.

[0038] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: a combined improvement method of "Sobol sequence optimization of initial population + real-time adjustment of inertia weight + sine-cosine replacement of learning factor" multi-strategy fusion is proposed, which makes the overall distribution of the initial population more uniform, laying the foundation for global search; a nonlinear decreasing inertia weight that can be adjusted in real time with the number of iterations is adopted, which improves the search ability of the algorithm at different stages; the sine-cosine factor is introduced to replace the learning factor, combined with the nonlinear decreasing inertia weight, to achieve the coordinated improvement of the search ability of the particle swarm algorithm at different stages, improve the calculation accuracy of the algorithm, and ensure the feasibility of the optimized scheduling results. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of the present invention;

[0040] Figure 2 Comparison chart for Sobol sequence and random population initialization;

[0041] Figure 3 is the convergence curve of the benchmark function;

[0042] Figure 4 This is a comparison chart of water extraction costs and unit water extraction costs of the two algorithms;

[0043] Figure 5 Figure 5 is the corresponding relationship diagram between the decision-making process of the two methods and the electricity price and lift in each period. Figure 5(a) is the corresponding relationship diagram between the electricity price and lift when the load is 60%; Figure 5(b) is the corresponding relationship diagram between the electricity price and lift when the load is 80%; Figure 5(c) is the corresponding relationship diagram between the electricity price and lift when the load is 100%. DETAILED DESCRIPTION

[0044] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0045] Taking the optimization dispatch of a single unit of a large variable frequency water pumping station as an example, the dynamic programming method and the particle swarm algorithm based on multi-strategy fusion improvement are used to solve the model, and the optimization results are compared and analyzed to provide a reference for the selection of optimization solution methods for single units or even multiple units. The single unit of the large variable frequency water pumping station is equipped with ten 2000ZLQ13.7-7.8 pumps with a rated speed of 214.3r / min. The discrete values ​​of the unit speed are 180, 185, 190, 195, 200, 205, 208, 211, 214.3, 218, 224. The design blade angle is +2°, and its adjustment range is [-6°, +4°]. The operating loads of the single unit are 60% load, 80% load, and 100% load (full load), and the corresponding water lifting capacity is 8.5×10 5 m 3 / d, 1.14×10 6 m 3 / d, 1.42×10 6 m 3 / d.

[0046] like Figure 1 As shown, this method description includes the following steps:

[0047] Step 1) Establish a model: Establish the objective function with the minimum total power consumption cost of startup operation. The specific optimization model is as follows:

[0048] Objective function:

[0049]

[0050] Constraint 1: The total amount of water pumped in one day is

[0051] Constraint 2: Unit rated power constraint, i.e. N i (n i (l) )≤N e (3)

[0052] Where: F represents the power consumption cost of a single unit (yuan); i represents the time period, i=1, 2, ..., SN (the total number of time periods in a day); They represent the ith speed value (r / min) and the corresponding water flow rate (m 3 / s); H i , ΔT i and P i They represent the average head (m), start-up time (h) and time-of-use electricity price (yuan / (kW·h)) in the i-th period respectively; Respectively represent the efficiency of single unit, motor efficiency, transmission efficiency and frequency conversion device efficiency; W dis the total daily water extraction volume of a single unit; N e Indicates rated power (kW); N i (n i (l) ) represents the actual operating power of the unit in the i-th period (kW).

[0053] Step 2) Solving the model: The particle swarm algorithm based on multi-strategy fusion is used for solving. The specific steps are as follows:

[0054] Step 1 Parameter setting: population number, maximum movement speed V max , Unit speed range n max and n min , the current number of iterations t and the maximum number of iterations T;

[0055] Step 2 Optimize the initial population: Introduce the Sobol sequence to initialize the generated population, such as Figure 2 As shown, the population position formula generated by Sobol sequence initialization is:

[0056] X i =(b u -b l )*r+b l (4)

[0057] Where: b u and b l are the upper and lower bounds of the search space respectively; r is a random number in the range of [0, 1] generated by the Sobol sequence.

[0058] Step 3: Calculate the fitness value and update the optimal position: Calculate the fitness value according to formula (1); and compare the current position with the previous optimal position pbest according to the fitness value. If it is better, update the current position to the local optimal position pbest; compare the current position with the historical optimal position gbest. If it is better, update the current position to the global optimal position gbest.

[0059] Step 4 Inertia weight improvement: Use an inertia weight that decreases nonlinearly with the number of iterations. The change rule of ω with the number of iterations is as follows:

[0060]

[0061] Where: max ,ω min is the maximum and minimum value of the inertia weight; t is the current iteration number, and T is the maximum iteration number. In general, ω max It can be 0.9, ω min 0.4 can be taken.

[0062] Step 5 Learning factor improvement: Introduce the sine and cosine factors in the SCA (sine and cosine algorithm) position update formula to replace the learning factor. At this time, the particle velocity update formula is as follows:

[0063]

[0064] r 1 =0.4(1-t / T) (7)

[0065] Where: t represents the number of iterations; V l it+1 、x li t+1 They represent the velocity and position of particle l in dimension i after iteration t+1 times; ω represents the inertia weight; pbest li t represents the optimal position of particle l in dimension i after t iterations; gbest i t represents the optimal position of the entire population on dimension i after t iterations; c 1 、c 2 Represent the learning factors of particles and populations respectively; r 1 、r 2 is a random number in the range [0, 1]. 1 The control algorithm is transformed from global search to local development; 2 describes the direction in which the individual moves, r 2 ∈[0, 2π], r 3 is the probability between sine update and cosine update, r 3 ∈[0, 1].

[0066] Step 6 Location Update:

[0067] Where: x li t 、x li t+1 They represent the position of particle l in dimension i after iterations t and t+1 respectively.

[0068] Step 7 determines the iteration condition: when t≤T, repeat Step 3-Step 6 until the iteration process is exited and the result is output. The result information includes the average head H i , speed n at different time periods i 、Period flow Q i 、Equipment efficiency η i , water extraction volume in each period and total operating electricity consumption minF.

[0069] Step 3) Operate the single unit of the pump station according to the above parameters.

[0070] Average head H in each period in step 1) i The determination process is to consider the typical tidal process of the river in the area where the pump station is located, the changes in different hydrological years, the tidal range in different periods and other factors, select a reasonable daily average head, and obtain the average head H for each period of the day based on the daily average head H and the river tidal level change curve. i .

[0071] The basis for dividing different time periods i in a day in step 1) is that according to the peak and valley electricity prices in the area where the pump station is located and the requirement that large pump stations should not be frequently started and stopped, a day is divided into SN time periods, and the duration of each time period is ΔT i , i∈[1,SN], the corresponding peak-valley electricity price P in the i-th period i .

[0072] In step 1), according to the performance curve equations of H~Q and Q~η, the flow rate, head and power of the pump meet the proportional rate within a certain speed change range, and the efficiency at similar operating points remains unchanged. Under the condition that the blade angle remains unchanged, according to the speed n e And the corresponding flow Q e And the device head H e , we can get different speed n i Average head H i The flow rate Q i 、Equipment efficiency η i .

[0073] In order to test the performance of the improved PSO algorithm, four benchmark functions were selected for testing. Matlab software was used to program and run 30 times independently. The corresponding optimal value, average value and standard deviation were obtained based on the 30 calculation results, and compared with the basic PSO algorithm. The convergence of each benchmark function is shown in the figure. Figure 3 shown.

[0074] According to the algorithm test results and function convergence, the improved PSO algorithm has good convergence. In the optimal value test, for the four functions, the accuracy of the improved PSO algorithm is significantly better than the basic PSO algorithm; in the average value test, the optimization accuracy of the Sphere function and the Step function is significantly higher than that of the basic PSO algorithm, and the accuracy of the Schwefel function and the Ackley function is also higher than that of the basic PSO algorithm; in the standard deviation test, the accuracy of the Sphere function and the Step function is significantly improved compared with the basic PSO algorithm, and the accuracy of the Schwefel function and the Ackley function is also greatly improved. It can be seen that for different benchmark functions, the improved PSO algorithm has a significant improvement in optimization accuracy.

[0075] By using the above-mentioned combination of improved particle swarm algorithm and Matlab software programming calculation, the single unit speed optimization operation plan under different operating load scenarios with an average daily head of 5.8m is obtained, as shown in Table 1; at the same time, the one-dimensional dynamic programming method and the basic particle swarm algorithm are used to compare and analyze the model solution. Here, taking the single unit 80% operating load scenario as an example, the single unit speed optimization decision process or startup plan obtained by calculating three methods (basic PSO, improved PSO and dynamic programming method) is given, as shown in Table 2.

[0076] Comparison and analysis of the results of the basic PSO and improved PSO algorithms show that under the conditions of single unit variable speed optimization operation with different loads, the water extraction volume, cost and corresponding unit water extraction cost (i.e. the ratio of power consumption cost to water extraction volume) of the two algorithms are shown in Table 3. Figure 4 As shown in the figure, when the single unit is running at 100% load, the difference between the two algorithms in water extraction costs is 57 yuan (0.3%), and when it is running at 60% load, the difference is 1056 yuan (11.4%); when it is running at 100% load, the deviation rate of unit water extraction costs of the two algorithms is -0.34%, and when it is running at 60% load, it reaches -11.93%. It can be seen that under different operating load conditions, the water extraction costs and unit water extraction costs of the variable speed optimization operation of the single unit are both better than the basic PSO algorithm. The lower the operating load, the larger the optimization space, and the greater the deviation of the optimal target value, that is, the more significant the improvement effect of the PSO algorithm combination.

[0077] Comparison and analysis of the results of the improved PSO algorithm and the dynamic programming method show that: under the condition of single unit with different load speed optimization operation, the corresponding relationship between the decision-making process of the two methods and the electricity price and head in each period is shown in Table 4. Figure 5 As shown by Figure 5 (a) It can be seen that when a single unit is running at low load, the decision-making process is mainly affected by the time-of-use electricity price, and generally the high-price unit will not be started; Figure 5 (b) and (c) show that when the time-of-use electricity price and lift are both low, the single unit is turned on and the water lifting volume is relatively large, while when the electricity price is low and the lift is high, the general unit also needs to be turned on, but the water lifting volume is relatively small at this time; Figure 5 (c) It can also be seen that when a single unit is running at full load, in order to meet the water lifting requirements, even if the electricity price and head are high, the unit generally needs to be started and operated.

[0078] Table 1 shows the electricity price and the corresponding average head for each period (average head per day 5.8m)

[0079]

[0080] Table 2 shows the single unit speed optimization decision process or startup plan calculated by three methods

[0081]

[0082] Table 3 is a comparative analysis of the unit water extraction costs of the two methods (basic PSO and improved PSO)

[0083]

[0084] Table 4 is a comparative analysis of the unit water extraction costs of the two methods (improved PSO and dynamic programming)

[0085]

[0086]

[0087] Therefore, in general, the optimization decision results obtained by the two calculation methods are basically the same, that is, during the variable speed optimization operation process, the single unit is generally not turned on when the electricity price and head are relatively high. Even if it is turned on, its speed is generally small, and the water lifting volume is relatively small; when the electricity price and head are relatively low, it is turned on and the speed is generally large, and the water lifting volume is relatively large. From the perspective of different operating loads, the variable speed optimization decision process of the single unit is greatly affected by the time-of-use electricity price. It can be seen from Table 2 that under three different operating load scenarios, the unit water lifting costs calculated according to the optimization results of the two methods are very close, and the absolute value of the deviation rate does not exceed 0.1%. It can be seen that the multi-strategy fusion improved PSO algorithm can be used as an effective method for solving the variable speed optimization model of the variable frequency pump station single unit.

Claims

1. A variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm, characterized in that: The steps include: (1) A model is established based on the data of a single unit in a pumping station. A variable frequency single unit speed optimization model is established with the goal of minimizing power consumption. The decision variable is the speed of the unit in each period, and the total amount of water pumped in a day and the operating power of the unit are used as constraints. (2) Based on the established variable frequency single unit speed optimization model, the particle swarm algorithm improved by multi-strategy fusion is used to solve the decision-making process and operating parameters of the pump station single unit; the operating parameters include the unit speed, flow, efficiency, water lifting volume, operating power consumption and average head per hour under different time periods; (3) Based on the decision-making process and operating parameters of step (2), the single unit operation of the pump station is realized.

2. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 1 is characterized in that: The variable speed optimization model of the variable frequency single unit in step (1) includes: Objective function: Constraint 1: The total amount of water pumped in one day is Constraint 2: Unit rated power constraint, i.e. N i (n i (l) )≤N e ; Wherein, F represents the power consumption cost of a single unit; i represents the time period, i=1, 2, ..., SN, SN represents the total number of time periods in a day; They represent the lth speed value and the corresponding water lifting flow rate in the i-th period respectively; H i , △T i and P i They represent the average lift, start-up time and time-of-use electricity price in the i-th period respectively; η z，i , η mot , η int , η f Respectively represent the efficiency of single unit, motor efficiency, transmission efficiency and frequency conversion device efficiency; W d is the total daily water extraction volume of a single unit; N e Indicates rated power; N i (n i (l) ) represents the actual operating power of the unit in the i-th period.

3. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 1 is characterized in that: The step (2) includes parameter setting, optimizing initial population improvement, solving fitness and updating optimal position, improving inertia weight, improving learning factor, updating position and judging iteration conditions.

4. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The parameter settings include: Set the population number and movement speed range V max and V min Unit speed range n max and n min , the current number of iterations t and the maximum number of iterations T.

5. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The optimization of the initial population improvement includes: The Sobol sequence is introduced to initialize the population. The population position formula generated by the Sobol sequence initialization is: X i =(b u -b l )*r+b l Among them, b u and b l are the upper and lower bounds of the search space respectively; r is a random number in the range of [0, 1] generated by the Sobol sequence.

6. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The fitness solution and optimal position update include: The fitness value is calculated according to the objective function; and according to the fitness value, the current position is compared with the previous optimal position pbest. If it is better, the current position is updated to the local optimal position pbest; the current position is compared with the historical optimal position gbest. If it is better, the current position is updated to the global optimal position gbest.

7. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The inertia weight improvements include: An inertia weight that decreases nonlinearly with the number of iterations is used; the change law of ω with the number of iterations is as follows: Among them, ω max ,ω m,in is the maximum and minimum value of the inertia weight; t is the current number of iterations, T is the maximum number of iterations; ω max Take 0.9, ω min Take 0.

4.

8. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The learning factor improvements include: The sine and cosine factors in the sine and cosine algorithm position update formula are introduced to replace the learning factor. At this time, the particle velocity update formula is as follows: ri=0.4(1-t / T) Where t represents the number of iterations; v li t+1 、x li t+1 They represent the velocity and position of particle l in dimension i after iteration t+1 times; ω represents the inertia weight; pbest li t represents the optimal position of particle l in dimension i after t iterations; gbest i t represents the optimal position of the entire population in dimension i after t iterations; c1 and c2 represent the learning factors of particles and populations respectively; r1 and r2 are random numbers in the range of [0, 1]; r1 controls the conversion of the algorithm from global search to local development; r2 describes the direction of individual movement, r2∈[0, 2π], and r3 is the probability between sine update and cosine update, r3∈[0, 1].

9. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The particle position of the position update is expressed as: Among them, x li t 、x li t+1 They represent the position of particle l in dimension i after iterations t and t+1 respectively.

10. The variable speed optimization scheduling method for variable frequency single unit based on multi-strategy fusion improved particle swarm algorithm according to claim 3 is characterized in that: The judgment iteration conditions include: When t≤T, repeat the steps of optimizing initial population improvement, inertia weight improvement, learning factor improvement and position update until exiting the iteration process and outputting the results, including the time-averaged lift H i , speed n at different time periods i 、Period flow Q i 、Equipment efficiency η i , water extraction volume in each period and total operating electricity consumption minF.

Images