A data-driven method for analyzing dynamic power supply margin in power grid

Abstract

The present invention discloses a data-driven method for analyzing the dynamic power supply margin of a power grid, comprising the following steps: obtaining 10 kV topological data, combing 10 kV large feeder and segment information, and counting the load information of each segment; iteratively setting the number of clusters according to the 24-hour load data of each segment, and calculating the optimal operation mode recommendation within a typical time interval; in each power grid operation mode, taking the load in each operation mode as known operation data, iteratively calculating the power supply margin of the power grid to distributed energy, and obtaining the final power supply margin result of the power grid to distributed energy; comprehensively considering all operation modes, obtaining the minimum value of the power supply margin of each device to distributed energy as the final calculation result of the power supply margin of the feeder to distributed energy, thereby analyzing the operation mode of the power grid in different time periods and the power supply margin of the power grid to distributed energy, and making the analysis of the dynamic power supply margin of the power grid more accurate and comprehensive.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply margin analysis, and in particular to a data-driven dynamic power supply margin analysis method for a power grid. Background Art

[0002] At present, the power grid structure is gradually changing from a single large power plant and substation power supply mode to a multi-power and distributed power supply mode. After the distributed energy is connected to the power grid, the power grid flow changes significantly. In addition, due to the influence of meteorological factors, the randomness of distributed energy power supply is very strong. In order to ensure the sustainable supply of clean energy and support the country's energy transformation tasks, it is necessary to propose a set of intelligent analysis methods for the distributed energy power supply margin of the existing power grid. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and to design a data-driven dynamic power supply margin analysis method for power grids.

[0004] To achieve the above-mentioned purpose, the technical solution of the present invention is a data-driven power grid dynamic power supply margin analysis method, comprising the following steps:

[0005] Step 1: Obtain 10 kV topology data, sort out 10 kV large feeder and segment information, and count the load information of each segment;

[0006] Step 2: According to the 24-hour load data of each segment, iteratively set the number of clusters, perform time-based cluster analysis, and obtain the clustering results of the typical time interval of the representative day;

[0007] Step 3: Consider the constraints on safe operation of the power grid and calculate the optimal operation mode recommendation within the typical time interval;

[0008] Step 4: Under each grid operation mode, taking the load under each operation mode as known operation data, iteratively calculate the power supply margin of the grid to the distributed energy, and obtain the final power supply margin result of the grid to the distributed energy;

[0009] Step 5: Comprehensively consider all operation modes and obtain the minimum value of the distributed energy power supply margin of each device as the final calculation result of the distributed energy power supply margin of the feeder.

[0010] As a further description of the technical solution, in step one, based on the 10 kV power grid topology data, a topology search is performed on the interconnected 10 kV lines with the 10 kV busbar of the upper-level substation as the boundary.

[0011] As a further description of the technical solution, in step one, the interconnected 10 kV lines are segmented into units with the segment switches as the boundaries, and the load data in each segment is counted with each segment unit as the object.

[0012] As a further description of the present technical solution, in step 2, the 24-hour load data counted in step 1 is clustered based on time using the k-means algorithm to analyze the time segments with the same load characteristics on the representative day, and each time segment corresponds to an operating mode.

[0013] As a further description of the technical solution, in step three, with feeder load balancing as the optimization goal and feeder load rate and voltage not exceeding the limit as constraints, the operating mode within a certain time interval is analyzed and calculated. The operating modes in all time intervals calculated in step two are merged, and the operating modes in adjacent time intervals and the same are merged.

[0014] As a further description of the technical solution, in step four, the analysis object is a single 10 kV feeder operating in radial operation under known grid operation conditions, and the equipment load rate, short-circuit current, and equipment voltage drop are respectively calculated to provide a power supply margin for distributed energy, and the power supply margin is logically allocated according to the parent-child relationship of the equipment under the radial operation state.

[0015] As a further description of the present technical solution, step five includes step A and step B.

[0016] As a further description of the present technical solution, in step A, based on the typical day operation mode scheme of the power grid analyzed in step three, the power supply margin calculation in step four is performed for each operation mode respectively, so as to obtain the power supply margin results of each node of the power grid for distributed energy under each operation mode.

[0017] As a further description of the technical solution, in step B, the power supply margin results of each node in the power grid and under each operating mode are compared, and the minimum power supply margin is selected as the final calculation result of the power supply margin of the feeder for distributed energy.

[0018] The beneficial effect is that this data-driven dynamic power supply margin analysis method for the power grid takes into account the state of the 10 kV section interconnecting switches and the operating mode between the feeders, the time characteristics of the distributed energy output and the load characteristics of each node in the power grid, thereby analyzing the operating mode of the power grid in different time periods and the power supply margin of the power grid to the distributed energy. The analysis of the dynamic power supply margin of the power grid is more accurate and comprehensive, and the analysis efficiency is also relatively high. DETAILED DESCRIPTION

[0019] First of all, the original intention of the design of the present invention is explained. At present, the power grid structure is gradually changing from a single large power plant and substation power supply mode to a multi-power source and distributed power supply mode. After the distributed energy is connected to the power grid, the power grid trend changes significantly. In addition, due to the influence of meteorological factors, the randomness of distributed energy power supply is very strong. For this reason, the dynamic power supply margin analysis of the power grid is inaccurate. Therefore, the present invention designs a data-driven dynamic power supply margin analysis method for the power grid.

[0020] The present invention will be described in detail below. A data-driven power grid dynamic power supply margin analysis method includes the following steps:

[0021] Step 1: Obtain 10 kV topology data, sort out 10 kV large feeder and segment information, and count the load information of each segment; In step 1, according to the 10 kV power grid topology data, take the 10 kV busbar of the upper substation as the boundary, perform topology search on the interconnected 10 kV lines, take the segment switch as the boundary, sort out the segment units inside the interconnected 10 kV lines, take each segment unit as the object, and count the load data in each segment. The specific implementation is as follows:

[0022] (1) Sorting out 10 kV isolated topological islands: Taking the outgoing line switch of the 10 kV line upstream substation as the boundary, the interconnected 10 kV equipment is divided into a topological island according to the topological search algorithm. Thus, the 10 kV distribution network is divided into N isolated 10 kV feeder topological islands.

[0023] (2) 10 kV segment combing: Take each isolated 10 kV topological island as the analysis object, and use the pole-mounted segmented circuit breakers and station segmented circuit breakers that support segmentation functions as boundaries to execute the topological search algorithm inside the topological island. The interconnected devices are divided into segmented topological islands with segment switches as boundaries.

[0024] (3) Statistics of 10 kV segment load: Taking the segment topology island as the object, statistics are made on the active power data of the representative day at 24 hours. The load data of each segment in the large feeder at 24 hours is obtained.

[0025] Step 2: According to the 24-hour load data of each segment, iteratively set the number of clusters, perform time-based cluster analysis, and obtain the typical time interval clustering results of the representative day; in step 2, the 24-hour load data of the segment counted in step 1 is clustered based on time using the k-means algorithm to analyze the time segments with the same load characteristics of the representative day. Each time segment corresponds to an operation mode. The specific implementation is as follows:

[0026] (1) Taking the minimum clustering error of all segments of the power grid in all time intervals as the objective function, analyze the optimal operation mode time interval distribution scheme. Set the objective function to minimize the error between the segment load and the average load in all time intervals:

[0027]

[0028] Where: P i is the load at a certain point in a certain period of time, P avg The average load for a period of time

[0029] is the error between the load at all times and the average load in a certain period of time,

[0030] is the sum of errors in all time periods, is the sum of all segment errors;

[0031] (2) Set the operation mode adjustment frequency to 3, and the number of time intervals under different operation modes is 3, that is, the number of clusters corresponding to k-means clustering is 3.

[0032] Step 3: Considering the constraints of safe operation of the power grid, calculate the optimal operation mode recommendation within the typical time interval; in step 3, take feeder load balancing as the optimization goal, take feeder load rate and voltage limit as constraints, analyze and calculate the operation mode within a certain time interval. The operation modes of all time intervals calculated in step 2 are merged, and the same operation modes in adjacent time intervals are merged. The specific implementation is as follows:

[0033] (1) For each operation mode time segment obtained in step 2, the average load of each time segment is calculated as the load data for operation mode analysis.

[0034] (2) Taking feeder load balancing as the optimization goal, the operation mode within a certain time interval is analyzed. The specific process is as follows:

[0035] a: Initialization: define the outgoing line switch path length: the sum of all segment loads on the path, close all segment switches in the feeder topology island,

[0036] b: Longest path search: Calculate the longest path between the outgoing line switches in the feeder topology island, record the segment switches on the path, turn on the segment switches,

[0037] c: Based on the principle of balancing the segment load to the 10 kV outgoing line, set the optimal segment switch in the longest path to open.

[0038] d: Determine whether the path lengths between all outgoing line switches are 0, that is, whether the feeders are all in the radiation operation state. If so, go to process e. If there are still connections between the outgoing line switches, go to process b.

[0039] e: The segment switches turned on during the recording process are the analysis results of the operation mode in this time period.

[0040] (3) For all feeder topology islands, the operation modes in all time intervals are calculated, and the same operation modes in adjacent time intervals are merged as the final operation mode analysis results.

[0041] Step 4: Under each grid operation mode, the load under each operation mode is used as the known operation data, and the power supply margin of the grid to the distributed energy is iteratively calculated to obtain the final power supply margin result of the grid to the distributed energy. In step 4, the analysis object is a single radially operated 10 kV feeder under the condition of known grid operation mode, and the power supply margin of the equipment load rate, short-circuit current, and equipment voltage drop to the distributed energy is calculated respectively, and the power supply margin logic is allocated according to the parent-child relationship of the equipment under the radial operation state. The specific implementation is as follows:

[0042] (1) In step 3, under the condition that the grid operation mode is known, the state of the open section switch is known, and each 10 kV feeder is in the radial operation state, each radial operation 10 kV feeder is analyzed.

[0043] (2) Calculate the power supply margin of distributed energy for each device load factor.

[0044]

[0045] Where: S β Provide power supply margin for the load rate of the grid node

[0046] β is the allowable load rate of reverse flow, the default value is 80%

[0047] S T Transformer capacity

[0048] The default power factor

[0049] U N Rated voltage 10 kV

[0050] I N Current carrying wire

[0051] (3) Calculate the power supply margin of distributed energy due to the voltage increase of each device.

[0052] After distributed energy is connected to the grid, in the full power state, the power flow is reversed, and the end is converted from the power consumption side to the power supply side, and the voltage will increase significantly. According to the voltage loss formula:

[0053] ΔU=(R L *ΔP+X L *ΔQ) / U N According to the upper limit of 10 kV voltage + 7%

[0054] Reversely infer the capacity of distributed energy that each node is allowed to access:

[0055]

[0056] Where: S V Provide voltage margin for grid nodes

[0057] ΔU is the voltage rise allowed at the node

[0058] R L The equivalent resistance of the 10 kV outgoing line at the node distance

[0059] X L is the equivalent reactance of the 10 kV outgoing line at the node distance

[0060] The default tangent value of the distributed energy output power factor

[0061] (4) Calculate the power supply margin of distributed energy for the short-circuit current of each device.

[0062] After the distributed energy is connected to the grid, the feeder operates in a multi-power state, causing the short-circuit current to be too large. The distributed energy power supply margin under the short-circuit current constraint of each node is calculated based on the 10 kV short-circuit current allowable value:

[0063]

[0064] That is: S i =(I m -I xz )*U N

[0065] Where: Im is the upper limit of short-circuit current allowed by the 10 kV power grid

[0066] Ixz is the calculated value of the current short-circuit current of the 10 kV power grid node

[0067] U N Rated voltage

[0068] (5) Considering the three power supply margins of equipment load rate, short-circuit current and voltage drop, the minimum value is taken as the power supply margin of distributed energy at each node of the power grid.

[0069] S pv =min(S β , S v , S i )

[0070] Where: S PV The final power supply margin allowed for a node in a 10 kV power grid

[0071] (6) According to the parent-child hierarchical relationship of the feeder equipment under the radiation operation state, the power supply margin of each feeder to the distributed energy is logically allocated.

[0072]

[0073] S′ pv The final power supply margin of the node

[0074] S′ pv Parent is the final power supply margin of the parent node

[0075] n is the number of child nodes under the parent node

[0076] K 同时 Simultaneous coefficient

[0077] S pv子 Provides power margin to child nodes before logic allocation.

[0078] Step 5: Comprehensively consider all operation modes and obtain the minimum value of the distributed energy power supply margin of each device as the final calculation result of the distributed energy power supply margin of the feeder.

[0079] Step five includes step A and step B.

[0080] In step A, based on the typical day operation mode plan of the power grid analyzed in step 3, the power supply margin calculation in step 4 is performed for each operation mode, thereby obtaining the power supply margin results of each node of the power grid to distributed energy under each operation mode.

[0081] In step B, the power supply margin results of each node in the power grid and under each operation mode are compared, and the minimum power supply margin is selected as the final calculation result of the power supply margin of the feeder to the distributed energy.

[0082] This data-driven dynamic power supply margin analysis method for the power grid takes into account the status of the 10 kV section interconnection switch and the operating mode between the feeders, the time characteristics of the distributed energy output and the load characteristics of each node in the power grid, so as to analyze the operating mode of the power grid in different time periods and the power supply margin of the power grid to the distributed energy. The analysis of the dynamic power supply margin of the power grid is more accurate and comprehensive, and the analysis efficiency is also relatively high.

[0083] The above technical solutions only reflect the preferred technical solutions of the technical solutions of the present invention. Some changes that may be made to certain parts thereof by technicians in this technical field all reflect the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. A data-driven power grid dynamic power supply margin analysis method, characterized in that: The following steps are involved: Step 1: Obtain 10 kV topology data, sort out 10 kV large feeder and segment information, and count the load information of each segment; Step 2: According to the 24-hour load data of each segment, iteratively set the number of clusters, perform time-based cluster analysis, and obtain the clustering results of the typical time interval of the representative day; Step 3: Consider the constraints on safe operation of the power grid and calculate the optimal operation mode recommendation within the typical time interval; Step 4: Under each grid operation mode, taking the load under each operation mode as known operation data, iteratively calculate the distributed energy power supply margin of the grid to obtain the final grid power supply margin result for distributed energy; Step 5: Comprehensively consider all operation modes and obtain the minimum value of the power supply margin of each device to distributed energy as the final calculation result of the power supply margin of the feeder to distributed energy; In the step 1, according to the 10 kV topology data, a topology search is performed on the interconnected 10 kV lines with the 10 kV busbar of the upper substation as the boundary; In step 1, the sectionalizing switch is used as the boundary to sort out the sectionalizing units inside the interconnected 10 kV line, and the load data in each sectionalizing unit is counted; In the step 2, the load data of the 24-hour segment counted in the step 1 is clustered based on time using the k-means algorithm to analyze the time segments with the same load characteristics on the representative day, and each time segment corresponds to an operation mode; In the step 3, taking feeder load balancing as the optimization goal and feeder load rate and voltage not exceeding the limit as the constraint conditions, the operation modes in all time intervals calculated in the step 2 are analyzed and calculated, and the operation modes in adjacent time intervals and the same are merged; In the step 4, the analysis object is a single 10 kV feeder in radial operation under the condition of known grid operation mode, and the power supply margin of distributed energy for equipment load rate, short-circuit current, and equipment voltage drop are calculated respectively, and the power supply margin is logically allocated according to the parent-child relationship of the equipment in the radial operation state; The step five includes step A and step B; In step A, based on the typical day operation mode of the power grid analyzed in step 3, the power supply margin calculation in step 4 is performed for each operation mode, thereby obtaining the power supply margin result of each node of the power grid to the distributed energy under each operation mode; In step B, the power supply margin results of each node in the power grid and under each operation mode are compared, and the minimum power supply margin is selected as the final calculation result of the power supply margin of the feeder to the distributed energy.

Images