Active decoupling optimal cross-section search method considering system primary frequency response characteristics

Abstract

The present invention belongs to the technical field of power system stability analysis, and specifically discloses a method for searching for the optimal section of active uncoupling taking into account the primary frequency response characteristics of the system. First, with the goal of minimizing the number of machine cuts, load cuts and disconnected lines and combining the relevant basic operating constraints of the system, a basic optimization solution model for active uncoupling is established by considering the coherence and connectivity constraints of the units. On this basis, a model of the primary frequency response characteristics of the system is established, and the frequency stability-related constraints of the system frequency such as the frequency change rate and the lowest frequency point are combined to ensure the frequency stability of the island operation after uncoupling. At the same time, the interruptible demand response load on the load side is modeled to improve the system frequency response characteristics and improve the system frequency stability. In order to cope with the adverse effects of the access of new energy to the power grid, a comprehensive model for searching for the optimal section of active uncoupling taking into account the uncertainty of the output of new energy is established to improve the stability of each island after uncoupling.

Description

Technical Field

[0001] The invention belongs to the technical field of power system stability analysis, and in particular relates to an active unloading optimal section search method taking into account the primary frequency response characteristics of the system. Background Art

[0002] Under the background of rapid development of power system, there is an obvious linkage effect between the construction of large power network and the interconnection between regional power grids. With the access of new energy, the inertia of the system is reduced, which greatly reduces its anti-interference performance. When an emergency occurs, it is more likely to expand local events, causing large-scale power outages, which will cause great damage to society and economy. In the stable operation of power system, active disconnection is the main means to prevent large-scale power outages and large-scale power outages in the power grid. It is the ultimate defense means to ensure the continuous safe and stable operation of the power grid within a certain range. The key is to quickly find the optimal disconnection section under the premise of meeting the corresponding operation constraints.

[0003] Determining a feasible active unloading strategy is a very complex process. It must meet the stable operating conditions and various dynamic constraints of all the islands formed by unloading to ensure that all the islands formed by unloading can operate stably. The key task is to determine the appropriate unloading section. Since the existing research results basically aim to minimize the unbalanced power, the impact of the cost of disconnecting the line is not considered. In order to increase the frequency support capacity of the system during the unloading process and reduce the economic losses of cutting the machine and load, it is also necessary to consider the primary frequency response capacity of each subsystem after unloading. At the same time, with the access of new energy, it is also necessary to comprehensively establish an active unloading section search model that takes into account the uncertainty of new energy output in order to deal with different output scenarios.

[0004] Therefore, it is of great significance to study and establish an active decoupling optimal section search method that can overcome the above-mentioned defects and take into account the primary frequency response characteristics of the system and the uncertainty of new energy. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method for searching the optimal section that takes into account the primary frequency response characteristics of the system and active disconnection. By using the optimization method, the goal is to minimize the number of machine cuts and load cuts and the number of disconnections. The primary frequency response model of the system is considered to regulate the unbalanced power of the island after disconnection. The frequency stability-related constraints such as the frequency change rate and the lowest frequency point are combined to ensure the frequency stability of the system operation after disconnection. At the same time, the load-side demand response model is introduced to reduce the interruptible load and reduce the shortage caused by the island, which is beneficial to the frequency stability of the island after disconnection. On this basis, the uncertainty of the output of new energy is comprehensively considered, and an optimal disconnection section search model considering the uncertainty of new energy is established and solved to cope with the impact of the uncertainty of the output of new energy on the system after disconnection.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] An active decoupling optimal section search method considering the primary frequency response characteristics of the system comprises the following steps:

[0008] Step 1: Establish a MILP model based on the objective function and basic constraints of active disconnection of the power system, and input the relevant operating data obtained when the power system is unstable into the MILP model to solve and obtain the initial active disconnection section set;

[0009] Step 2: Considering the primary frequency response capability of the system, a primary frequency response model is established, including the constraints of the lowest frequency point and the highest frequency point, the primary frequency response reserve capacity constraint, and the frequency change rate constraint, and the initial active disconnection section set is constrained and screened to obtain the intermediate disconnection section set;

[0010] Step 3: Model the interruptible power load, introduce the demand response model, filter the intermediate disconnection section set, and obtain the active disconnection optimal section set considering the primary frequency response characteristics of the system;

[0011] Step 4: Considering the uncertainty of renewable energy output, construct renewable energy output and its probability model, and obtain a complete set of optimal sections for active decoupling that considers the primary frequency response characteristics of the system and the uncertainty of renewable energy.

[0012] Furthermore, the MILP model established in step 1 includes:

[0013] (1) The objective function is:

[0014]

[0015] Where: Ω L is the set of lines; Ω D is the set of loads; Ω G is the set of generators; ΩW is a collection of wind farms; is the cost of disconnecting the line; d is the load shedding penalty coefficient; g is the machine-cutting penalty coefficient; w is the penalty coefficient for cutting off the wind farm; x ij is a binary variable, indicating the connectivity status of branch ij, x ij = 0 means that branch ij is disconnected by decoupling, x ij =1 means branch ij is connected; and They are the actual load shedding amount d of island k, the generator shedding amount g and the wind farm shedding amount w;

[0016] (2) Generator coherence clustering model:

[0017]

[0018]

[0019]

[0020]

[0021] Where: Ω N represents a set of nodes; Ω K represents the set of different islands generated by decoupling; i and j represent the first and last nodes of branch l respectively; x i,k 、x j,k Represent the state variables of nodes i and j respectively. If x i,k =1,x j,k =1, which means that nodes i and j belong to the island k; if x i,k =0,x j,k =0, which means nodes i and j do not belong to island k; t ij,k Represents auxiliary variables, used to linearize two 0-1 variables x i,k and x j,k multiply;

[0022] (3) System connectivity model after generator synchronization grouping

[0023]

[0024] -M·x ij ≤P′ ij ≤M·x ij

[0025] P′ ref ≥1

[0026] Where: P′ij Represents the active power of the virtual branch ij; P′ jk represents the active power of virtual branch jk; P′ ref Represents the active power generated by the virtual generator, M is a sufficiently large number; Ω j Represents a set of devices connected to node j.

[0027] Furthermore, the MILP model also includes:

[0028] The MILP model also includes:

[0029] (1) Node balance constraints:

[0030]

[0031]

[0032] 0≤v d ≤P d

[0033]

[0034]

[0035] Where: P g , Q g Respectively represent the active and reactive output of generator g; P w represents the dispatch value of wind farm w; P d , Q d They represent the active and reactive power of load d after considering the load side demand response; v d 、v g and v w They represent the power imbalance of load d, generator g and wind farm w in the island respectively; P ij , Q ij are respectively the active and reactive power flows of branch ij; Q b Represents the reactive power dispatch value of capacitor group b; Q s Represents the reactive power dispatch value of the static VAR compensator s; Indicates the power factor of load d; P jk and Q jk Indicates the active and reactive power flows of branch jk; Indicates the maximum output of generator g; It represents the predicted output value of wind farm w;

[0036] (2) Branch flow constraints:

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045] Where: (·) * are the initial values ​​of the variables given for linearizing the AC power flow; Respectively represent the maximum transmission capacity of active and reactive power of the line; v i and v j are the voltage amplitudes of nodes i and j, respectively, ij is the phase angle difference; b ij , g ij are the susceptance and conductance of branch ij respectively; As auxiliary variables, the AC power flow model is used for calculation; and Respectively represent the active and reactive losses of branch ij;

[0046] (3) Generator output constraints:

[0047]

[0048]

[0049]

[0050] Where: and They represent the upper and lower limits of the active output of generator g respectively; and They represent the upper and lower limits of the reactive power output of the generator g respectively; ε represents the adjustment range when the generator is disconnected; Indicates the output of the generator before disconnection;

[0051] (4) Constraints on new energy output:

[0052]

[0053]

[0054] In the formula, σ represents the regulation range when the wind farm is disconnected; Indicates the output of the wind farm before disconnection;

[0055] (5) Node voltage and node phase angle constraints:

[0056]

[0057]

[0058] Where: and Respectively represent the upper and lower limits of the voltage amplitude of node i; θ i is the phase angle of node i; θ ref Indicates the voltage phase angle value of the reference node;

[0059] (6) Constraints on capacitor banks and static VAR compensators:

[0060]

[0061]

[0062]

[0063] Where: Ω b is the collection of capacitor banks; Ω s is a collection of static VAR compensators; Represents the unit reactive power of capacitor bank b; y b is an integer variable representing the number of capacitor banks put into use, subject to the maximum number of restrictions; and They respectively represent the upper and lower limits of the reactive power dispatch of the static VAR compensator.

[0064] Furthermore, the primary frequency response model in step 2 includes:

[0065] (1) Constraints on the lowest and highest frequency points:

[0066]

[0067]

[0068]

[0069]

[0070] Where: It represents the surplus of power generation over load in island k; It indicates the shortfall of power generation in island k below the load; and They respectively represent the upper reserve of the primary frequency response of the generator when a power shortage occurs and the lower reserve of the primary frequency response of the generator when a power surplus occurs; Indicates the lowest frequency point when power shortage occurs; Indicates the highest frequency point when power surplus occurs; c g Indicates the governor ramp rate of generator g; H k represents the equivalent inertia of island k; f0 represents the rated frequency of the system; f db Indicates the dead zone of the speed regulator;

[0071] (2) Primary frequency response reserve capacity constraint:

[0072]

[0073]

[0074]

[0075]

[0076] Where: and They represent the maximum upper reserve and maximum lower reserve of the primary frequency response of the generator respectively; is the decision variable of the cutting machine, Indicates the removal of the generator;

[0077] (3) Frequency change rate constraint:

[0078]

[0079]

[0080] Where: γ max Indicates the maximum operating frequency change rate of the generator set; H g represents the inertia time constant of the generator g.

[0081] Furthermore, the primary frequency response model in step 2 also includes other constraints:

[0082] The surplus of island k after system disconnection and generator shedding or shortfall The model is as follows:

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097] Where: It represents the surplus after the load is cut off by the k island cutting machine; represents the power deficit after the k island is cut off and the load is shed; δ is a binary variable, which limits the relationship between the power surplus and the power deficit in the island; and is the decision variable of the cutting machine, Indicates the removal of the generator. Indicates the removal of wind farms; P w Represents the dispatch value of wind farm w.

[0098] Furthermore, the demand response model described in step 3 is as follows:

[0099]

[0100]

[0101] Where: and They represent the predicted load value of load d and the load value participating in demand response respectively; α d Indicates the load d interruptible load ratio.

[0102] Furthermore, the overall optimization model expression considering the uncertainty of new energy in step 4 is specifically as follows:

[0103]

[0104] Where: vector x represents the binary variables in the model; vector y represents the continuous variables in the model; a and c represent the coefficient vectors of the variables to be solved in the objective function; A, B, C represent matrices in abstract form, h and f represent vectors in abstract form, representing the coefficients of the relevant constraints; ρ ξ is the probability of scene ξ appearing; y ξ is a continuous variable in different scenarios.

[0105] Compared with the prior art, the present invention has the following beneficial effects:

[0106] 1) The model of the present invention considers the situation that the primary frequency response reserve of the generator in the island after decoupling is used to balance the surplus or shortage of the island, and limits the frequency change rate and the lowest frequency point of the system to ensure the frequency stability of the system. The decoupling scheme obtained can restore each island to a stable operating state under the condition of minimum load shedding of the total generator.

[0107] 2) Reducing part of the load by responding to a certain proportion of load-side demand can not only reduce the amount of load shedding, making the resulting splitting solution better, but also effectively reduce the power shortage in each island after splitting, which is more conducive to ensuring the stability of the frequency of each island.

[0108] 3) The model takes into account the uncertainty of renewable energy. Generators can balance the surplus or shortage in different output scenarios through primary frequency response and work together with the loads participating in demand response to cope with the uncertainty of their output. The proposed model can obtain the optimal decoupling result under the operating conditions of renewable energy output uncertainty. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 It is a flow chart of the steps of the method of the present invention.

[0110] Figure 2 The modified New England 39-node system structure diagram used to verify the present invention.

[0111] Figure 3 1 is a comparison diagram of the transient process of the New England 39-node system after decoupling after modification in Example 1 of the present invention; (a) is the generator power angle curve before decoupling; (b) is the bus voltage curve before decoupling; (c) is the generator frequency curve before decoupling; (d) is the generator power angle curve after decoupling; (e) is the bus voltage curve after decoupling; (f) is the generator frequency curve after decoupling.

[0112] Figure 4 1 is a comparison of the frequency curves of island 2 in Example 2 and Example 3 of the present invention.

[0113] Figure 51 and 10 are frequency curves of the generator G5 in different scenarios after the decoupling of Example 2 and Example 7 of the present invention; (a) is the frequency curve of the generator G5 in Example 7 after the decoupling; (b) is the frequency curve of the generator G5 in different scenarios after the decoupling of Example 2.

[0114] Figure 6 is the demand response participation amount of each scenario in Example 8 of the present invention; (a) is the demand response participation amount of each scenario in Island 1; (b) is the demand response participation amount of each scenario in Island 2. DETAILED DESCRIPTION

[0115] In order to fully illustrate the technical solution disclosed by the present invention, the present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0116] The present invention discloses a method for searching the optimal cross section of active decoupling taking into account the primary frequency response characteristics of the system and the uncertainty of new energy. The specific implementation steps are as follows: Figure 1 As shown, the technical solution of the present invention comprises the following steps:

[0117] Step 1: According to the objective function and basic constraints of active disconnection of the power system, a MILP (Mixed integer linear programming) model is established, and the relevant operating data obtained when the power system is unstable is input into the MILP model to solve the initial active disconnection section set.

[0118] (1.1)Objective function:

[0119]

[0120] Where: Ω L is the set of lines; Ω D is the set of loads; Ω G is the set of generators; Ω W is a collection of wind farms; is the cost of disconnecting the line; d is the load shedding penalty coefficient; g is the machine-cutting penalty coefficient; w is the penalty coefficient for cutting off the wind farm; x ij is a binary variable, indicating the connectivity status of branch (i, j). If x ij = 0, indicating that branch (i, j) is disconnected by decoupling, otherwise, x ij =1, the branch is connected; and They are the actual load shedding amount d of island k, the generator shedding amount g and the wind farm shedding amount w.

[0121] (1.2) Generator coherence clustering model:

[0122]

[0123]

[0124]

[0125]

[0126] Where: Ω N represents a set of nodes; Ω K represents the set of different islands generated by decoupling; i and j represent the first and last nodes of branch l respectively; x i,k 、x j,k Represent the state variables of nodes i and j respectively. If x i,k =1,x j,k = 1, then it means that nodes i and j belong to the points in the island k. If x i,k =0,x j,k = 0, it means that nodes i and j do not belong to island k; x ij Indicates the disconnection status of the line; t ij,k Represents auxiliary variables, used to linearize two 0-1 variables x i,k and x j,k Multiply.

[0127] (1.3) System connectivity model after generator synchronization grouping:

[0128]

[0129] -M·x ij ≤P′ ij ≤M·x ij

[0130] P′ ref ≥1

[0131] Where: P′ ij Represents the active power of the virtual branch; P′ ref It represents the active power generated by the virtual generator. Its value is greater than or equal to 1 to ensure that there are at least two nodes in each island to prevent only one isolated generator node in the island.

[0132] (1.4) Node balance constraint:

[0133]

[0134]

[0135] 0≤v d ≤P d

[0136]

[0137]

[0138] Where: Ω j P represents a series of devices connected to node j; g , Q g Respectively represent the active and reactive output of generator g; P w represents the dispatch value of wind farm w; P d , Q d They represent the active and reactive power of load d after considering the load side demand response; v d 、v g and v w They represent the power imbalance of load d, generator g and wind farm w in the island respectively; is the predicted output value of wind farm w; P ij , Q ij are the active and reactive power flows of branch (i, j) respectively; Q b Represents the reactive power dispatch value of capacitor group b; Q s Represents the reactive power dispatch value of the static VAR compensator s; Indicates the power factor of load d.

[0139] (1.5) Branch flow constraints:

[0140]

[0141]

[0142]

[0143]

[0144]

[0145]

[0146]

[0147]

[0148] Where: M is a sufficiently large number; (·) * are the initial values ​​of the variables given for linearizing the AC power flow; Respectively represent the maximum transmission capacity of active and reactive power of the line; v i and v j are the voltage amplitudes of nodes i and j, respectively, ijis the phase angle difference; b ij , g ij are the susceptance and conductance of branch (i, j) respectively; As auxiliary variables, the AC power flow model is used for calculation; and Respectively represent the active and reactive line losses of branch (i, j);

[0149] (1.6) Generator output constraints:

[0150]

[0151]

[0152]

[0153] Where: and They represent the upper and lower limits of the active output of generator g respectively; and They represent the upper and lower limits of the reactive power output of the generator g respectively; ε represents the adjustment range when the generator is disconnected; Indicates the output of the generator before de-energization.

[0154] (1.7) Constraints on new energy output:

[0155]

[0156]

[0157] In the formula, σ represents the regulation range when the wind farm is disconnected; Indicates the output of the wind farm before disconnection.

[0158] (1.8)Node voltage and node phase angle constraints:

[0159]

[0160]

[0161] In the formula, and Respectively represent the upper and lower limits of the voltage amplitude of node i; θ i is the phase angle of node i; θ ref Indicates the voltage phase angle value of the reference node.

[0162] (1.9) Constraints on capacitor banks and static VAR compensators:

[0163]

[0164]

[0165]

[0166] In the formula, Ω b is the collection of capacitor banks; Ω s is a collection of static VAR compensators; Represents the unit reactive power of capacitor bank b; y b is an integer variable representing the number of capacitor banks put into use, subject to the maximum number of restrictions; and They respectively represent the upper and lower limits of the reactive power dispatch of the static VAR compensator.

[0167] Step 2: Consider the primary frequency response capability of the system and establish primary frequency response constraints, including the frequency minimum and maximum point constraints, primary frequency response reserve capacity constraints, and frequency change rate constraints. After constraint screening of the initial active disconnection section set, the intermediate disconnection section set is obtained.

[0168] (2.1) Constraints on the lowest and highest frequency points:

[0169]

[0170]

[0171]

[0172]

[0173] Where: It represents the surplus of power generation over load in island k; It indicates the shortfall of power generation in island k below the load; and They respectively represent the upper reserve of the primary frequency response of the generator when a power shortage occurs and the lower reserve of the primary frequency response of the generator when a power surplus occurs; Indicates the lowest frequency point when power shortage occurs; Indicates the highest frequency point when power surplus occurs; c g Indicates the governor ramp rate of generator g; H k represents the equivalent inertia of island k.

[0174] (2.2) Primary frequency response reserve capacity constraint:

[0175]

[0176]

[0177]

[0178]

[0179] Where: and Respectively represent the maximum upper and lower reserves of the primary frequency response of the generator; P g Indicates the output of generator g; Indicates the maximum output of generator g; is the decision variable of the cutting machine, Indicates that the generator is removed.

[0180] (2.3) Rate of change of frequency (RoCoF) constraint:

[0181]

[0182]

[0183] Where: γ max Indicates the maximum operating frequency change rate of the generator set; H g represents the inertia time constant of the generator g.

[0184] (2.4) Other constraints: The surplus of island k after system decoupling and generator shedding or shortfall The model is as follows:

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194]

[0195]

[0196]

[0197]

[0198]

[0199] Where: It represents the surplus after the load is cut off by the k island cutting machine; represents the power deficit after the k island is cut off and the load is shed; δ is a binary variable, which limits the relationship between the power surplus and the power deficit in the island; and is the decision variable of the cutting machine, Indicates the removal of the generator. Indicates the removal of wind farms.

[0200] Step 3: Model the interruptible power load, introduce the demand response model, filter the intermediate disconnection section set, and obtain the optimal section set for active disconnection that considers the primary frequency response characteristics of the system.

[0201]

[0202]

[0203] Where: and They represent the predicted load value of load d and the load value participating in demand response respectively; α d Indicates the load d interruptible load ratio.

[0204] Step 4: Considering the uncertainty of renewable energy output, construct renewable energy output and its probability model, and obtain a complete set of optimal sections for active decoupling that considers the primary frequency response characteristics of the system and the uncertainty of renewable energy.

[0205]

[0206] Where: vector x represents the binary variables in the model; vector y represents the continuous variables in the model; a and c represent the coefficient vectors of the variables to be solved in the objective function; A, B, C represent matrices in abstract form, h and f represent vectors in abstract form, representing the coefficients of the relevant constraints; ρ ξ is the probability of scene ξ appearing; y ξ is a continuous variable in different scenarios.

[0207] The effects of the present invention are described in detail below through specific embodiments.

[0208] (1) Example introduction.

[0209] like Figure 2As shown in the figure, a modified New England 39-node system is used to construct an active power system decoupling example. The system includes 10 generators, 46 branches, and 150MW wind turbines are connected at nodes 5 and 19. The test tool uses Matlab2016a programming software and GUROBI8.1 commercial solver.

[0210] (2) Implementation scenario description.

[0211] In order to verify the effectiveness of the proposed optimal unloading section search model and the influence of primary frequency response constraints and demand response on it, the following examples 1-3 are set; in order to verify the influence of wind power generation uncertainty on the unloading results and the effectiveness of the proposed model, examples 4-6 are obtained after taking into account the uncertain wind power output scenario on the basis of examples 1-3.

[0212] Example 1: Only consider the system decoupling under the AC power flow model;

[0213] Example 2: Based on Example 1, consider the primary frequency response characteristics of the system;

[0214] Example 3: Based on Example 2, load side demand response is considered.

[0215] Example 4: Based on Example 1, the uncertain wind power output scenario is taken into account;

[0216] Example 5: Based on Example 2, the uncertain wind power output scenario is taken into account;

[0217] Example 6: Based on Example 3, the uncertain wind power output scenario is taken into account;

[0218] (3) Analysis of the results of the implementation examples.

[0219] Table 1 gives the results of the load shedding of Examples 1-3, including the load shedding section, unbalanced power, generator shedding, and load shedding. Table 2 further gives the results of Examples 2 and 3, comparing the impact of the load participating in demand response on the surplus and deficit of each island, and gives the maximum value of the frequency change rate of each island calculated. Figure 3 The dynamic simulation results before and after decomposition of Example 2 are shown in Figure 2. Figure 4 The frequency curves of the generator in island 2 in Example 2 and Example 3 are compared. It can be concluded that the system that loses stability after a serious fault is decoupled according to the optimal decoupling scheme obtained by solving the model proposed in this paper. The decoupled subsystem can quickly recover to a stable state, and demand response can reduce part of the load to reduce the shortage in the island, which helps to ensure the frequency stability of the system.

[0220] Table 1 Comparison of search results for the decomposition section of Example 1-3

[0221]

[0222]

[0223] Table 2 Comparison of results of examples 2-3

[0224]

[0225] Table 3 gives the wind power output scenarios and probabilities, and Table 4 gives the decomposition results of Examples 6-8. Figure 5 The frequency curves of generator G5 in different scenarios after disconnection in Example 2 and Example 7 are compared. Figure 6 The participation amount of demand response in each scenario in Example 8 is compared. It can be concluded that by reducing part of the load through demand response, the unbalanced power generated can be effectively reduced, and the remaining surplus or deficit in the island can be adjusted through the primary frequency response of the system to reduce the amount of cutting of the generator and load according to the constraints related to the primary frequency response established. The combination of the two can make the system recover safe and stable operation better and faster after the decoupling.

[0226] Table 3 Wind power output scenarios and probabilities

[0227]

[0228]

[0229] Table 4 Solution results of example 6-8

[0230]

[0231] The above description is only a specific embodiment of the present invention, but it does not limit the patent protection scope of the present invention. Any equivalent changes or substitutions made by using the contents of the present invention and the drawings, directly or indirectly applied to other related technical fields, should be included in the protection scope of the present invention.

Claims

1. A method for searching for the optimal cross section of active decoupling considering the primary frequency response characteristics of the system, characterized in that: The following steps are involved: Step 1: A mixed integer linear programming (MILP) model is established based on the objective function and basic constraints of active disconnection of the power system, and the relevant operating data obtained when the power system is unstable is input into the MILP model to solve the initial active disconnection section set; Step 2: Considering the primary frequency response capability of the system, a primary frequency response model is established, including the constraints of the lowest frequency point and the highest frequency point, the primary frequency response reserve capacity constraint, and the frequency change rate constraint, and the initial active disconnection section set is constrained and screened to obtain the intermediate disconnection section set; Step 3: Model the interruptible power load, introduce the demand response model, filter the intermediate disconnection section set, and obtain the active disconnection optimal section set considering the primary frequency response characteristics of the system; Step 4: Considering the uncertainty of new energy output, construct the new energy output and its probability model, and obtain the complete active decoupling optimal section set considering the primary frequency response characteristics of the system and the uncertainty of new energy; The MILP model established in step 1 includes: (1) The objective function is: Where: Ω L is the set of lines; Ω D is the set of loads; Ω G is the set of generators; Ω W is a collection of wind farms; is the cost of disconnecting the line; d is the load shedding penalty coefficient; g is the machine-cutting penalty coefficient; w is the penalty coefficient for cutting off the wind farm; x ij is a binary variable, indicating the connectivity status of branch ij, x ij = 0 means that branch ij is disconnected by decoupling, x ij =1 means branch ij is connected; and They are the actual load shedding amount d of island k, the generator shedding amount g and the wind farm shedding amount w; (2) Generator coherence clustering model: Where: Ω N represents a set of nodes; Ω K represents the set of different islands generated by decoupling; i and j represent the first and last nodes of branch l respectively; x i,k 、x j,k Represent the state variables of nodes i and j respectively. If x i,k =1,x j,k =1, which means that nodes i and j belong to the island k; if x i,k =0,x j,k =0, which means nodes i and j do not belong to island k; t ij,k Represents auxiliary variables, used to linearize two 0-1 variables x i,k and x j,k multiply; (3) System connectivity model after generator synchronization grouping -M·x ij ≤P′ ij ≤M·x ij P′ ref ≥1 Where: P′ ij Represents the active power of the virtual branch ij; P′ jk represents the active power of virtual branch jk; P′ ref Represents the active power generated by the virtual generator, M is a sufficiently large number; Ω j Represents a set of devices connected to node j.

2. The method for searching the optimal cross section for active decoupling considering the primary frequency response characteristics of the system according to claim 1, characterized in that: The MILP model also includes: (1) Node balance constraints: 0≤v d ≤P d Where: P g , Q g Respectively represent the active and reactive output of generator g; P w represents the dispatch value of wind farm w; P d , Q d They represent the active and reactive power of load d after considering the load side demand response; v d 、v g and v w They represent the power imbalance of load d, generator g and wind farm w in the island respectively; P ij , Q ij are respectively the active and reactive power flows of branch ij; Q b Represents the reactive power dispatch value of capacitor group b; Q s Represents the reactive power dispatch value of the static VAR compensator s; Indicates the power factor of load d; P jk and Q jk Indicates the active and reactive power flows of branch jk; Indicates the maximum output of generator g; It represents the predicted output value of wind farm w; (2) Branch flow constraints: Where: (·) * are the initial values ​​of the variables given for linearizing the AC power flow; Respectively represent the maximum transmission capacity of active and reactive power of the line; v i and v j are the voltage amplitudes of nodes i and j, respectively, ij is the phase angle difference; b ij , g ij are the susceptance and conductance of branch ij respectively; As auxiliary variables, the AC power flow model is used for calculation; and Respectively represent the active and reactive losses of branch ij; (3) Generator output constraints: Where: and They represent the upper and lower limits of the active output of generator g respectively; and They represent the upper and lower limits of the reactive power output of the generator g respectively; ε represents the adjustment range when the generator is disconnected; Indicates the output of the generator before disconnection; (4) Constraints on new energy output: In the formula, σ represents the regulation range when the wind farm is disconnected; Indicates the output of the wind farm before disconnection; (5) Node voltage and node phase angle constraints: Where: and Respectively represent the upper and lower limits of the voltage amplitude of node i; θ i is the phase angle of node i; θ ref Indicates the voltage phase angle value of the reference node; (6) Constraints on capacitor banks and static VAR compensators: Where: Ω b is the collection of capacitor banks; Ω s is a collection of static VAR compensators; Represents the unit reactive power of capacitor bank b; y b is an integer variable representing the number of capacitor banks put into use, subject to the maximum number of restrictions; and They respectively represent the upper and lower limits of the reactive power dispatch of the static VAR compensator.

3. The method for searching the optimal cross section of active decoupling considering the primary frequency response characteristics of the system according to claim 1, characterized in that: The primary frequency response model in step 2 includes: (1) Constraints on the lowest and highest frequency points: Where: It represents the surplus of power generation over load in island k; It indicates the shortfall of power generation in island k below the load; and They respectively represent the upper reserve of the primary frequency response of the generator when a power shortage occurs and the lower reserve of the primary frequency response of the generator when a power surplus occurs; Indicates the lowest frequency point when power shortage occurs; Indicates the highest frequency point when power surplus occurs; c g Indicates the governor ramp rate of generator g; H k represents the equivalent inertia of island k; f0 represents the rated frequency of the system; f db Indicates the dead zone of the speed regulator; (2) Primary frequency response reserve capacity constraint: Where: and They represent the maximum upper reserve and maximum lower reserve of the primary frequency response of the generator respectively; is the decision variable of the cutting machine, Indicates the removal of the generator; (3) Frequency change rate constraint: Where: γ max Indicates the maximum operating frequency change rate of the generator set; H g represents the inertia time constant of the generator g.

4. The method for searching the optimal cross section of active decoupling considering the primary frequency response characteristics of the system according to claim 2, characterized in that: The primary frequency response model described in step 2 also includes other constraints: The surplus of island k after system disconnection and generator shedding or shortfall The model is as follows: Where: It represents the surplus after the load is cut off by the k island cutting machine; represents the power deficit after the k island is cut off and the load is shed; δ is a binary variable, which limits the relationship between the power surplus and the power deficit in the island; and is the decision variable of the cutting machine, Indicates the removal of the generator. Indicates the removal of wind farms; P w Represents the dispatch value of wind farm w.

5. The method for searching the optimal cross section of active decoupling considering the primary frequency response characteristics of the system according to claim 1, characterized in that: The demand response model described in step 3 is as follows: Where: and They represent the predicted load value of load d and the load value participating in demand response respectively; α d Indicates the load d interruptible load ratio.

6. The method for searching the optimal cross section of active decoupling considering the primary frequency response characteristics of the system according to claim 1, characterized in that: The overall optimization model expression considering the uncertainty of new energy output described in step 4 is specifically: Where: vector x represents the binary variables in the model; The vector y represents the continuous variables in the model; a and c represent the coefficient vectors of the variables to be solved in the objective function; A, B, C represent matrices in abstract form, h and f represent vectors in abstract form, representing the coefficients of the relevant constraints; ρ ξ is the probability of scene ξ appearing; y ξ is a continuous variable in different scenarios.

Images