Offshore wind power distributed generation planning method based on hybrid improved seagull algorithm

Abstract

The present invention discloses a method for planning an offshore wind power distributed power source based on a hybrid improved seagull algorithm, comprising: optimizing the basic seagull algorithm to obtain a hybrid improved seagull algorithm; initializing the parameters of the hybrid improved seagull algorithm; using the initialized hybrid improved seagull algorithm to solve the upper optimization model, obtain the distributed power site selection and capacity information and input it into the lower optimization model for solving, and obtain the distributed power planning information; returning the distributed power planning information to the upper optimization model; calculating whether the total cost meets the conditions, if so, outputting the distributed power planning information, if not, repeating the above steps. The present invention takes demand response into consideration during the planning process, effectively improves the amount of distributed energy consumption, reduces network losses, reduces annual comprehensive costs, realizes economical and efficient, safe and stable operation of the system, and has higher optimization accuracy and faster optimization speed in solving the problem of double-layer planning of offshore wind power distributed power sources.

Description

Technical Field

[0001] The present invention belongs to the technical field related to distribution network planning, and in particular relates to an offshore wind power distributed power source planning method based on a hybrid improved Seagull algorithm. Background Art

[0002] The construction of new power systems based on new energy sources is advancing rapidly, and the market-oriented reform of distribution networks is deepening, providing an opportunity for access to the system. Offshore wind power is clean, low-carbon and efficient. In addition, my country's offshore wind energy resources are large, so there is great potential for development. At the same time, various types of flexible regulation resources are increasing, and the electricity demand of power grid users is gradually showing diversified characteristics, which makes the traditional distribution network gradually transform into a multi-source, active distribution network with a higher level of observable control and active coordination and optimization management capabilities. If various resources such as offshore wind power distributed power sources and demand response can be planned in a coordinated manner under the new power system environment, and friendly and coordinated interactions can be carried out, the economy and stability of the power grid can be effectively improved. Therefore, in view of the problem of offshore wind power consumption under the new power system environment, it is of great significance to coordinate the planning of the active distribution network "source-grid-load" connected to the offshore wind power distributed power sources.

[0003] At present, the research on the collaborative planning of active distribution networks is mainly divided into two parts: model research and algorithm research. In terms of model research, the active distribution network planning model can be mainly divided into single-layer planning model and double-layer planning model. In terms of algorithm, the use of intelligent algorithms to solve the double-layer planning model is widely used. Common ones include particle swarm algorithm, genetic algorithm, gray wolf algorithm, etc. The quality of the planning result depends on the performance of the intelligent algorithm. Compared with other intelligent algorithms, the Seagull algorithm has the advantages of simple structure, strong adaptability, and fast convergence speed, but it still has defects such as single population search method, over-reliance on the optimal individual, and poor local search ability. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an offshore wind power distributed power planning method based on a hybrid improved Seagull algorithm to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides an offshore wind power distributed power planning method based on a hybrid improved seagull algorithm, comprising:

[0006] S1. Optimize the basic Seagull algorithm to obtain the hybrid improved Seagull algorithm;

[0007] S2, initialize the parameters of the hybrid improved Seagull algorithm;

[0008] S3, using the initialized hybrid improved Seagull algorithm to solve the upper optimization model and obtain the site selection and capacity determination information of distributed generation;

[0009] S4, inputting the distributed power site selection and capacity determination information into the lower optimization model for solving, and obtaining distributed power planning information;

[0010] S5. Return the distributed power planning information to the upper optimization model to calculate whether the total cost meets the conditions. If so, output the distributed power planning information. If not, repeat the above steps.

[0011] Optionally, the process of obtaining the hybrid improved Seagull algorithm in S1 includes: optimizing the algorithm structure of the basic Seagull algorithm based on collaborative evolution and optimizing the search method based on adaptive mutation.

[0012] Optionally, the process of optimizing the basic Seagull algorithm in S1 includes:

[0013] S101, setting algorithm parameters and randomly generating an initial population;

[0014] S102, calculating the fitness value of the initial population and sorting it, marking the best individual;

[0015] S103, dividing the seagull group into an elite seagull group and an ordinary seagull group according to the sorting situation, and the ratio of the elite seagulls to the ordinary seagulls is generated based on preset parameters;

[0016] S104, updating the control factor according to the cosine control strategy;

[0017] S105. Update the positions of elite seagulls and ordinary seagulls according to the dual-group co-evolution strategy;

[0018] S106, calculating the fitness value of the new population, updating the ranking of the seagull group, and updating the optimal individual;

[0019] S107, performing mutation and update operations on the optimal individual;

[0020] S108, determine whether the maximum number of iterations has been reached, if so, stop the iteration and output the optimal value, otherwise jump to step S103 to continue the iteration.

[0021] Optionally, the process of optimizing the algorithm structure of the basic Seagull algorithm based on co-evolution includes:

[0022] The fitness values ​​are sorted and the seagull group is divided into elite seagulls and ordinary seagulls. During the iteration process, the elite seagulls conduct local exploration by performing small-scale spiral movements at the original position.

[0023] Elite Seagull Position PosE i The update formula is: PosE i (t+1)=PosS i (t)×x×y×z;

[0024] The position of the common seagull is updated according to the following formula: i The update formula is:

[0025] PosS i (t+1)=Dis(t)×x×y×z+Pos best (t)

[0026] Among them, Dis(t) represents the distance between the individual seagull and the optimal seagull, Pos best (t) represents the optimal position; x, y, and z are the horizontal, vertical, and vertical coordinates, respectively;

[0027] An improved method based on the cosine function is introduced to adjust the change mode of the control factor. The improved control factor expression is:

[0028]

[0029] Among them, k c (t) represents the control factor of the Seagull algorithm at the tth iteration, k max is the maximum value of the control factor, k min is the minimum value of the control factor.

[0030] Optionally, the Cauchy mutation and Gaussian mutation are combined to obtain an adaptive mutation strategy, a mutation operation is performed based on the adaptive mutation strategy, and a greedy selection strategy is used to define an evolutionary direction for the optimal individual, wherein the adaptive mutation strategy is expressed as:

[0031] Pos best.V (t) = Pos best (t)

[0032] α=[1+γCauchy(0,1)+(1-γ)Gauss(0,1)]

[0033] Among them, Pos best.V (t) represents the position of the optimal individual after mutation, Cauchy(0,1) and Gauss(0,1) are standard Gaussian mutation and standard Cauchy mutation respectively; γ=1-t 2 / maxIter 2 is the variation control coefficient, and γ is adaptively adjusted with the number of iterations.

[0034] Optionally, the upper-layer optimization model in S3 takes minimizing the total system cost as the first objective function and adopts a first constraint condition, wherein the first constraint condition includes an access capacity constraint condition of a distributed power source, and the access capacity constraint condition of the distributed power source is:

[0035] 0≤Si ≤S i,max

[0036] Among them, S i,max is the maximum access capacity of distributed generation.

[0037] Optionally, the total system cost includes the annual investment cost of distributed power sources, the annual operation and maintenance cost of distributed power sources, the cost of purchasing electricity from superiors, the annual network loss cost and the compensation cost of interruptible loads;

[0038] The annual investment cost calculation formula of the distributed power source is:

[0039]

[0040] Among them, c inv is the investment cost of distributed power generation per unit capacity; N i is the number of nodes to be connected to the distributed power source; is the capacity of the i-th distributed generation; r and y are the discount rate and life cycle of the distributed generation, respectively;

[0041] The calculation formula for the annual operation and maintenance cost of the distributed power source is:

[0042]

[0043] Among them, N t is the total number of time periods in a day; is the actual output of the i-th distributed generation unit in period t; c m The unit operating cost of distributed power generation;

[0044] The formula for calculating the cost of purchasing electricity from the superior is:

[0045]

[0046] in, and are the transmission power and electricity purchase price in time period t respectively;

[0047] The annual network loss cost calculation formula is:

[0048]

[0049] in, is the network loss at time t, c loss The price of grid loss electricity;

[0050] The calculation formula for the compensation cost of the interruptible load is:

[0051]

[0052] Among them, NDR is the number of users on the demand side; and are the electricity consumption and compensation electricity price of the jth user in period t respectively.

[0053] Optionally, the lower-level optimization model in S4 takes minimizing network loss as the second objective function and adopts a second constraint condition; wherein the second constraint condition includes power balance constraint, node voltage constraint, branch current constraint, node voltage amplitude constraint, branch capacity constraint, and distributed power supply operation constraint.

[0054] Optionally, the network loss calculation formula is:

[0055]

[0056] in, is the branch current between node i and node j at time t, R ij is the branch resistance between node i and node j.

[0057] Compared with the prior art, the present invention has the following advantages and technical effects:

[0058] The present invention takes demand response into consideration during the planning of offshore wind power distributed power sources, which can effectively increase the consumption of distributed energy, reduce network losses, reduce annual comprehensive costs, and achieve economical, efficient, safe and stable operation of the system; the hybrid seagull algorithm takes into account both global search and local exploration, has good optimization performance, and has higher optimization accuracy and faster optimization speed in solving the problem of double-layer planning of offshore wind power distributed power sources. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0060] Figure 1 A flow chart of model solving in an embodiment of the present invention;

[0061] Figure 2 A schematic diagram of a seagull foraging for food according to an embodiment of the present invention;

[0062] Figure 3 The standard Gaussian distribution and the standard Cauchy distribution diagram of the embodiment of the present invention;

[0063] Figure 4 The convergence curves of the Seagull Algorithm, the Grey Wolf Algorithm, and the Hybrid Improved Seagull Algorithm in the standard example of the embodiment of the present invention;

[0064] Figure 5 is a simulation convergence characteristic curve of an embodiment of the present invention;

[0065] Figure 6 Schematic diagram showing the voltage distribution of each node in scenes 1 and 2 at the same time according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0067] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0068] Embodiment 1

[0069] like Figure 1 As shown, this embodiment provides an offshore wind power distributed power planning method based on a hybrid improved seagull algorithm, comprising the following steps:

[0070] S1. Initialize the parameters of the hybrid improved Seagull algorithm, and use the original data of the distribution network, system economic indicators, typical daily load, and distributed power output data as the input of the model;

[0071] S2. Solve the upper optimization model through the hybrid improved Seagull algorithm to obtain the site selection and capacity determination information of distributed generation;

[0072] (I) Upper model

[0073] The upper layer is the planning layer, which makes decisions on the location and capacity of distributed power sources from the perspective of economy to achieve the goal of minimizing the total cost of the system. Its objective function is as follows:

[0074] C all =C inv +C m +C buy +C loss +C DR

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] Where: Call is the total system cost; C inv is the annual investment cost; C m is the annual operation and maintenance cost; C s The cost of purchasing electricity from the superior; C lass is the annual network loss cost; C DR is the compensation cost for interruptible load. inv is the investment cost of distributed power generation per unit capacity; N i is the number of nodes to be connected to the distributed power source; is the capacity of the i-th distributed generation; r and y are the discount rate and life cycle of the distributed generation, respectively; N t is the total number of time periods in a day; is the actual output of the i-th distributed generation unit in period t; c m The unit operating cost of distributed power generation; and are the transmission power and electricity purchase price in time period t respectively; is the network loss at time t, c loss is the grid loss electricity price; N DR is the number of users on the demand side; and are the electricity consumption and compensation electricity price of the jth user in period t respectively.

[0081] Access capacity constraints of distributed generation:

[0082] 0≤S i ≤S i,max

[0083] Where: S i,max is the maximum access capacity of distributed generation.

[0084] (II) Hybrid Seagull Optimization Algorithm

[0085] The hybrid improved Seagull algorithm is an optimization of the traditional Seagull algorithm based on the algorithm structure optimization based on collaborative evolution and the search method optimization based on adaptive mutation. The optimization steps include:

[0086] S201, setting algorithm parameters and randomly generating an initial population;

[0087] S202, calculating the fitness value of the initial population and sorting them, marking the best individual;

[0088] S203, dividing the seagull group into an elite seagull group and an ordinary seagull group according to the ranking;

[0089] S204, updating the control factor according to the cosine control strategy;

[0090] S205. Update the positions of elite seagulls and ordinary seagulls according to the dual-group co-evolution strategy;

[0091] S206, calculating the fitness value of the new population, updating the ranking of the seagull group, and updating the optimal individual;

[0092] S207, performing mutation and update operations on the optimal individual;

[0093] S208, determine whether the maximum number of iterations has been reached, if so, stop the iteration and output the optimal value, otherwise jump to step S203 to continue the iteration.

[0094] I. Seagull Optimization Algorithm

[0095] The Seagull Algorithm is an optimization algorithm that simulates the foraging behavior of seagulls in nature. It is used to solve optimization problems. The idea of ​​the algorithm is derived from the migratory and foraging behavior of seagulls when looking for food. The migratory behavior of seagulls is to obtain sufficient and abundant food. In the process of migration, in order to quickly obtain food resources, each seagull will approach the individual with the highest fitness in the group. The foraging behavior of seagulls means that when seagulls fly over the sea, they will randomly select a starting point and then circle around this starting point until they find food or the foraging range exceeds their flight radius.

[0096] The basic seagull optimization algorithm includes random population initialization, fitness value calculation and sorting, and position update operations in the migration and foraging stages. The specific process is as follows:

[0097] (1) Random initialization

[0098] First, the location of the seagull group is expressed in the form of a two-dimensional vector group, which is:

[0099]

[0100] Where: Pos i,j represents the position component of the i-th (1≤i≤N) seagull in the j-th (1≤j≤N) dimension. Each seagull is displaced in any dimension to simulate the spatial position change of the seagull. The position initialization of each seagull can be achieved by a uniform distribution function:

[0101] Pos i,j =lb j +rand×(ub-lb)

[0102] Where: lb j andub j are the lower and upper limits of the domain of the j-th dimension, respectively, and rand is a random number in the range of [0,1] that follows a uniform distribution.

[0103] (2) Fitness calculation and ranking

[0104] By substituting the position vector of the seagull into the fitness calculation function f of the actual problem, the fitness value Fitness of the seagull group can be obtained. The calculation formula is:

[0105]

[0106] After each calculation, the fitness is sorted to select the location of the optimal solution.

[0107] (3) Migration stage

[0108] The seagull algorithm achieves global search by simulating the behavioral characteristics of seagulls during their migration. During the migration of seagulls, in order to avoid collisions between individuals, each seagull will choose a new location to ensure that there is no conflict with other seagulls. To achieve this, the algorithm controls the factor k c To calculate the new position of each seagull:

[0109] Pos new (t) = k c ×Pos(t)

[0110] Where: Pos new (t) is the new position of the seagull without conflict with other seagulls, Pos(t) is the position of the seagull before the migration period, and t is the number of iterations. Control factor k c The calculation formula is as follows:

[0111]

[0112] Where: maxIter is the maximum number of iterations.

[0113] After avoiding collisions with other seagulls, each seagull will look for the direction of the optimal position and move towards that direction. best The expression of (t) is:

[0114] Dir best (t) = k b ×[Pos best (t)-Pos(t)]

[0115] Where: Pos best (t) represents the optimal position; k b is the search balance factor, which is expressed as follows:

[0116]

[0117] Where: r d is a random number in the range [0,1] that follows a uniform distribution.

[0118] The distance Dis(t) between an individual seagull and the optimal seagull is:

[0119] Dis(t)=|Pos new (t)+Dir best (t)|

[0120] (4) Foraging stage

[0121] The local development phase in the Seagull algorithm is similar to the behavior of seagulls when foraging. This phase occurs in the second half of the search process and aims to use search history and experience to adjust the search scope and improve search accuracy. Figure 2 A schematic diagram of it.

[0122] In the Seagull algorithm, the local development phase also uses search history and experience to more accurately find the optimal solution by adjusting the search scope and strategy. Its movement trajectory is as follows:

[0123] x=r×cosθ

[0124] y=r×sinθ

[0125] z=r×θ

[0126] r=u×e θv

[0127] Where r is the radius of the spiral flight of the seagull in the air; θ is a random number in the range [0,2π] that obeys a uniform distribution, indicating the attack angle of the seagull; u and v are spiral coefficients, both of which are constants of 1; x, y, and z are the horizontal, vertical, and vertical coordinates, respectively.

[0128] Its foraging location is:

[0129] Pos(t+1)=Dir best (t)×x×y×z+Pos best (t)

[0130] Where: Pos(t+1) is the new position after iteration.

[0131] As an emerging swarm intelligence optimization algorithm, the Seagull algorithm has the advantages of strong global search capability, few algorithm parameters, parallel computing, and strong adaptability. However, due to its over-reliance on population initialization, single search method, and easy to fall into local optimality, it is necessary to optimize its population initialization, search method, and parameters to obtain better optimization results.

[0132] II. Algorithm structure optimization based on co-evolution

[0133] In terms of algorithm structure, this embodiment introduces the idea of ​​co-evolution, optimizes the population update mechanism of the traditional seagull algorithm, and proposes a dual-group co-evolution strategy. This strategy first divides the seagull group into elite seagulls and ordinary seagulls according to the fitness value, where the proportion of elite seagulls is determined by the preset parameter R elite to control.

[0134] During the iteration process, the elite seagulls use a local search method to perform local exploration by performing a small-scale spiral motion at the original position, while the ordinary seagulls continue to use the position update formula of the original algorithm.

[0135] Elite Seagull Position PosE i The update formula is:

[0136] PosE i (t+1)=PosS i (t)×x×y×z

[0137] Common Seagull Position PosS i The update formula is:

[0138] PosS i (t+1)=Dis(t)×x×y×z+Pos best (t)

[0139] Such a dual-group co-evolution strategy can better balance the capabilities of local search and global search, thereby improving the optimization performance and convergence speed of the Seagull algorithm.

[0140] Control factor k c It plays a role in balancing global exploration and local development. c In order to solve the problem that the linear change does not match the actual search situation, this embodiment introduces an improved method based on the cosine function to adjust k c The cosine function has good smoothness and periodicity in periodic changes, which will help to make the change of the control factor more stable and better adapt to the search process of the population. The improved control factor expression is:

[0141]

[0142] Where: k c (t) represents the control factor of the Seagull algorithm at the tth iteration, k max is the maximum value of the control factor, k min is the minimum value of the control factor.

[0143] III. Search method optimization based on adaptive mutation

[0144] In terms of search methods, this embodiment proposes an adaptive mutation seagull optimization strategy based on the combination of Gaussian mutation and Cauchy mutation to enrich the search methods of the population. The formula of Gaussian mutation operation is as follows:

[0145] P i G =P i (1+Gauss(0,σ 2 ))

[0146] Where: P i G represents the position after Gaussian mutation operation; P i is the position before mutation; Gauss(0,σ 2 ) is a Gaussian mutation operator, which is used to generate a random number that obeys the standard normal distribution.

[0147] The formula for the Cauchy mutation operation is as follows:

[0148] P i C =P i (1+Cauchy(0,ρ))

[0149] Where: P i C Represents the position after the Cauchy mutation operation; Cauchy(0,ρ) is the Cauchy mutation operator, which generates a random number that obeys the standard Cauchy distribution.

[0150] The standard normal distribution curve and the standard Cauchy distribution curve are as follows Figure 3 As shown, it can be seen that the local search ability of Gaussian mutation is stronger, but the global search ability is relatively weak; the Cauchy mutation has better global search ability due to its wider distribution range and is easier to jump out of the local trap.

[0151] According to the different advantages and characteristics of Cauchy mutation and Gaussian mutation, this embodiment combines these two mutation methods and introduces an adaptive mutation strategy, which acts on the optimal seagull:

[0152] Pos best.V (t) = Pos best (t)

[0153] α=[1+γCauchy(0,1)+(1-γ)Gauss(0,1)]

[0154] Where: Pos best.V (t) represents the position of the optimal seagull after mutation, Cauchy (0,1) and Gauss (0,1) are standard Gauss mutation and standard Cauchy mutation respectively; γ = 1-t 2 / maxIter2 is the variation control coefficient, and γ is adaptively adjusted with the number of iterations, so that the algorithm focuses on global exploration in the early stage and local development in the later stage.

[0155] The mutated position is not necessarily better than the original position. A greedy selection strategy is needed to define a good evolutionary direction for the optimal seagull. The position update formula is:

[0156]

[0157] S3. The distributed generation site selection and capacity determination information obtained by the upper optimization model is input into the lower optimization model, and solved by the interior point method to obtain the distributed generation consumption, system network loss, power purchase from the upper level, and demand response.

[0158] The lower layer is the operation layer, which optimizes the system operation status through the coordinated control of the actual output of distributed power sources and demand response, with the minimum network loss as the objective function, and its expression is:

[0159]

[0160] Where: is the branch current between node i and node j at time t, R ij is the branch resistance between node i and node j.

[0161] The lower optimization model considers the system operation status and satisfies the following constraints:

[0162] (1) Power balance constraints

[0163]

[0164] Where: and Respectively represent the active and reactive power values ​​of node i; U i,t and U j,t are the voltage amplitudes of nodes i and j respectively; G ij and B ij is the admittance between node i and node j; θ ij is the phase angle difference between node i and node j.

[0165] (2) Node voltage constraints

[0166]

[0167] Where: and Represent the voltage amplitude of nodes i and j respectively, r ij and x ij is the impedance between node i and node j.

[0168] (3) Branch current constraints

[0169]

[0170] (4) Node voltage amplitude constraint

[0171]

[0172] Where: and They represent the upper and lower limits of the voltage amplitude at node i respectively.

[0173] (5) Branch capacity constraints

[0174]

[0175] Where: L ij represents the capacity of the line between node i and node j; represents the rated capacity of the line between node i and node j.

[0176] (6) Distributed power generation operation constraints

[0177]

[0178] S4. The information obtained by the lower optimization model is returned to the upper optimization model to calculate the total system cost and determine whether the termination conditions are met. If so, the planning result is output; otherwise, return to S2.

[0179] To prove the effectiveness of the method, the offshore wind power distributed generation two-layer planning model and the hybrid seagull optimization algorithm proposed in this embodiment are simulated and analyzed through the IEEE33-node distribution network example. The number of distributed generation access points is set to 3, and the upper limit of the capacity of a single distributed generation is 1500kVA. The economic indicators of the system are shown in Table 1.

[0180] Table 1

[0181]

[0182] First, the Grey Wolf Optimizer (Grey Wolf Optimizer) widely used in the electrical field is compared with the traditional Seagull algorithm and the hybrid improved Seagull algorithm proposed in this embodiment in a single peak test example. The population size of the three optimization methods is set to 30, the maximum number of iterations is 500, and the convergence condition is set to. The optimization results of the three optimization methods are as follows: Figure 4 As shown in the figure, compared with the Seagull algorithm, which requires 400 iterations to converge, the Grey Wolf algorithm requires 36 iterations, while the Hybrid Improved Seagull Algorithm only requires 7 iterations to converge.

[0183] The three algorithms were subjected to 30 optimization experiments, and the results are shown in Table 2. As can be seen from the table, compared with the Seagull algorithm, the optimal value of the Gray Wolf algorithm is smaller, indicating that it has higher optimization accuracy, and the mean value and standard deviation are also smaller, indicating that it has a faster convergence speed. The hybrid improved Seagull algorithm has better optimization performance than the Gray Wolf algorithm.

[0184] Table 2

[0185]

[0186] The Grey Wolf Algorithm with better optimization performance and the Hybrid Improved Seagull Algorithm are compared in solving AND-BPM optimization problems. To ensure the fairness of the comparison, the maximum number of iterations and the population size of the two algorithms are set to 75 and 30 respectively. The optimization results are shown in Table 3, and the simulation convergence characteristic curve is shown in Figure 5 As shown in the figure, the Grey Wolf Algorithm fell into local convergence at the 37th iteration and could not obtain the optimal result. The hybrid improved Seagull Algorithm finally converged at the 44th iteration, and the total cost of the optimized system was reduced by 120,600 yuan (0.71%) compared with the Grey Wolf Algorithm. The total cost of the optimized system was lower under the same number of iterations, showing better optimization ability.

[0187] Table 3

[0188]

[0189] In order to analyze the impact of demand response on the total system cost and network loss, two different schemes are set up and the optimal solution is obtained by solving the hybrid improved Seagull algorithm.

[0190] Solution 1: Based on the classic scenario, select the site and capacity of distributed power sources;

[0191] Solution 2: Based on the classic scenario and taking demand response into consideration, the site selection and capacity determination of distributed power sources are carried out.

[0192] The optimization results of the two different schemes are shown in Table 4. Comparing Scheme 1 and Scheme 2, it can be seen that after the demand response is connected, the consumption of distributed power generation is increased by 3.25MWh (7.9%). Through the allocation and scheduling of distributed power generation and demand response, the uncertainty problem of new energy generation is effectively solved, thereby improving the flexibility and efficiency of the system; the network loss is reduced by 6.8MWh (6.13%), and the overall energy efficiency of the system is improved; the total system cost is reduced by 1.44 million yuan (7.8%), which effectively improves the economy of the system.

[0193] Table 4

[0194]

[0195] Figure 6The voltage distribution of each node in scenario 1 and scenario 2 at the same time. As can be seen from the figure, the voltage amplitude deviation of each node considering demand response is smaller, which can effectively improve the power quality of the system and reduce the risk of equipment damage, thereby achieving safe and stable operation of the system.

[0196] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for planning distributed power sources for offshore wind power based on a hybrid improved seagull algorithm, characterized in that: The following steps are involved: S1. Optimize the basic Seagull algorithm to obtain the hybrid improved Seagull algorithm; S2, initialize the parameters of the hybrid improved Seagull algorithm; S3, using the initialized hybrid improved Seagull algorithm to solve the upper optimization model and obtain the site selection and capacity determination information of distributed generation; The upper-level optimization model in S3 takes the minimum total system cost as the first objective function and adopts the first constraint condition, wherein the first constraint condition includes the access capacity constraint condition of the distributed power source, and the access capacity constraint condition of the distributed power source is: 0≤S i ≤S i,max Among them, S i,max is the maximum access capacity of distributed power sources; The total system cost includes the annual investment cost of distributed power generation, the annual operation and maintenance cost of distributed power generation, the cost of purchasing electricity from the superior, the annual network loss cost and the compensation cost of interruptible load; The annual investment cost calculation formula of the distributed power source is: Among them, c inv N is the investment cost of distributed power generation per unit capacity; i is the number of nodes to be connected to the distributed power source; is the capacity of the i-th distributed generation; r and y are the discount rate and life cycle of the distributed generation, respectively; The calculation formula for the annual operation and maintenance cost of the distributed power source is: Among them, N t is the total number of time periods in a day; is the actual output of the i-th distributed generation unit in period t; c m The unit operating cost of distributed power generation; The formula for calculating the cost of purchasing electricity from the superior is: in, and are the transmission power and electricity purchase price in time period t respectively; The annual network loss cost calculation formula is: in, is the network loss at time t, c loss The price of grid loss electricity; The calculation formula for the compensation cost of the interruptible load is: Among them, N DR is the number of users on the demand side; and are the electricity consumption and compensation price of the jth user in period t respectively; S4, inputting the distributed power site selection and capacity determination information into the lower optimization model for solving, and obtaining distributed power planning information; The lower optimization model in S4 takes minimizing network loss as the second objective function and adopts the second constraint condition; wherein the second constraint condition includes power balance constraint, node voltage constraint, branch current constraint, node voltage amplitude constraint, branch capacity constraint, and distributed power supply operation constraint; S5. Return the distributed power planning information to the upper optimization model to calculate whether the total cost meets the conditions. If so, output the distributed power planning information. If not, repeat the above steps.

2. The offshore wind power distributed power planning method based on the hybrid improved seagull algorithm according to claim 1 is characterized in that: The process of obtaining the hybrid improved Seagull algorithm in S1 includes: optimizing the algorithm structure of the basic Seagull algorithm based on collaborative evolution and optimizing the search method based on adaptive mutation.

3. The offshore wind power distributed power planning method based on the hybrid improved seagull algorithm according to claim 2 is characterized in that: The process of optimizing the basic Seagull algorithm in S1 includes: S101, setting algorithm parameters and randomly generating an initial population; S102, calculating the fitness value of the initial population and sorting it, marking the best individual; S103, dividing the seagull group into an elite seagull group and an ordinary seagull group according to the sorting situation, and the ratio of the elite seagulls to the ordinary seagulls is generated based on preset parameters; S104, updating the control factor according to the cosine control strategy; S105. Update the positions of elite seagulls and ordinary seagulls according to the dual-group co-evolution strategy; S106, calculating the fitness value of the new population, updating the ranking of the seagull group, and updating the optimal individual; S107, performing mutation and update operations on the optimal individual; S108, determine whether the maximum number of iterations has been reached, if so, stop the iteration and output the optimal value, otherwise jump to step S103 to continue the iteration.

4. The offshore wind power distributed power planning method based on the hybrid improved seagull algorithm according to claim 3 is characterized in that: The process of optimizing the algorithm structure of the basic Seagull algorithm based on co-evolution includes: The fitness values ​​are sorted and the seagull group is divided into elite seagulls and ordinary seagulls. During the iteration process, the elite seagulls conduct local exploration by performing small-scale spiral movements at the original position. Elite Seagull Position PosE i The update formula is: PosE i (t+1)=PosS i (t)×x×y×z; The position of the common seagull is updated according to the following formula: i The update formula is: PosS i (t+1)=Dis(t)×x×y×z+Pos best (t) Among them, Dis(t) represents the distance between the individual seagull and the optimal seagull, Pos best (t) represents the optimal position; x, y, and z are the horizontal, vertical, and vertical coordinates, respectively; An improved method based on the cosine function is introduced to adjust the change mode of the control factor. The improved control factor expression is: Among them, k c (t) represents the control factor of the Seagull algorithm at the tth iteration, k max is the maximum value of the control factor, k min is the minimum value of the control factor.

5. The offshore wind power distributed power planning method based on the hybrid improved seagull algorithm according to claim 3 is characterized in that: The adaptive mutation strategy is obtained by combining Cauchy mutation and Gaussian mutation. The mutation operation is performed based on the adaptive mutation strategy, and the greedy selection strategy is used to define the evolution direction for the optimal individual. The adaptive mutation strategy is expressed as: Pos best.V (t)=Pos best (t) α=[1+γCauchy(0,1)+(1-γ)Gauss(0,1)] Among them, Pos best.V (t) represents the position of the optimal individual after mutation, Cauchy(0,1) and Gauss(0,1) are standard Gaussian mutation and standard Cauchy mutation respectively; γ=1-t 2 / maxIter 2 is the variation control coefficient, and γ is adaptively adjusted with the number of iterations.

6. The offshore wind power distributed power planning method based on the hybrid improved seagull algorithm according to claim 1 is characterized in that: The network loss calculation formula is: in, is the branch current between node i and node j at time t, R ij is the branch resistance between node i and node j.

Images