A method for power network topology optimization considering source and load uncertainty

Abstract

A method for optimizing power network topology taking into account source-load uncertainty belongs to the technical field of power system optimization methods. Input candidate disconnection lines; use the continuous power flow method to find the operating point and load margin of the power system under the prediction scenario; use the scenario reduction method to reduce it and obtain representative scenarios; identify effective switching lines, calculate the load margin under each representative scenario after switching the lines, and sort the candidate lines; for the top-ranked candidate lines, accurately calculate the load margin of the system after it is disconnected under the prediction scenario; output the results. The present invention reduces a large number of uncertain scenarios to a few representative scenarios, reduces the complexity of calculation, maximizes the load margin of the predicted power system, and makes the load margin of all possible scenarios reach above the preset ideal threshold, thereby improving the static voltage stability of the power system considering source-load uncertainty.

Description

Technical Field

[0001] The invention discloses a power network topology optimization method taking into account source-load uncertainty, belonging to the technical field of power system operation optimization. Background Art

[0002] The latest data from the International Renewable Energy Agency (IRENA) shows that by the end of 2020, the global renewable energy generation capacity reached 2799GW, an increase of 260GW. Among them, photovoltaic power generation and wind power generation dominated renewable energy generation, accounting for 91% of the total net increase in renewable energy in 2020. The large-scale injection of renewable energy into the power grid has greatly reduced carbon emissions and achieved clean power generation. However, due to its inherent nature, renewable energy cannot be dispatched in the traditional sense. Renewable energy brings the following challenges to the power system: (1) The output of renewable energy is closely related to weather conditions and cannot be accurately predicted, which brings more uncertainty to the power system; (2) There is a limit on the grid-connected power of renewable energy. If this limit is exceeded, the constraints of current, voltage or static / transient stability will be violated. Conversely, these constraints will also limit the penetration level of renewable energy; (3) When the injection rate of renewable energy is high, its reactive power demand may not be met. The above challenges may cause the power system to operate near the limit or stability boundary, or even cause system collapse. In addition, the uncertainty of renewable energy itself plus the uncertainty of load bring new challenges to the safe and stable operation of the power system. Therefore, it is imperative to study the impact of renewable energy and load uncertainties on static voltage stability and provide appropriate preventive control for power systems.

[0003] With the continuous expansion of the scale of power grids and the continuous development of new power systems, a large number of renewable energy sources have been connected to the grid, and the demand-side load has increased significantly, resulting in a serious shortage of reactive power in the system and the collapse of the system. It is becoming more and more common, and it is becoming increasingly difficult for the system to maintain voltage stability on its own. Therefore, analyzing the voltage instability mechanism of new power systems is helpful to further study measures to improve static voltage stability.

[0004] As one of the ways of topology optimization, transmission line disconnection has been generally regarded as an economical and effective measure. It is used to alleviate transmission congestion, alleviate line overload and voltage violations, and improve stability. However, considering the uncertainty of source and load, the research on improving static voltage stability by line disconnection has not been solved. According to the above analysis, how to improve the load margin of the power system with dual uncertainty of renewable energy and load by line disconnection, so as to improve its static voltage stability, is the focus of this invention.

[0005] The existing transmission line disconnection has the following problems:

[0006] 1) The existing technology does not comprehensively consider the impact of source-load uncertainty on static voltage bifurcation (saddle-node bifurcation (SNB) and structure-induced bifurcation (SIB));

[0007] 2) The classic scenario reduction method aims to minimize the probability distance between the retained scenario set and the original scenario. Although it has a certain theoretical basis, the reduced scenarios cannot target specific research problems and are less representative.

[0008] 3) The line disconnection research for static voltage stability problem does not take into account the dual uncertainties of renewable energy and load. Summary of the invention

[0009] The technical problem to be solved by the present invention is: to overcome the shortcomings of the prior art and provide a power network topology optimization method that determines an optimal network topology by disconnecting effective lines, maximizes the load margin of the power system under predicted data, and ensures that the load margin of the power system in all possible scenarios reaches the ideal value pre-set by the operator.

[0010] The technical solution adopted by the present invention to solve the technical problem is: the power network topology optimization method taking into account source and load uncertainty is characterized by comprising the following steps:

[0011] S1 is given the current network topology, the current operating point of the system, the power generation plan, the maintenance plan, the historical forecast data of renewable energy and load and the forecast error, and inputs the candidate disconnection line;

[0012] S2 uses the continuation power flow method to find the operating point and load margin of the expected power system (i.e., the power system before disconnecting the line) under the forecast scenario;

[0013] S3 uses a scenario-based approach to characterize the uncertainties of renewable energy and loads and generates a set of joint scenarios;

[0014] S4 uses the clustering method to reduce the scenarios according to the dual indicators of scenario distance and load margin distance to obtain representative scenarios;

[0015] S5 uses a weighted sensitivity method to identify effective disconnection lines, retains the lines that increase the load margin after disconnection, and forms a candidate line set;

[0016] S6 uses a nonlinear look-ahead method to calculate the load margin of each disconnected line in the candidate line set under various representative scenarios, and deletes the lines that violate the operation constraints; and sorts all candidate disconnected lines by weighted margin indicators;

[0017] S7 uses the continuous power flow method to accurately calculate the load margin of the system after the candidate lines ranked high in the previous stage are disconnected, and removes the lines that do not meet the threshold constraints from the candidate line set; and calculates the load margin of the system after the candidate lines are disconnected in the prediction scenario;

[0018] S8 outputs the disconnected line solution, the load margin after the line is disconnected in the predicted scenario, and the load margin of all representative scenarios after each line is disconnected. The disconnected line with the largest load margin in the predicted scenario is the best solution.

[0019] Preferably, the method further comprises that a mathematical model of the line switching problem for enhancing static voltage stability taking into account source and load uncertainty is:

[0020] maxλ 0 (N b );

[0021] Among them, N b Determine the optimal network topology by switching valid lines, λ 0 To maximize the prediction scenario s 0 Load margin of the power system;

[0022] The continuous power flow balance equation of the power system after the transmission line is disconnected is:

[0023]

[0024] The load margin constraint of the power system under all uncertain scenarios is:

[0025]

[0026] The operation constraints of the power system after the transmission line is disconnected are:

[0027]

[0028]

[0029]

[0030] The allowed number of disconnected lines is constrained as follows:

[0031] NE(NN b )=1;

[0032] in, Represents the network topology N b Next, scenes m The nonlinear continuous power flow equilibrium equation is: is the vector of state variables; h represents the variable parameters in the system, including the uncertainty caused by prediction error and line disconnection admittance, λ th is the load margin threshold set in advance by the operator, V i,sm 、V i,min and V i,max It's scenes m The voltage amplitude of the lower bus i and its upper and lower limits, and S ij,max are the apparent power between busbars i and j and its upper and lower limits, Q G,i,min and Q G,i,max The scenes are m The reactive power output of the generator on the lower bus i and its upper and lower limits, B and B G are the set of buses and the set of generator buses, and B G ∈B, NE(·) represents the current network topology N and the optimal network topology N b The network topology differences caused by the disconnection of transmission lines.

[0033] Preferably, the method further comprises, according to the Copula theory and the Latin hypercube sampling method, generating a renewable energy scenario set S R and load scenario set S L , and generate a joint scenario set Ω; Ω is the multivariate relationship between renewable energy and load scenario set, that is, the renewable energy scenario set S R and load scenario set S L The total number of Ω is n R ×n L ;

[0034]

[0035] in, are the samples of the i-th renewable energy source and the j-th load respectively; n R and n L are the number of scenario sets for renewable energy and load, respectively.

[0036] Preferably, the method further comprises reducing the scenes according to the dual indicators of scene distance and load margin distance to obtain representative scenes;

[0037] 4.1 Cluster all scenes into several groups according to the proximity of the scene distances, and determine the center of each group as the central scene:

[0038]

[0039] in, For scenesm and n When r=2, it represents the Euclidean distance between the two scenes.

[0040] 4.2 Apply the continuous power flow method to calculate the load margin, bifurcation point type, and non-zero vector w for each central scenario;

[0041] Select a group and calculate and predict scenarios 0 The load margin changes caused by the differences between them are used to estimate the load margins of all group members;

[0042] At the saddle node bifurcation point, the scenario s m With prediction scenarios 0 The load margin change caused by the difference between for:

[0043]

[0044] Among them, ΔP spec and ΔQ spec are the changes in active power and reactive power injected into the generator node respectively; ΔP inc and ΔQ inc Respectively represent the changes in active power and reactive power; w is the Jacobian matrix A nonzero left eigenvector of a zero eigenvalue; λ 0 To predict the scene 0 System load margin under inc and Q inc Respectively represent the increase in active power and reactive power;

[0045] At the structure-induced bifurcation point, scenario s m With prediction scenarios 0 The load margin change caused by the difference between for:

[0046]

[0047] Where Q k,inc and ΔQ k,inc are the reactive power increase and its change corresponding to the kth node respectively;

[0048] Therefore, for any bifurcation (SNB or SIB), scenario s m Load margin under for:

[0049]

[0050] Repeat step 4.2 until all groups have been analyzed;

[0051] 4.3 According to the closeness of the load margins of all scenarios, all scenarios are grouped, and the central scenario of each group is selected as the representative scenario to form the representative scenario set RS; Indicates scenes m and n The degree of closeness of the load margin of the power system is:

[0052]

[0053] in, and Respectively represent scenes m and n The estimated load margin.

[0054] Preferably, the method further comprises screening effective disconnection lines that can improve the system load margin after disconnection by weighted sensitivity;

[0055] The weighted sensitivity H is: H = J × M;

[0056] in, is a weighted vector whose elements represent the predicted load margin associated with the representative scenario RS, and M is the sensitivity matrix;

[0057]

[0058] Among them, n l is the total number of candidate switching lines; It's scenes m Lower disconnect line k Load margin of the rear system, s m ∈s RS ; is the change in load margin between the two power systems;

[0059]

[0060] in, The scenes are m Lower disconnect line k Before, the active power and reactive power from node i to node j; The scenes are m Lower disconnect line k Then the active power and reactive power from node j to node i; is a vector corresponds to the element in the active power balance equation of node i, is a vector The element in the reactive power balance equation corresponding to node i, is a vector corresponds to the element in the active power balance equation at node j, is a vector corresponds to the element in the active power balance equation of node j; Indicates that in scenes m Next, disconnect the line k The load margin of the power system can be increased, otherwise, the load margin of the system can be reduced;

[0061] Select the lines with weighted sensitivity greater than 0 (i.e., the load margin increases after disconnection) to form the candidate line set CA;

[0062] CA={l k |H(l k )≥0}.

[0063] Preferably, the method further comprises: using a nonlinear look-ahead method to calculate the load margin of each disconnected line in the candidate line set under each representative scenario, and deleting the line that violates the operation constraint; and sorting all candidate disconnected lines by a weighted margin index; the weighted margin index is:

[0064]

[0065] in, For scenes m Lower disconnect line k Load margin after It's scenes m The corresponding elements of matrix J.

[0066] Preferably, the method also includes evaluating the top-ranked candidate lines one by one through a continuous power flow method; checking whether all scenarios after disconnecting the line meet the load margin constraints, if so, retaining the line; if not, removing the line from the candidate lines; and calculating the load margin of the system after each candidate line is disconnected in the predicted scenario.

[0067] Preferably, the method also includes output line disconnection solutions, load margins after line disconnection in predicted scenarios, and load margins of all representative scenarios after each line is disconnected; the disconnected line with the largest load margin after line disconnection in predicted scenarios is the solution to the problem.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] This power network topology optimization method taking into account source-load uncertainty proposes a scenario generation and reduction method based on the dual indicators of load margin and scenario distance, which reduces a large number of uncertain scenarios into a few representative scenarios and reduces the complexity of the calculation; a staged power grid topology optimization method is proposed, which determines an optimal power network topology structure by disconnecting transmission lines, maximizes the load margin of the expected power system, and makes the load margin of all possible scenarios reach above the ideal threshold set in advance by the operator, thereby improving the static voltage stability of the power system taking into account source-load uncertainty; in addition, this method does not require additional investment, is highly economical and easy to operate. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 The overall flow chart of the power network topology optimization method considering source and load uncertainty;

[0071] Figure 2 Comparison of PV curves before and after switching line 8-5 in the prediction scenario without considering reactive power limit;

[0072] Figure 3 The load margin of all scenarios of disconnecting line 8-5 without considering reactive power limit;

[0073] Figure 4 Comparison of PV curves before and after switching line 17-31 in the prediction scenario when considering reactive power limit;

[0074] Figure 5 Load margin for all scenarios considering reactive power limits. DETAILED DESCRIPTION

[0075] like Figure 1 As shown: A method for optimizing power network topology taking into account source and load uncertainty comprises the following steps:

[0076] S1 Given the current network topology, the current operating point of the system, the generation plan, the maintenance plan, the historical forecast data and forecast errors of renewable energy and load, the candidate disconnection lines are input.

[0077] The objective function of the mathematical model for the line disconnection problem of enhancing static voltage stability considering source and load uncertainty is:

[0078] maxλ 0 (N b ); (1)

[0079] Among them, N b Determine the optimal network topology by switching valid lines, λ 0 To maximize the prediction scenario s 0 Load margin of the power system;

[0080] The continuous power flow balance equation of the power system after the transmission line is disconnected is:

[0081]

[0082] The load margin constraint of the power system under all uncertain scenarios is:

[0083]

[0084] The operation constraints of the power system after the transmission line is disconnected are:

[0085]

[0086]

[0087]

[0088] The allowed number of disconnected lines is constrained as follows:

[0089] NE(NN b )=1; (7)

[0090] in, Represents the network topology N b Next, scenes m The nonlinear continuous power flow equilibrium equation is: is a vector of state variables. h represents the variable parameters in the system, including the uncertainty caused by prediction error and line disconnection admittance, λ th It is the load margin threshold set in advance by the operator. V i,min and V i,max It's scenes m The voltage amplitude of the lower bus i and its upper and lower limits, and S ij,max are the apparent power between busbars i and j and its upper and lower limits, Q G,i,min and Q G,i,max The scenes are m The reactive power output of the generator on the lower bus i and its upper and lower limits, B and B G are the set of buses and the set of generator buses, and B G ∈B, NE(·) represents the current network topology N and the optimal network topology N b The network topology differences caused by the disconnection of transmission lines.

[0091] S2 uses the continuation power flow method to find the operating point and load margin of the expected power system (i.e., the power system before disconnecting the line) under the prediction scenario.

[0092] S3 uses a scenario-based approach to characterize the uncertainties of renewable energy and loads and generates a set of joint scenarios;

[0093] Furthermore, the specific process of step S3 includes the following steps:

[0094] S3.1 Input the historical forecast error of renewable energy into R language to obtain the correlation coefficient between renewable energy output. Input the forecast value of renewable energy and load.

[0095] S3.2 Generate the renewable energy scenario set SR and load scenario set S respectively according to the Copula theory and Latin hypersampling method. L , and generate a joint scenario set Ω, where Ω is the multivariate relationship between renewable energy and load scenario sets, that is, the renewable energy scenario set S R and load scenario set S L The total number of Ω is n R ×n L .

[0096]

[0097] in, are the samples of the i-th renewable energy source and the j-th load respectively; n R and n L are the number of scenario sets for renewable energy and load, respectively.

[0098] S4 uses the clustering method to reduce the scenarios according to the dual indicators of scenario distance and load margin distance to obtain representative scenarios;

[0099] Furthermore, the specific process of step S4 includes the following steps:

[0100] S4.1 Cluster all scenes into several groups according to the proximity of the scene distances, and determine the center of each group as the central scene:

[0101]

[0102] in, For scenes m and n When r = 2, it represents the Euclidean distance between the two scenarios. S4.2 Apply the continuous power flow method to calculate the load margin, bifurcation point type, and non-zero vector w for each central scenario;

[0103] S4.3 Select a group and calculate and predict the scenario s 0 The load margin changes caused by the differences between the two are used to estimate the load margins of all members of the group until all groups are analyzed;

[0104] Furthermore, the specific process of step S4.3 includes the following steps:

[0105] For Saddle Node Bifurcation (SNB)

[0106] S4.3.1 At the saddle node bifurcation point, the steady-state model describing the power system is:

[0107]

[0108] Among them, P spec and Q spec are the active power and reactive power injected into the generator node respectively; calc and Q calc are the calculated active power and reactive power of the transmission line respectively; P inc and Q inc Represent the increase in active power and reactive power respectively.

[0109] S4.3.2 Taking into account the uncertainty of source and load, the steady-state model of the power system is:

[0110]

[0111] Among them, Δx, Δh, and Δλ are the changes in system state quantity, control parameter, and load margin caused by source load uncertainty, respectively.

[0112] S4.3.3 At the saddle node bifurcation point, the first-order Taylor expansion of the steady-state model of the power system taking into account the uncertainty of sources and loads is:

[0113]

[0114] Among them, the Jacobian matrix is singular. At the saddle node bifurcation point, there exists a matrix corresponding to the Jacobian matrix The non-zero eigenvector w of zero eigenvalue is such that

[0115] S4.3.4 The load margin change Δλ caused by the change in parameter Δh is:

[0116]

[0117] S4.3.5 Scenarios m With prediction scenarios 0 The load margin change caused by the difference between for:

[0118]

[0119] For structure-induced bifurcation (SIB)

[0120] At the structurally induced bifurcation point, an additional equilibrium equation should be added to the steady-state model. Assuming that the generator on bus k is at the bifurcation point, its reactive power reaches the limit, that is, the Q-constraint equation Q k,max -Q G,k =0 holds true.

[0121] S4.3.1 At the structurally induced bifurcation point, the steady-state model of the power system is:

[0122]

[0123] S4.3.2 At the structural induced bifurcation point, taking into account the uncertainty of source and load, the steady-state model of the power system is:

[0124]

[0125] S4.3.3 At the structure-induced bifurcation point, the first-order Taylor expansion of the steady-state model of the power system taking into account the uncertainty of sources and loads is:

[0126]

[0127] S4.3.4 The load margin change Δλ caused by the change in parameter Δh is:

[0128]

[0129] S4.3.5 Scenarios m With prediction scenarios 0 The load margin change caused by the difference between for:

[0130]

[0131] S4.4 For any bifurcation (SNB or SIB), scenario s m Load margin under for:

[0132]

[0133] S4.5 Group all scenarios according to the closeness of their load margins, and select the central scenario of each group as the representative scenario. Indicates scenes m and n The distance to lower the power system load margin.

[0134]

[0135] in, and Respectively represent scenes m and n The estimated load margin.

[0136] S5 identifies effective disconnection lines through a weighted sensitivity method, retains the lines that increase the load margin after disconnection, and forms a candidate line set.

[0137] Furthermore, the specific process of step S5 is as follows:

[0138] In order to quickly identify the lines that can increase the load margin of representative scenarios after disconnection, a weighted sensitivity is proposed to select effective disconnection lines to increase the load margin to reach the ideal value. The weighted sensitivity H is:

[0139] H=J×M; (22)

[0140] in, is a weighted vector whose elements represent the predicted load margin associated with the representative scenario RS, and M is the sensitivity matrix;

[0141]

[0142] Among them, n l is the total number of candidate switching lines; It's scenes m Lower disconnect line k Load margin of the rear system, s m ∈s RS ; is the change in load margin between the two power systems;

[0143]

[0144] in, The scenes are m Lower disconnect line k Before, the active power and reactive power from node i to node j; The scenes are m Lower disconnect line k Then the active power and reactive power from node j to node i; is a vector corresponds to the element in the active power balance equation of node i, is a vector The element in the reactive power balance equation corresponding to node i, is a vector corresponds to the element in the active power balance equation at node j, is a vector Corresponds to the element in the active power balance equation of node j. Indicates that in scenes m Next, disconnect the line k The load margin of the power system can be increased, otherwise, the load margin of the system will be reduced.

[0145] Select the lines with weighted sensitivity greater than 0 (i.e., the load margin increases after disconnection) to form the candidate line set CA;

[0146] CA={l k |H(l k )≥0}; (25)

[0147] S6 uses a nonlinear look-ahead method to calculate the load margin of each disconnected line in the candidate line set under each representative scenario, deletes the lines that violate the operation constraints, and sorts all candidate disconnected lines by weighted margin indicators. The weighted margin indicators are:

[0148]

[0149] in, For scenes m Lower disconnect line k The load margin after. It's scenes m The corresponding elements of matrix J.

[0150] S7 uses the continuous power flow method to accurately calculate the load margin of the candidate lines ranked high in the previous stage after disconnection, and removes the lines that do not meet the threshold constraints from the candidate line set; and calculates the load margin of the system after each candidate line is disconnected in the prediction scenario;

[0151] S8 outputs the disconnection line solution, the load margin after disconnection under the predicted scenario, and the load margin of all representative scenarios after each line is disconnected. The solution with the largest load margin after disconnection under the predicted scenario is the "optimal solution".

[0152] At this point, the detailed calculation process of the network topology optimization method taking into account source-load uncertainty involved in the present invention is completed.

[0153] The method of the present invention is verified by simulation of the IEEE 118-node system example. The test system has a total of 186 branches, node 69 is a balancing node, and there are 177 disconnectable lines. In this test example, 32 loads on nodes 35 to 67 increased by 50%, and all load increases were provided by 3 generators on busbars 4, 31 and 49. The load margin of the expected power system under the predicted scenario is 2.9505 (per unit). Before disconnecting the line, the load margin in the worst scenario is 2.5138 (per unit). Assume that the load margin required for all scenarios is set to λ without considering the reactive power limit of the generator. th =3 (per unit value). Using the network topology optimization method taking into account the uncertainty of source and load proposed in the present invention, the load margin in the prediction scenario before and after the candidate line is disconnected is shown in Table 1. In the prediction scenario, the load margin of the power system is increased by 32.44% after disconnecting line 8-5. In the prediction scenario, the PV curve of the expected power system before and after switching line 8-5 is as follows: Figure 2 The load margin of all scenarios of switching line 8-5 is shown as follows Figure 3 shown.

[0154] Table 1 Load margin and line disconnection solution without considering reactive power limit

[0155]

[0156] Considering the reactive power limits of all generators: In the forecast scenario, the load margin of the expected power system in the forecast scenario is 1.6249 (per unit) due to the reactive power limit of the generators at nodes 34, 56, 36, 15, 19, 18, 62, 55, and 65, which leads to the occurrence of SIB. Before disconnecting the line, the load margin in the worst scenario is 0.068 (per unit). Assume that the load margin required for all scenarios when considering the reactive power limit of the generator is set to λ th =1.1 (per unit value). Using the network topology optimization method taking into account the uncertainty of source and load proposed in the present invention, the load margin in the prediction scenario before and after the candidate line is disconnected is shown in Table 2. In the prediction scenario, the load margin of the power system is increased by 13.47% after disconnecting line 17-31. In the prediction scenario, the PV curve of the expected power system before and after switching line 17-31 is as follows: Figure 4 The load margin of all scenarios of switching lines 17-31 is shown as follows Figure 5 shown.

[0157] Table 2 Load margin and line disconnection solutions considering reactive power limit

[0158]

Claims

1. A method for optimizing power network topology taking into account source-load uncertainty, characterized in that: The steps include: S1 is given the current network topology, the current operating point of the system, the power generation plan, the maintenance plan, the historical forecast data of renewable energy and load and the forecast error, and inputs the candidate disconnection line; S2 uses the continuation power flow method to find the expected power system under the forecast scenario, that is, the operating point and load margin of the power system before disconnecting the line; S3 uses a scenario-based approach to characterize the uncertainties of renewable energy and loads and generates a set of joint scenarios; S4 uses the clustering method to reduce the scenarios according to the dual indicators of scenario distance and load margin distance to obtain representative scenarios; S5 uses a weighted sensitivity method to identify effective disconnection lines, retains the lines that increase the load margin after disconnection, and forms a candidate line set; S6 uses a nonlinear look-ahead method to calculate the load margin of each disconnected line in the candidate line set under various representative scenarios and deletes the lines that violate the operation constraints; And all candidate disconnection lines are ranked by weighted margin index; S7 uses the continuous power flow method to accurately calculate the load margin of the system after the top candidate lines in the previous stage are disconnected, and removes the lines that do not meet the threshold constraints from the candidate line set; And calculate the load margin of the system after each candidate line is disconnected in the prediction scenario; S8 outputs the disconnected line solution, the load margin after the line is disconnected in the predicted scenario, and the load margin of all representative scenarios after each line is disconnected. The disconnected line with the largest load margin in the predicted scenario is the best solution.

2. The method for optimizing power network topology taking into account source-load uncertainty according to claim 1, characterized in that: The method further includes that a mathematical model of the line switching problem for enhancing static voltage stability taking into account source and load uncertainty is: ； in, Determine the optimal network topology by switching valid lines. To maximize the prediction scenario Load margin of the power system; The continuous power flow balance equation of the power system after the transmission line is disconnected is: ； The load margin constraint of the power system under all uncertain scenarios is: ； The operation constraints of the power system after the transmission line is disconnected are: ； ； ； The allowed number of disconnected lines is constrained as follows: ； in, Represents the network topology Next, scene The nonlinear continuous power flow equilibrium equation is: is the vector of state variables; represents the variable parameters in the system, including uncertainties due to prediction errors and line interruption admittance, It is the load margin threshold set in advance by the operator. , and It's a scene Lower bus The voltage amplitude and its upper and lower limits, and Busbar and The apparent power between and its upper and lower limits, , and The scenes are Lower bus The reactive power output of the upper generator and its upper and lower limits, and are the set of buses and the set of generator buses respectively, and , Represents the current network topology and the optimal network topology The network topology differences caused by the disconnection of transmission lines between Representing scenes The estimated load margin is For the joint scene set.

3. The method for optimizing power network topology taking into account source-load uncertainty according to claim 1, characterized in that: The method also includes generating renewable energy scenario sets according to Copula theory and Latin hypercube sampling method and load scenario sets , and generate a joint scene set ; is the multivariate relationship between renewable energy and load scenario set, i.e., the renewable energy scenario set and load scenario collection The Cartesian product of The total number is ; ； in, , Respectively i Renewable energy and j A sample of loads; and are the number of scenario sets for renewable energy and load, respectively.

4. The method for optimizing power network topology taking into account source-load uncertainty according to claim 1, characterized in that: The method further includes reducing the scenarios according to dual indicators of scenario distance and load margin distance to obtain representative scenarios; 4.1 Cluster all scenes into several groups according to the proximity of the scene distances, and determine the center of each group as the central scene: ； in, For the scene and Between r -Distance; when r =2, it indicates the Euclidean distance between the two scenes; 4.2 Calculation of load margin, bifurcation point type, and non-zero vector for each central scenario using the continuous power flow method ; Select a group and calculate and predict the scenario The load margin changes caused by the differences between them are used to estimate the load margins of all group members; At the saddle node bifurcation point, the scene With prediction scenario The load margin change caused by the difference between for: ； in, and are the changes of active power and reactive power injected into the generator node respectively; and Respectively represent the changes in active power and reactive power; is the Jacobian matrix corresponding to Non-zero left eigenvector of zero eigenvalue; For prediction scenarios System load margin under ; and Respectively represent the increase in active power and reactive power; At the structure-induced bifurcation point, the scene With prediction scenario The load margin change caused by the difference between for: ； in and Respectively k The reactive power increase and change corresponding to each node; Therefore, for SNB bifurcation or SIB bifurcation, the scenario Load margin under for: ； Repeat step 4.2 until all groups have been analyzed; 4.3 According to the closeness of the load margins of all scenarios, all scenarios are grouped, and the central scenario of each group is selected as the representative scenario to form a representative scenario set. ; Representation scene and The degree of closeness of the load margin of the power system is: ； in, and Representing scenes and The estimated load margin.

5. The method for optimizing power network topology taking into account source-load uncertainty according to claim 1, characterized in that: The method further includes screening, by weighted sensitivity, effective disconnection lines that can improve system load margin after disconnection; Weighted sensitivity for: ; in, is a weighted vector whose elements represent the scene The associated forecast load margin, is the sensitivity matrix, It is the load margin threshold set in advance by the operator; ； in, is the total number of candidate switching lines; It's a scene Lower disconnect line The load margin of the rear system, ; is the change in load margin between the two power systems, To maximize the prediction scenario Load margin of the power system; ； in, , Respectively for scenes Lower disconnect line Before, from the node To Node Active power and reactive power; , Respectively for scenes Lower disconnect line Back Slave Node To Node Active power and reactive power; , is a vector Corresponding to the node The elements in the active power balance equation are: is a vector Corresponding to the node The elements in the reactive power balance equation are: is a vector Corresponding to the node The elements in the active power balance equation are: is a vector Corresponding to the node Elements in the active power balance equation; Indicates in the scene Next, disconnect the line The load margin of the power system can be increased, otherwise, the load margin of the system can be reduced. To disconnect the line; Select the lines with weighted sensitivity greater than 0, that is, the lines that increase the load margin after disconnection to form the candidate line set ; ； in, To disconnect the line The weighted sensitivity.

6. The method for optimizing power network topology taking into account source-load uncertainty according to claim 1, characterized in that: The method further includes calculating the load margin under each representative scenario for each disconnected line in the candidate line set using a nonlinear look-ahead method, and deleting lines that violate operation constraints; And all candidate disconnection lines are ranked by weighted margin index; the weighted margin index is: ； in, For the scene Lower disconnect line The load margin after It's a scene The corresponding matrix elements.

7. The method for optimizing power network topology taking into account source and load uncertainty according to claim 1, characterized in that: The method further includes evaluating the top-ranked candidate routes one by one by a continuous power flow method; Check whether all scenarios after disconnecting the line meet the load margin constraint. If so, keep the line. If not, remove the line from the candidate lines. And calculate the load margin of the system after each candidate line is disconnected in the prediction scenario.

8. The method for optimizing power network topology taking into account source and load uncertainty according to claim 1, characterized in that: The method also includes outputting the line breaking solution, the load margin after breaking the line in the predicted scenario, and the load margin of all representative scenarios after each line breaking; The disconnected line with the largest load margin after disconnecting the line in the prediction scenario is the best solution.

Images