Vehicle charging disturbance resisting elastic power system operation method and elastic power system

Abstract

The invention discloses an elastic electric power system operation method and an elastic electric power system.The elastic electric power system operation method for resisting vehicle charging disturbance comprises the following steps that 1, node loads are divided into basic loads and electric vehicle charging loads, respectively establishing removal loss influence evaluation models of the basic load and the electric vehicle charging load; the cut-off loss of the charging load of the electric vehicle comprises instant loss and subsequent loss; step 2, based on the optimal power flow problem model, designing and setting a target function by taking the minimum load shedding loss as a target; 3, a mixed dragonfly-particle swarm optimization algorithm is adopted for solving, and a load removal scheme is obtained; the method has the advantages of being fast in solving and small in influence caused by load shedding.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and specifically relates to a flexible power system operation method and a flexible power system capable of resisting vehicle charging disturbances. Background Art

[0002] In recent years, extreme disturbance accidents in power systems have occurred frequently. Disturbance accidents will continue to spread and worsen within the system, and may eventually cause serious losses. In order to enable the power system to have the ability to adapt to disturbances, the traditional power system needs to evolve into a resilient power system. Resilience refers to the ability of the power grid to resist, absorb and quickly recover from low-probability, high-impact disturbance accidents. The improvement of resilience needs to be designed in combination with factors that have a direct impact on the system. Today, the large-scale penetration of electric vehicle charging demand has increased the operating pressure of the power system, and there is a rising trend. How to effectively reduce the impact of electric vehicle charging loads and improve the resilience of the power system has become an urgent problem to be solved.

[0003] In the prior art, the specific difficulty in designing a resilient power system that can resist disturbances caused by charging power of electric vehicles lies in the fact that load shedding is often used in the disturbance resistance process, but there is a lack of assessment of the impact of load shedding on losses. If load shedding is performed blindly, some loads may be excessively shedding, resulting in paralysis of corresponding social functions. Summary of the invention

[0004] The purpose of the present invention is to provide a flexible power system operation method and a flexible power system that can resist vehicle charging disturbances in response to the problems existing in the prior art. First, a load shedding loss impact assessment model is established to analyze the electric vehicle charging load, and the impact loss of the electric vehicle charging load is divided into immediate loss and subsequent loss according to the time and method of affecting travel users. Then, based on the optimal power flow model, a disturbance resistance scheme with the minimum load shedding loss as the goal is designed, and a hybrid dragonfly-particle swarm algorithm is proposed, which can be solved in a short time and reduce the impact caused by load shedding.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a method for operating a flexible power system to resist vehicle charging disturbance, comprising the following steps:

[0006] Step 1: Divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss;

[0007] Step 2, based on the optimal power flow problem model, the objective function is set with the minimum load shedding loss as the goal design;

[0008] Step 3: Use the hybrid dragonfly-particle swarm optimization algorithm to solve and obtain the load shedding plan.

[0009] In the above scheme, step 1 distinguishes the types of node loads and establishes a corresponding removal loss impact assessment model to conduct targeted assessments, and subdivides the removal loss of electric vehicle charging loads into immediate loss and subsequent loss to more accurately reflect the impact of load removal loss. In step 2, the minimum load removal loss is achieved under the constraints of satisfying the normal operation of the node and various safety indicators. Step 3 is solved by the hybrid dragonfly-particle swarm optimization algorithm to obtain the load removal scheme.

[0010] Furthermore, the hybrid dragonfly-particle swarm optimization algorithm comprises the following steps:

[0011] Step 3.1, set the particle population, dragonfly population and maximum number of iterations, traverse the dragonfly population and calculate the individual fitness through the objective function, select the individual with the highest fitness as the initial optimal position, and enter the iteration process;

[0012] Step 3.2, use the dragonfly algorithm to obtain the solution space of the historical optimal position;

[0013] Step 3.3, using the particle swarm algorithm to update the historical optimal position obtained by the dragonfly algorithm;

[0014] Step 3.4, increase the number of iterations. If the maximum number of iterations is reached, return to the global optimal position; if not, return to step 3 and enter the next iteration.

[0015] Furthermore, the basic load removal loss impact assessment model includes: calculating the loss impact of basic load removal according to the basic load value vector that needs to be removed at each node in the distribution network at time t, the total demand basic load value of distribution network node i at time t, the minimum basic load demand value, the distribution network node set, the duration of the disturbance resistance process, and the loss cost, where t and i are set constants.

[0016] The loss impact of base load removal is calculated based on the parameters in the model using the formula, which facilitates intuitive evaluation and judgment of the impact size.

[0017] Furthermore, the instantaneous loss impact of the electric vehicle charging load removal is calculated based on the following parameters: the electric vehicle charging load removal value vector of each traffic network node under the distribution network, the numerical value of the electric vehicle charging load that needs to be removed by the charging station of the traffic network node coupled to the distribution network node i at time t, the number of traffic network nodes coupled to the distribution network, a judgment variable for judging whether there is a charging station at the traffic network node coupled to the distribution network node i, the active state of the k-th charging pile in the charging station of the traffic network node coupled to the distribution network node i at time t, the set of all charging piles in the charging station of the traffic network node coupled to the distribution network node i, the instantaneous loss cost of the electric vehicle user served by the k-th charging pile in the station at time t, and the output power size that needs to be adjusted for the k-th charging pile in the charging station of the traffic network node coupled to the distribution network node i at time t.

[0018] Furthermore, the subsequent loss impact of the electric vehicle charging load removal is calculated based on the following parameters: a numerical function of the reduction in travel capacity after the charging load is removed from the charging station selected for stopping during the starting point-destination trip, and a starting point-destination trip distribution of the traveling users within the disturbance resistance time range; wherein the starting point-destination trip is the optimal path for electric vehicle travel.

[0019] The above parameters are used to calculate the immediate loss impact and subsequent loss impact of electric vehicle charging load removal, which is convenient for intuitive evaluation and judgment of the impact size and subsequent processing.

[0020] Furthermore, the relevant constraints include:

[0021]

[0022] Among them, P i,t , and They represent the active power injected into the distribution network node i at time t, the active power output of the power generation, the active power demand of the load and the active power value of the removed load, respectively. i,t , and They represent the reactive power injected by the distribution network node i at time t, the reactive power output of the generator, the reactive power demand of the load and the reactive power value of the removed load, respectively. i,t represents the voltage amplitude of the distribution grid node i at time t, Ω DN With Ω DL Respectively represent the nodes and lines in the distribution network, P ab,t With Q ab,t They represent the active power flow and reactive power flow carried by the distribution network line (a, b) at time t, respectively. i With B i They represent the parallel conductance and parallel susceptance of the distribution network node i, Gab , B ab With θ ab,t Respectively represent the conductance, susceptance and voltage phase difference of the distribution network line (a, b), where θ ab,t =θ a,t -θ b,t ,θ a,t With θ b,t They represent the phase angles of the nodes at both ends of the line, and T is the protection time range.

[0023] Combine the above constraints to generate the objective function and screen out individuals with high fitness.

[0024] Furthermore, the formula for moving and updating the position of the dragonfly algorithm in the search space in step 3.2 is:

[0025]

[0026] Where Levy(d) is a d-dimensional column value vector, r1 and r2 are random numbers in the interval [0,1], and β is a custom constant. Indicates that at t i The current position of particle p at iteration .

[0027] Furthermore, the formula for adjusting the position and speed in each iteration of the particle swarm algorithm in step 3.3 is:

[0028]

[0029] in Indicates that at t i The velocity vector of particle p at the iteration, Indicates that at t i The individual best historical position of particle p at the iteration, Indicates that at t i The current position of particle p at iteration , Indicates that at t i The best historical position of the population at the iteration, Indicates that at t i The inertia weight at the iteration, C1 and C2 are the self-learning factor and the group learning factor respectively, rand1 and rand2 represent random numbers in the range, and the particle position is adjusted by moving in the direction of the speed in this iteration based on the position of the previous iteration.

[0030] Furthermore, step 3.1 includes the following steps: initializing the position, velocity and optimal historical position of the particle group, initializing the position and velocity of the dragonfly group, traversing the dragonfly population to calculate the individual fitness through the objective function, selecting the individual with the highest fitness as the initial optimal position, and entering the iterative process.

[0031] The present invention also discloses a flexible power system for resisting automobile charging disturbance, comprising:

[0032] The first main module is used to divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss;

[0033] The second main module is used to set the objective function based on the optimal power flow problem model and to design the minimum load shedding loss as the goal;

[0034] The third main module is used to solve the problem using a hybrid dragonfly-particle swarm optimization algorithm and obtain a load shedding solution.

[0035] The present invention also discloses an operating device for a flexible power system capable of resisting automobile charging disturbances, comprising a memory and a processor, wherein the memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the steps of the above method.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. In the process of load shedding, the loss impact of different types of load shedding is considered, and an evaluation model for the loss impact of electric vehicle charging load shedding is established, which can more realistically reflect the loss impact caused by load shedding and more effectively guide load shedding. The method proposed in the present invention can effectively distinguish the loss impact of different types of load shedding, and can design different load shedding schemes according to different disturbance scenarios;

[0038] 2. A hybrid dragonfly-particle swarm algorithm is proposed, which can overcome the defects of the particle swarm algorithm that will cause it to fall into the local optimum and the dragonfly algorithm will take up more time and computing resources. It can ensure that the travel of electric vehicle users is not overly affected. In addition, the hybrid dragonfly-particle swarm algorithm has certain advantages in solution accuracy and solution feasibility, and has a fast solution speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 : Workflow diagram provided by an embodiment of the present invention;

[0040] Figure 2 : Schematic diagram of the IEEE 33-node power distribution system provided by an embodiment of the present invention;

[0041] Figure 3 : Schematic diagram of a distribution network coupled with a transportation network provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The Dragonflies Algorithm (DA) is a new meta-heuristic optimization technique. This algorithm is derived from the static and dynamic clustering behaviors of dragonflies in nature, including their social interactions when searching for food, navigating, and avoiding predators. The Dragonflies Algorithm simulates these behaviors and designs two basic stages: optimization, exploration, and exploitation.

[0043] Algorithm concept: The dragonfly algorithm simulates five behaviors of dragonflies: separation, alignment, cohesion, food source attraction, and enemy distraction. These behaviors are described by specific mathematical models and used to guide the movement of dragonflies in the search space.

[0044] Particle Swarm Optimization (PSO), also known as particle swarm algorithm, comes from the simulation of a simplified social model. Particles are a compromise choice because it is necessary to describe the members of the group as having no mass and no volume, and also to describe its speed and acceleration state.

[0045] The Optimal Power Flow (OPF) model refers to the power flow distribution that makes a certain performance indicator of the system (such as power generation cost or network loss) reach the optimal value by adjusting the available control variables (such as generator output power and adjustable transformer taps) under given power system structural parameters and load conditions, while satisfying all operating constraints. The Optimal Power Flow model is a typical nonlinear optimization problem. Due to the complexity of its constraints, the calculation is difficult. The mathematical description of the Optimal Power Flow model usually includes the objective function and constraints.

[0046] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention are arbitrarily combined with each other to form a new technical solution. This combination is not restricted by the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in this field to achieve. When the combination of technical solutions is contradictory or cannot be achieved, it should be considered that this combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0047] like Figure 1 As shown, a method for operating a flexible power system to resist vehicle charging disturbances includes the following steps:

[0048] Step 1: Divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss;

[0049] Step 2, based on the optimal power flow problem model, the objective function is set with the minimum load shedding loss as the goal design;

[0050] Step 3: Use the hybrid dragonfly-particle swarm optimization algorithm to solve and obtain the load shedding plan.

[0051] In the above scheme, step 1 distinguishes the types of node loads and establishes a corresponding removal loss impact assessment model to conduct targeted assessments, and subdivides the removal loss of electric vehicle charging loads into immediate loss and subsequent loss to more accurately reflect the impact of load removal loss. In step 2, the minimum load removal loss is achieved under the constraints of satisfying the normal operation of the node and various safety indicators. Step 3 is solved by the hybrid dragonfly-particle swarm optimization algorithm to obtain the load removal scheme.

[0052] Furthermore, the hybrid dragonfly-particle swarm optimization algorithm comprises the following steps:

[0053] Step 3.1, set the particle population, dragonfly population and maximum number of iterations, traverse the dragonfly population and calculate the individual fitness through the objective function, select the individual with the highest fitness as the initial optimal position, and enter the iteration process;

[0054] Step 3.2, use the dragonfly algorithm to obtain the solution space of the historical optimal position;

[0055] Step 3.3, using the particle swarm algorithm to update the historical optimal position obtained by the dragonfly algorithm;

[0056] Step 3.4, increase the number of iterations. If the maximum number of iterations is reached, return to the global optimal position; if not, return to step 3 and enter the next iteration.

[0057] In the above scheme, the initialization operation is performed before step 3.1. In step 3.1, the parameters related to the system operation are set, and then the individual with the highest fitness is selected as the initial optimal position through the objective function calculation; step 3.2 starts the iteration process, first using the dragonfly algorithm to obtain the historical optimal position, and then using the particle swarm algorithm to update the historical optimal position obtained in step 3.2 through step 3.3. In step 3.4, when it is judged that the maximum number of iterations has not been reached, the updated historical optimal position of step 3.3 is substituted into step 3.2 for iteration. When the maximum number of iterations is reached, the updated historical optimal position of step 3.3 is returned to the global optimal position. This solution contains the optimal load shedding scheme, which can know which nodes are selected for load shedding and the specific amount of shedding, so as to achieve the goal of minimizing load shedding losses.

[0058] Using the formula, the loss impact of base load removal is calculated according to the parameters in the model, which is convenient for intuitive evaluation and judgment of the impact. The loss impact model of base load removal is:

[0059]

[0060] in, is the basic load value that needs to be removed at each node of the distribution network at time t The vector composed of is the total demand base load value of distribution network node i at time t, is the minimum base load requirement value. Ω DN is the set of distribution network nodes, T is the discrete time set considered in the disturbance resistance process until the system fully recovers the pre-disturbance stable level, and |T| is the duration of the disturbance resistance process. The loss cost is expressed as the sum of the product of the VoLL of each node and the removal duration. In order to facilitate the unification of the dimensions of other indicators defined later, the function The loss impact of base load removal is defined in terms of maximum and minimum normalization.

[0061] Furthermore, the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss. The disturbance resistance strategy is designed with the minimum load removal loss as the goal.

[0062] In order to unify the dimensions, the present invention normalizes the function, and the instantaneous loss impact function of the removal of the electric vehicle charging load is for:

[0063]

[0064] in, is the electric vehicle charging load removal value vector of each traffic network node under the distribution network, expressed as It represents the value of the electric vehicle charging load that needs to be removed by the charging station of the traffic network node coupled to the distribution network node i at time t, |Ω DN | is the number of traffic network nodes coupled to the distribution network, SE i To determine whether there is a charging station in the traffic network node coupled to the distribution network node i, Indicates the active state of charging pile No. k in the charging station of the traffic network node coupled to the distribution network node No. i at time t, represents the set of all charging piles in the charging station of the traffic network node coupled to the distribution network node i, The instant loss cost of an electric vehicle user served by charging pile No. k in the station at time t is: Indicates the output power that needs to be adjusted for charging pile No. k in the charging station (if any) of the traffic network node coupled to distribution network node No. i at time t.

[0065] The present invention defines the optimal path plan for electric vehicle users as the origin-destination (OD) trip, and the subsequent loss impact function of the electric vehicle charging load removal described in step 1 is for:

[0066]

[0067] in, For the starting point-destination trip with n as the starting point and m as the destination, the charging station selected for the stop is removed. Numerical function of travel capacity reduction after charging load, CN T is the distribution of the starting point-destination trips of the users within the disturbance resistance time range T. This is a square matrix with the number of nodes of the traffic network coupled with the distribution network as the scale. Its matrix elements are It means the number of electric vehicle users who travel from the starting point n to the destination m within the time range T.

[0068] The above parameters are used to calculate the immediate loss impact and subsequent loss impact of electric vehicle charging load removal, which is convenient for intuitive evaluation and judgment of the impact size and subsequent processing.

[0069] Furthermore, the relevant constraints include:

[0070]

[0071] Among them, P i,t , and They represent the active power injected into the distribution network node i at time t, the active power output of the power generation, the active power demand of the load and the active power value of the removed load, respectively. i,t , and They represent the reactive power injected by the distribution network node i at time t, the reactive power output of the generator, the reactive power demand of the load and the reactive power value of the removed load, respectively. i,t represents the voltage amplitude of the distribution grid node i at time t, Ω DN With Ω DL Respectively represent the nodes and lines in the distribution network, P ab,t With Q ab,t They represent the active power flow and reactive power flow carried by the distribution network line (a, b) at time t, respectively. i With B i They represent the parallel conductance and parallel susceptance of the distribution network node i, G ab , B ab With θ ab,t Respectively represent the conductance, susceptance and voltage phase difference of the distribution network line (a, b), where θ ab,t =θ a,t -θ b,t ,θ a,t With θ b,t Respectively represent the phase angle of the nodes at both ends of the line, V i min and V i max They represent the minimum and maximum voltage amplitudes of the distribution network node i, respectively, i Gmin , P i Gmax , They represent the minimum and maximum values ​​of active power and reactive power output of distribution network node i, respectively. and They respectively represent the minimum and maximum values ​​of the active line flow and the minimum and maximum values ​​of the reactive line flow that the distribution network lines (a, b) can withstand.

[0072] The penalty function of the inequality constraint described in step 2 is:

[0073] l i (x,u)=max(0,-g i (x,u)) g i (x,u)≤0 (14)

[0074] The penalty function for the equality constraint described in step 2 is:

[0075] l j (x,u)=max(0,|h j(x,u)|-ε) h j (x,u)=0 (15)

[0076] The objective function with penalty function is set in combination with relevant constraints as described in step 2:

[0077]

[0078] Among them, the inequality constraint F I The various lim values ​​in are:

[0079]

[0080] Furthermore, the formula for moving and updating the position of the dragonfly algorithm in the search space in step 3.2 is:

[0081]

[0082] Where Levy(d) is a d-dimensional column value vector, r1 and r2 are random numbers in the interval [0,1], and β is a custom constant. Indicates that at t i The current position of particle p at iteration .

[0083] Define the step vector for each iteration The calculation method is:

[0084]

[0085] Among them, s, a, c, f and e are the weights of the five behaviors of separation, parallel, aggregation, attraction by food and staying away from enemies, respectively. t is the inertia weight, which is calculated as follows:

[0086]

[0087] F p =X + -X p (twenty three)

[0088] E p =X - +X p (twenty four)

[0089]

[0090] Among them, Iter and Iter max are the current number of iterations and the maximum number of iterations respectively. The updating process of the position vector is:

[0091]

[0092] Furthermore, the formula for adjusting the position and speed in each iteration of the particle swarm algorithm in step 3.3 is:

[0093]

[0094] in Indicates that at t i The velocity vector of particle p at the iteration, Indicates that at t i The individual best historical position of particle p at the iteration, Indicates that at t i The current position of particle p at iteration , Indicates that at t i The best historical position of the population at the iteration, Indicates that at t i The inertia weight at the iteration, C1 and C2 are the self-learning factor and the group learning factor respectively, rand1 and rand2 represent random numbers in the range, and the particle position is adjusted by moving in the direction of the speed in this iteration based on the position of the previous iteration.

[0095] Furthermore, step 3.1 includes the following steps: initializing the position, velocity and optimal historical position of the particle group, initializing the position and velocity of the dragonfly group, traversing the dragonfly population to calculate the individual fitness through the objective function, selecting the individual with the highest fitness as the initial optimal position, and entering the iterative process.

[0096] The present invention also discloses a flexible power system for resisting automobile charging disturbance, comprising:

[0097] The first main module is used to divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss;

[0098] The second main module is used to set the objective function based on the optimal power flow problem model and to design the minimum load shedding loss as the goal;

[0099] The third main module is used to solve the problem using a hybrid dragonfly-particle swarm optimization algorithm and obtain a load shedding solution.

[0100] The present invention also discloses an operating device for a flexible power system capable of resisting automobile charging disturbances, comprising a memory and a processor, wherein the memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the steps of the above method.

[0101] Preferably, step 2 includes setting various default parameters such as the basic load removal loss impact weight and the electric vehicle charging load removal loss impact weight. Obtain the current distribution network operation data, including the basic load and electric vehicle charging load power requirements of each node, the minimum power requirement, etc., obtain the basic load removal loss impact function at the current moment according to the node basic load power demand, obtain the current electric vehicle charging load removal immediate loss impact function and subsequent loss impact function according to the traffic network topology and user travel characteristics, and set the objective function with penalty function in combination with relevant constraints.

[0102] As a preferred method, the number of active electric vehicles in the area is 5,000 and the configuration parameters are the same. The battery capacity is set to 10kW·h, the power consumption per kilometer is 0.2kW·h / km, and the battery capacity at the start of daily travel is 70% of the full capacity. The purpose of EV travel is strictly in accordance with the travel chain. It is set that there is only a simple chain (HPH) in the area, which includes residential area-industrial area-residential area and residential area-commercial area-residential area. The ratio of these two travel chains is 52.8% and 47.2%. When setting the starting point and destination of electric vehicle users, the travel chain is first confirmed and then randomly selected with equal probability in the corresponding regional nodes according to the travel chain. The travel timeout threshold is set to 1.3 times the ideal starting point-destination travel time. The simulation cycle is 24 hours a day, the sampling time interval is 15 minutes, and the maximum number of iterations is 1000. During the solution process, the penalty factor of the objective function is 800. The self-learning factor C1 and the group learning factor C2 of the particle swarm algorithm are both 0.5. The inertia weight decreases with the increase of the number of iterations, and the update method is shown in formula (1.1); the behavior weights and inertia weights in the dragonfly algorithm also decrease with the increase of the number of iterations. The behavior weight update method is shown in formula (1.2), where rand(1) represents a random number in (0,1], and the inertia weight is the same as the particle swarm algorithm. The PSO population is set to 100 and the DA population is set to 25. The population consists of two parts: the basic load removal vector and the electric vehicle charging load removal vector. The maximum number of iterations is 500.

[0103]

[0104] As a preferred method, the dragonfly algorithm is used in step 3.2 to obtain the solution space. First, the position with the best fitness of the previous generation of dragonfly population is used as the food position, and the position with the worst fitness is used as the enemy position. The neighbor radius, the weight of each behavior, and the inertia weight are updated according to the current number of iterations. Traverse the dragonfly population, update the behavior parameters of the corresponding individuals, and determine whether there are neighbors within the radius. If so, update the step vector using formula (1.3) and update the individual position using formula (1.4); if not, update the individual position using formula (1.5);

[0105]

[0106] Calculate the fitness based on the current individual position and update the historical optimal position. After traversing all M dragonfly individuals, obtain the historical optimal position obtained in this iteration

[0107] As a preferred method, in step 3.3, the particle swarm algorithm is used to update the historical optimal position obtained by the dragonfly algorithm, and the speed and position of the particle swarm are updated by substituting into equations (1.6) and (1.7);

[0108]

[0109] Traverse the particle swarm individuals to calculate fitness and update the individual's historical optimal position After completion, compare the highest particle fitness obtained in this iteration with the historical optimal fitness. If it is better, replace the historical optimal position X gbest .

[0110] In this embodiment, the distribution network uses an improved IEEE 33-node distribution system, which is connected to a certain transmission network. The active power and reactive power of each node load are adjusted based on the original data. The specific node electrical data are shown in Table 1, and the topological structure is shown in Figure 2 . The traffic network uses Figure 3 In the road distribution structure of a certain city shown in the figure, each node is equipped with a charging station, and each charging station is equipped with 15 charging piles with a maximum charging power of 25kW. The roads between nodes are all two-way streets, and the impact of road congestion and traffic flow on vehicle driving speed is not considered. The lengths of various roads are shown in Table 2. There is power supply coupling between the distribution network and the traffic network, in which a single distribution network node will only supply electricity to a single traffic network node area at most. The detailed node correspondence is shown in Table 3. In addition, in order to make the basic load power of the distribution system closer to reality, the present invention sets the node basic load power to a time-varying state, with the initial electrical data as the reference value. The changes in power at each moment are shown in Table 4.

[0111] Table 1: IEEE 33-node power distribution system node electrical information

[0112] Node Number Power demand (MW) Maximum voltage (pu) Minimum voltage (pu) 1 0 1.1 0.9 2 0.5 1.1 0.9 3 0.45 1.1 0.9 4 0.6 1.1 0.9 5 0.3 1. 1 0.9 6 0.3 1.1 0.9 7 1 1.1 0.9 8 1 1.1 0.9 9 0.3 1.1 0.9 10 0.3 1.1 0.9 11 0.225 1.1 0.9 12 0.3 1.1 0.9 13 0.3 1.1 0.9 14 0.6 1.1 0.9 15 0.3 1.1 0.9 16 0.3 1.1 0.9 17 0.3 1.1 0.9 18 0.45 1.1 0.9 19 0.45 1.1 0.9 20 0.45 1.1 0.9 21 0.45 1.1 0.9 22 0.45 1.1 0.9 23 0.45 1.1 0.9 24 2.1 1.1 0.9 25 2.1 1.1 0.9 26 0.3 1. 1 0.9 27 0.3 1.1 0.9 28 0.3 1.1 0.9 29 0.6 1.1 0.9 30 1 1.1 0.9 31 0.75 1.1 0.9 32 1.05 1.1 0.9 33 0.3 1.1 0.9

[0113] Table 2: Road length information of a certain traffic road network

[0114] Starting point end Road length (km) Starting point end Road length (km) 1 2 10.2 13 15 10.5 1 4 10.4 14 15 9.5 2 3 10.3 14 28 9.7 2 5 10.5 15 16 8.6 2 6 10.6 15 28 8.8 3 4 10.4 16 17 10.7 3 7 10.7 16 29 8.9 4 8 10.8 17 30 11.0 4 9 9.9 18 19 8.9 5 6 10.6 18 23 10.3 5 23 10.3 18 27 9.7 6 7 9.7 19 20 10.0 6 24 9.4 19 24 10.4 7 8 10.8 19 28 8.8 7 25 9.5 20 21 9.1 8 10 10.0 20 25 10.5 8 26 9.6 20 29 8.9 9 10 9.9 21 22 10.2 9 11 10.1 21 25 10.5 10 11 9.0 22 30 11.0 10 21 10.1 23 24 10.4 11 22 10.2 24 25 9.5 12 13 10.3 25 26 9.6 12 16 10.6 27 28 9.8 12 17 10.7 28 29 8.9 13 14 10.4 29 30 11.0

[0115] Table 3: Correspondence between distribution network nodes and traffic network nodes

[0116] Power distribution node Traffic nodes Power distribution node Traffic nodes Power distribution node Traffic nodes 1 1 12 12 23 4 2 2 13 13 24 9 3 3 14 14 25 11 4 6 15 15 26 25 5 24 16 16 27 26 6 19 17 29 28 10 7 20 18 28 29 8 8 21 19 5 30 7 9 22 20 23 31 - 10 30 21 18 32 - 11 17 22 27 33 -

[0117] Table 4: Ratio of load power in each area compared to the baseline value from 4 to 24 o'clock

[0118] Time (hr) Residential area Industrial Area Commercial District time Residential area Industrial Area Commercial District 4 0.226 0.483 0.417 14.25 0.403 0.914 0.729 4.25 0.226 0.483 0.417 14.5 0.403 0.914 0.729 4.5 0.226 0.483 0.417 14.75 0.403 0.914 0.729 4.75 0.226 0.483 0.417 15 0.403 1 0.833 5 0.226 0.552 0.479 15.25 0.403 1 0.833 5.25 0.226 0.552 0.479 15.5 0.403 1 0.833 5.5 0.226 0.552 0.479 15.75 0.403 1 0.833 5.75 0.226 0.552 0.479 16 0.419 0.966 0.813 6 0.226 0.621 0.479 16.25 0.419 0.966 0.813 6.25 0.226 0.621 0.479 16.5 0.419 0.966 0.813 6.5 0.226 0.621 0.479 16.75 0.419 0.966 0.813 6.75 0.226 0.621 0.479 17 0.565 0.966 0.938 7 0.323 0.67 0.396 17.25 0.565 0.966 0.938 7.25 0.323 0.67 0.396 17.5 0.565 0.966 0.938 7.5 0.323 0.67 0.396 17.75 0.565 0.966 0.938 7.75 0.323 0.67 0.396 18 0.726 0.862 0.896 8 0.355 0.759 0.458 18.25 0.726 0.862 0.896 8.25 0.355 0.759 0.458 18.5 0.726 0.862 0.896 8.5 0.355 0.759 0.458 18.75 0.726 0.862 0.896 8.75 0.355 0.759 0.458 19 0.935 0.759 0.979 9 0.403 0.931 0.438 19.25 0.935 0.759 0.979 9.25 0.403 0.931 0.438 19.5 0.935 0.759 0.979 9.5 0.403 0.931 0.438 19.75 0.935 0.759 0.979 9.75 0.403 0.931 0.438 20 1 0.67 0.958 10 0.484 0.948 0.521 20.25 1 0.67 0.958 10.25 0.484 0.948 0.521 20.5 1 0.67 0.958 10.5 0.484 0.948 0.521 20.75 1 0.67 0.958 10.75 0.484 0.948 0.521 21 0.919 0.586 1 11 0.403 1 0.708 21.25 0.919 0.586 1 11.25 0.403 1 0.708 21.5 0.919 0.586 1 11.5 0.403 1 0.708 21.75 0.919 0.586 1 11.75 0.403 1 0.708 22 0.645 0.534 0.938 12 0.403 0.914 0.542 22.25 0.645 0.534 0.938 12.25 0.403 0.914 0.542 22.5 0.645 0.534 0.938 12.5 0.403 0.914 0.542 22.75 0.645 0.534 0.938 12.75 0.403 0.914 0.542 23 0.484 0.517 0.521 13 0.387 0.897 0.469 23.25 0.484 0.517 0.521 13.25 0.387 0.897 0.469 23.5 0.484 0.517 0.521 13.5 0.387 0.897 0.469 23.75 0.484 0.517 0.521 13.75 0.387 0.897 0.469 24 0.403 0.5 0.479 14 0.403 0.914 0.729

[0119] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for operating a flexible power system to resist vehicle charging disturbances, characterized in that: The steps include: Step 1: Divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss; Step 2, based on the optimal power flow problem model, the objective function is set with the minimum load shedding loss as the goal design; Step 3: Use the hybrid dragonfly-particle swarm optimization algorithm to solve and obtain the load shedding plan.

2. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 1, characterized in that: The hybrid dragonfly-particle swarm optimization algorithm comprises the following steps: Step 3.1, set the particle population, dragonfly population and maximum number of iterations, traverse the dragonfly population and calculate the individual fitness through the objective function, select the individual with the highest fitness as the initial optimal position, and enter the iteration process; Step 3.2, use the dragonfly algorithm to obtain the solution space of the historical optimal position; Step 3.3, using the particle swarm algorithm to update the historical optimal position obtained by the dragonfly algorithm; Step 3.4, increase the number of iterations, and return to the global optimal position if the maximum number of iterations is reached; If not reached, go back to step 3 and enter the next iteration.

3. The method for operating a flexible power system against vehicle charging disturbances according to claim 1, characterized in that: The basic load removal loss impact assessment model includes: calculating the loss impact of basic load removal according to the basic load value vector that needs to be removed at each node in the distribution network at time t, the total demand basic load value of distribution network node i at time t, the minimum basic load demand value, the distribution network node set, the duration of the disturbance resistance process, and the loss cost, where t and i are set constants.

4. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 1, characterized in that: The instantaneous loss impact of the electric vehicle charging load removal is calculated based on the following parameters: the electric vehicle charging load removal value vector of each traffic network node under the distribution network, the numerical value of the electric vehicle charging load that needs to be removed by the charging station of the traffic network node coupled to the distribution network node i at time t, the number of traffic network nodes coupled to the distribution network, a judgment variable for judging whether there is a charging station at the traffic network node coupled to the distribution network node i, the active state of the k-th charging pile in the charging station of the traffic network node coupled to the distribution network node i at time t, the set of all charging piles in the charging station of the traffic network node coupled to the distribution network node i, the instantaneous loss cost of the electric vehicle user served by the k-th charging pile in the station at time t, and the output power size that needs to be adjusted for the k-th charging pile in the charging station of the traffic network node coupled to the distribution network node i at time t.

5. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 1, characterized in that: The subsequent loss impact of the removal of the electric vehicle charging load is calculated based on the following parameters: a numerical function of the travel capacity reduction after the charging load is removed from the charging station selected for the stop during the start-destination journey, and the start-destination journey distribution of the travel users within the disturbance resistance time range; The starting point-destination journey is the optimal route for electric vehicles.

6. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 1, characterized in that: In step 2, the objective function is set in combination with constraints, wherein the constraints include: Among them, P i,t , and They represent the active power injected into the distribution network node i at time t, the active power output of the power generation, the active power demand of the load and the active power value of the removed load, respectively. i,t , and They represent the reactive power injected by the distribution network node i at time t, the reactive power output of the generator, the reactive power demand of the load and the reactive power value of the removed load, respectively. i,t represents the voltage amplitude of the distribution grid node i at time t, Ω DN With Ω DL Respectively represent the nodes and lines in the distribution network, P ab,t With Q ab,t They represent the active power flow and reactive power flow carried by the distribution network line (a, b) at time t, respectively. i With B i They represent the parallel conductance and parallel susceptance of the distribution network node i, G ab , B ab With θ ab,t Respectively represent the conductance, susceptance and voltage phase difference of the distribution network line (a, b), where θ ab，t =θ a，t -θ b，t ,θ a,t With θ b,t They represent the phase angles of the nodes at both ends of the line, and T is the protection time range.

7. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 2, characterized in that: The formula for moving and updating the position of the dragonfly algorithm in the search space in step 3.2 is: Where Levy(d) is a d-dimensional column value vector, r1 and r2 are random numbers in the interval [0,1], and β is a custom constant. Indicates that at t i The current position of particle p at iteration .

8. The method for operating a flexible power system to resist vehicle charging disturbance according to claim 2, characterized in that: The formula for adjusting position and speed in each iteration of the particle swarm algorithm in step 3.3 is: in Indicates that at t i The velocity vector of particle p at the iteration, Indicates that at t i The individual best historical position of particle p at the iteration, Indicates that at t i The current position of particle p at iteration , Indicates that at t i The best historical position of the population at the iteration, Indicates that at t i The inertia weight at the iteration, C1 and C2 are the self-learning factor and the group learning factor respectively, rand1 and rand2 represent random numbers in the range, and the particle position is adjusted by moving in the direction of the speed in this iteration based on the position of the previous iteration.

9. A resilient power system for resisting automobile charging disturbances, characterized in that: include: The first main module is used to divide the node load into two categories: basic load and electric vehicle charging load, and establish the removal loss impact assessment model of the basic load and the electric vehicle charging load respectively; the removal loss of the electric vehicle charging load includes two parts: immediate loss and subsequent loss; The second main module is used to set the objective function based on the optimal power flow problem model and to design the minimum load shedding loss as the goal; The third main module is used to solve the problem using a hybrid dragonfly-particle swarm optimization algorithm and obtain a load shedding solution.

10. An operating device for a flexible power system that resists automobile charging disturbances, characterized in that: The method comprises a memory and a processor, wherein the memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the steps of the method according to any one of claims 1 to 8.

Images