A method and system for evaluating the accuracy of power grid simulation based on fault inversion
Through the combination of Frechet distance and electrical distance, the waveform difference quantification problem in the accuracy evaluation of the grid simulation model is solved, the accuracy evaluation of the grid simulation model is improved, and more accurate reduction of the operating status of the grid is realized and the simulation results can be compared. sex.
Patent Information
- Application Number
- CN202411067257.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-06
AI Technical Summary
The prior art cannot effectively quantify the differences between the fault recording waveforms and simulation waveforms of different nodes, which affects the accuracy evaluation of the power grid simulation model.
Using the Frechet distance method, the waveform difference between the voltage recording waveforms and simulation waveforms of the fault node and the first-level cross-section node is calculated. The voltage simulation waveforms of the power grid node at the time of failure are obtained through online offline fusion technology, and the simulation accuracy is calculated using Frechet distance. Indicators, weighted averages combined with electrical distances to evaluate grid simulation accuracy.
By quantifying the evaluation of the simulation results and fault recording differences between the transient voltage of the AC bus, the waveform timing characteristics are fully considered, the accuracy evaluation results of the simulation model are improved, and the contribution of non-fault nodes to the overall simulation model of the power grid is comprehensively considered.
Smart Images

Figure CN118611056B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power grid analysis, and in particular to a method and system for evaluating the accuracy of power grid simulation based on fault inversion. Background Art
[0002] The accuracy of power grid simulation directly determines the safety and economy of power system operation. On the one hand, accurate power grid simulation results help power system dispatchers identify risk factors in the power grid, provide risk warnings, and avoid accidents. On the other hand, accurate power grid simulation results help determine whether the current power grid operation status is close to the safety and stability limit, fully tap the network's power transmission capacity, and improve the power system operation efficiency. Therefore, it is necessary to quantitatively evaluate the accuracy of power grid simulation and provide technical guidance for the safe and stable operation of the power grid.
[0003] The accuracy of power grid simulation is mainly evaluated by comparing the difference between fault recording and simulation results. Commonly used methods include methods based on waveform similarity and methods based on electrical characteristic quantity comparison. In the prior art, the paper "Verification Method for Short-Circuit Current Calculation Accuracy Based on Fault Recording Data" (Huazhong Electric Power, Zhang Xiaodong, Shi Guang, Henan Electric Power Research Institute, 2008) uses fault recording data as the standard for verifying the short-circuit current calculation value, and verifies the accuracy of the calculated value by comparing it with the calculated value through corresponding transformation. Since this verification method extracts the maximum instantaneous value and maximum effective value of the short-circuit current from the fault recording data, it does not make full use of the waveform information rich in the fault recording data, which is bound to affect the verification result of the accuracy of the calculated value. The paper "Research on Short-Circuit Current Characteristics of Doubly Fed Wind Turbine Groups and Their Impact on Line Protection" (Luo Meiling, North China Electric Power University, 2015) verifies the accuracy of the simulation model by comparing the fault recording waveform with the simulation waveform. The qualitative comparison method is used, and the similarity of the waveform cannot be quantified. The paper "Research and Application of High-power Hybrid Digital-Analog Simulation Interface System" (Leng Feng, Huazhong University of Science and Technology, 2021) uses Hausdorff distance to quantify the similarity between the fault recording waveform and the simulation waveform to evaluate the accuracy of the model. Since the Hausdorff distance can only represent the maximum distance between the two waveforms, it only focuses on the local extreme differences between the two waveforms and ignores the overall shape information between the waveforms. If the two waveforms are similar overall, but there is local misalignment or noise, this indicator cannot accurately measure the similarity between the two waveforms. Summary of the invention
[0004] The technical problem to be solved by the present invention is how to quantify the difference between the recorded waveforms and the simulated waveforms of faults at different nodes, so as to improve the accuracy of the evaluation of the power grid simulation model.
[0005] The present invention solves the above technical problems through the following technical solutions: a method for evaluating the accuracy of power grid simulation based on fault inversion, comprising the following steps:
[0006] Step 1: Based on the online and offline fusion operation mode at the time of the fault, the voltage simulation waveform of each node in the power grid at the time of the fault is obtained after fault inversion;
[0007] Step 2: According to the fault inversion results, the fault node is extracted from the voltage simulation waveform of each node in the power grid at the time of the fault. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes;
[0008] Step 3: Calculate the faulty node based on the Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node j The node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance ;
[0009] Step 4: Based on Frechet distance Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index ;
[0010] Step 5: Calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A .
[0011] The present invention uses Frechet distance to measure the difference between the fault recording waveform and the simulation waveform of different nodes, which is used to evaluate the accuracy of the simulation model. Based on the Frechet distance, the quantitative difference between the simulation result and the fault recording of the AC bus transient voltage is obtained, and the timing characteristics between different waveforms are fully considered, so that the quantitative evaluation result is more accurate.
[0012] Preferably, the step 1 comprises:
[0013] 1.1. Select the historical fault, obtain the fault time and fault node number, and export the online power flow mode file at the fault time from the full-time and space platform;
[0014] 1.2. Use the online and offline fusion technology to splice the node connection relationship, unit startup status, and load level represented by the online power flow mode file at the time of the fault into the typical offline power flow mode file to obtain the spliced typical offline power flow mode file;
[0015] 1.3. Using the power grid simulation analysis platform, perform power flow calculation and stability calculation on the spliced typical offline power flow mode file and the typical offline stability mode file to obtain the voltage simulation waveform of each node in the power grid at the time of the fault.
[0016] Preferably, the faulty node in step 3 i The recorded waveform and simulation waveform Frechet distance The calculation process includes:
[0017] 3.1.1. The waveform of the node voltage recording is The node voltage simulation waveform is For each sampling point, establish the connection matrix D:
[0018] (1)
[0019] 3.1.2. Calculate the Euclidean distance of each element in the connection matrix D Distance from Frechet :
[0020] when =1, =1, we have:
[0021] (2)
[0022] when From 2 to M, =1, we have:
[0023] (3)
[0024] when =1, When it grows from 2 to M, we have:
[0025] (4)
[0026] According to formulas (2)-(4), the Frechet distance of each element in the first column and the first row of the connection matrix D is calculated;
[0027] In the connection matrix D, from =2, =2, and search forward according to formula (5):
[0028] (5)
[0029] until =M, =M, get the faulty node i The recorded waveform and simulation waveform Frechet distance ,
[0030] .
[0031] Preferably, the first-level section node in step 3 j The recorded waveform and simulation waveform Frechet distance The calculation process includes:
[0032] 3.2.1. The waveform of node voltage recording is The node voltage simulation waveform is Each sampling point establishes a connection matrix :
[0033] (7)
[0034] 3.2.2 Calculate the connection matrix The Euclidean distance of each element in Distance from Frechet :
[0035] when =1, =1, we have:
[0036] (8)
[0037] when From 2 to M, =1, we have:
[0038] (9)
[0039] when =1, When it grows from 2 to M, we have:
[0040] (10)
[0041] The connection matrix is calculated according to formulas (8)-(10): The Frechet distance between the first column and the first row of each element;
[0042] In the connection matrix In, from =2, =2, and search forward according to formula (11):
[0043] (11)
[0044] until =M, =M, get the first-level section node j The recorded waveform and simulation waveform Frechet distance , .
[0045] Preferably, the faulty node in step 4 i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node jThe recorded waveform and simulation waveform Simulation accuracy index .
[0046] Preferably, in step 5, the first-level section node j To the failed node i The electrical distance is expressed based on impedance.
[0047] Preferably, the first-level section node j To the failed node i The electrical distance is calculated as: ,in, Indicates a faulty node i With the first-level section node j The equivalent impedance between and Represents the faulty nodes i With the first-level section node j The self-impedance, and Indicates a faulty node i With the first-level section node j mutual impedance.
[0048] Preferably, the power grid simulation accuracy index A It is calculated by:
[0049] ,in, is an intermediate variable, representing the sum of the inverses of the equivalent impedances from the fault node to each primary section node. a For the failed node i The recorded waveform and simulation waveform Simulation accuracy index The weight of b For all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index The weight of .
[0050] Preferably, the faulty node i The recorded waveform and simulation waveform Simulation accuracy index Weight a The value is 1 / 2, all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index Weightb The value is 1 / 2.
[0051] The present invention also provides a power grid simulation accuracy evaluation system based on fault inversion, which adopts the power grid simulation accuracy evaluation method based on fault inversion, including:
[0052] The fault inversion module is used for the operation mode at the time of fault based on online and offline fusion. After fault inversion, the voltage simulation waveform of each node in the power grid at the time of fault is obtained;
[0053] The waveform extraction module is used to extract the fault node from the voltage simulation waveform of each node in the power grid at the time of the fault according to the fault inversion results. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes;
[0054] Frechet distance calculation module, used to calculate fault nodes based on Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node j The node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance ;
[0055] Node simulation accuracy index calculation module for Frechet distance-based Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index ;
[0056] Power grid simulation accuracy index calculation module, used to calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A .
[0057] The advantages provided by the present invention are:
[0058] 1. The present invention uses Frechet distance to measure the difference between the fault recording waveform and the simulation waveform of different nodes, which is used to evaluate the accuracy of the simulation model. Based on the Frechet distance, the quantitative difference between the simulation result and the fault recording of the AC bus transient voltage is obtained, and the timing characteristics between different waveforms are fully considered, so that the quantitative evaluation result is more accurate.
[0059] 2. The present invention uses the recorded data of multiple non-fault nodes to comprehensively evaluate the simulation accuracy, and performs weighted averaging on the results based on the electrical distance from the non-fault node to the fault node, fully considering the contribution of the waveform differences of the non-fault nodes to the accuracy of the overall power grid simulation model.
[0060] 3. The present invention is based on the online and offline fusion fault time operation mode generation technology, which can solve the problem that the power grid operation mode is difficult to obtain at the time of a fault, and accurately restore the operation state of the power grid at the time of the fault as the initial value of the simulation model, so that the simulation results are comparable with the fault recording. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Figure 1 A flow chart of a method for evaluating power grid simulation accuracy based on fault inversion provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0062] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the technical solution of the present invention is clearly and completely described below in combination with specific embodiments and with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0063] like Figure 1 As shown, this embodiment provides a method for evaluating the accuracy of power grid simulation based on fault inversion, comprising the following steps:
[0064] Step 1: Based on the online and offline fusion fault operation mode, the voltage simulation waveform of each node in the power grid at the fault time is obtained after fault inversion. Step 1 specifically includes the following steps:
[0065] 1.1. Select the historical fault, obtain the fault time and fault node number, and export the online power flow mode file at the time of the fault from the full-time and space platform.
[0066] 1.2. Use the online and offline fusion technology to splice the node connection relationship, unit startup status, and load level represented by the online power flow mode file at the time of the fault in step 1.1 into the typical offline power flow mode file to obtain the spliced typical offline power flow mode file.
[0067] 1.3. Using the power grid simulation analysis platform, perform power flow calculation and stability calculation on the typical offline power flow mode file and the typical offline stability mode file spliced in step 1.2 to obtain the voltage simulation waveform of each node in the power grid at the time of the fault.
[0068] Step 1 is based on the online and offline fusion of the fault time operation mode generation technology, which can solve the problem that the power grid operation mode is difficult to obtain at the time of the fault, accurately restore the power grid operation state at the time of the fault as the initial value of the simulation model, and make the simulation results comparable with the fault recording.
[0069] Step 2: According to the fault inversion results, the fault node is extracted from the voltage simulation waveform of each node in the power grid at the time of the fault. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes.
[0070] Step 3: Calculate the faulty node based on the Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node j The node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance .
[0071] For failed nodes i , the node voltage recording waveform is , the node voltage simulation waveform is , using the recursive method to calculate the faulty nodes i The node voltage recording waveform The node voltage simulation waveform is The Frechet distance includes the following steps:
[0072] 3.1.1. The waveform of the node voltage recording is The node voltage simulation waveform is For each sampling point, a connection matrix D is established.
[0073] (1)
[0074] 3.1.2. Calculate the Euclidean distance and Frechet distance of each element in the connection matrix D. Among them, Indicates sampling point The node voltage recording waveform With sampling point Node voltage simulation waveform The Euclidean distance between . Indicates the node voltage recording waveform Before The subsequence composed of sampling points and the node voltage simulation waveform forward The Frechet distance between subsequences consisting of sampling points, Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1, ..., M, where M is a natural number.
[0075] when =M, =M, This indicates a faulty node i The recorded waveform and simulation waveform Frechet distance .
[0076] when =1, =1, we have:
[0077] (2)
[0078] At this time, each of the two subsequences contains only one sampling point, and the node voltage recording waveform is A subsequence of a sampling point is expressed as , node voltage simulation waveform A subsequence of a sampling point is expressed as , two subsequences and The Frechet distance is equal to the Euclidean distance of the sampling points.
[0079] when From 2 to M, =1, we have:
[0080] (3)
[0081] when =1, When it grows from 2 to M, we have:
[0082] (4)
[0083] According to formulas (2)-(4), the Frechet distance of each element in the first column and the first row of the connection matrix D is obtained by recursive calculation.
[0084] In the connection matrix D, from =2, =2, and search forward according to formula (5):
[0085] (5)
[0086] until =M, =M, get the faulty node i The recorded waveform and simulation waveform Frechet distance .
[0087] (6)
[0088] The first-level section node of the present invention j The recorded waveform and simulation waveform Frechet distance The calculation method and fault node i The recorded waveform and simulation waveform Frechet distance The same calculation method is used to finally obtain the first-level section node j The recorded waveform and simulation waveform Frechet distance .
[0089] , the specific calculation process is as follows:
[0090] For the first-level section node j , j =1, ..., J, where J represents the number of nodes in the first-level section. The waveform of the node voltage recording is , the node voltage simulation waveform is , the recursive method is used to calculate the first-level section nodes j The node voltage recording waveform The node voltage simulation waveform is Frechet distance.
[0091] 3.2.1. The waveform of node voltage recording is The node voltage simulation waveform is Each sampling point establishes a connection matrix .
[0092] (7)
[0093] 3.2.2 Calculate the connection matrix The Euclidean distance and Frechet distance of each element in . Indicates sampling point The node voltage recording waveform With sampling point Node voltage simulation waveform The Euclidean distance between . Indicates the node voltage recording waveform Before The subsequence composed of sampling points and the node voltage simulation waveform forward The Frechet distance between subsequences consisting of sampling points, Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1, ..., M, where M is a natural number.
[0094] when =M, =M, It can represent the first-level section node j The recorded waveform and simulation waveform Frechet distance .
[0095] when =1, =1, we have:
[0096] (8)
[0097] At this time, each of the two subsequences contains only one sampling point, and the node voltage recording waveform is A subsequence of a sampling point is expressed as , node voltage simulation waveform A subsequence of a sampling point is expressed as , two subsequences and The Frechet distance is equal to the Euclidean distance of the sampling points.
[0098] when From 2 to M, =1, we have:
[0099] (9)
[0100] when =1, When it grows from 2 to M, we have:
[0101] (10)
[0102] According to formulas (8)-(10), the connection matrix is obtained by recursive calculation The Frechet distance between the first column and the first row of each element in .
[0103] In the connection matrix In, from =2, =2, and search forward according to formula (11):
[0104] (11)
[0105] until =M, =M, get the first-level section node j The recorded waveform and simulation waveform Frechet distance .
[0106] (12)
[0107] Step 4: Based on Frechet distance Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index .
[0108] The smaller the Frechet distance between the recorded waveform and the simulated waveform, the higher the accuracy of the power grid simulation and the more accurate the fault node is. i The recorded waveform and simulation waveform Simulation accuracy index , .
[0109] Similarly, the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index , , j=1,…,J.
[0110] Step 5: Calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A .
[0111] First-level section node j To the failed node i The electrical distance is expressed based on impedance, ,in, Indicates a faulty node i With the first-level section node j The equivalent impedance between and Represents the faulty nodes i With the first-level section node j The self-impedance, and Indicates a faulty node i With the first-level section node j The present invention uses electrical distance to measure the electrical connection tightness between nodes. The smaller the electrical distance, the tighter the connection.
[0112] Assume a failed node i The recorded waveform and simulation waveform Simulation accuracy index The weight of a , all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index The weight of b , weight value a and b The fault node is adjusted according to the voltage level, load level and importance of the first-level section node. i The recorded waveform and simulation waveform Simulation accuracy index The weight is 1 / 2, all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index The weight of the simulation accuracy index is 1 / 2. The weight distribution of the simulation accuracy index of each first-level section node is determined by the inverse ratio of the electrical distance. ACalculated by the following formula:
[0113] ,in, It is an intermediate variable, representing the sum of the reciprocals of the equivalent impedance from the fault node to each primary section node, and is used to calculate the weight of the simulation accuracy index.
[0114] The present invention uses the Frechet distance to measure the difference between the fault recording waveform and the simulation waveform of different nodes, which is used to evaluate the accuracy of the simulation model. Based on the Frechet distance, the simulation result of the AC bus transient voltage and the quantitative difference between the fault recording are obtained. Compared with the traditional method, the timing characteristics between different waveforms can be fully considered, so that the quantitative evaluation result is more accurate. In addition, the recording data of multiple non-fault nodes are used to comprehensively evaluate the simulation accuracy, and the results are weighted averaged based on the electrical distance from the non-fault node to the fault node, fully considering the contribution of the waveform difference of the non-fault node to the accuracy of the overall simulation model of the power grid.
[0115] The present invention also provides a power grid simulation accuracy evaluation system based on fault inversion, which adopts a power grid simulation accuracy evaluation method based on fault inversion, including:
[0116] The fault inversion module is used for the operation mode at the time of fault based on online and offline fusion. After fault inversion, the voltage simulation waveform of each node in the power grid at the time of fault is obtained;
[0117] The waveform extraction module is used to extract the fault node from the voltage simulation waveform of each node in the power grid at the time of the fault according to the fault inversion results. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes;
[0118] Frechet distance calculation module, used to calculate fault nodes based on Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node jThe node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance ;
[0119] Node simulation accuracy index calculation module for Frechet distance-based Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index ;
[0120] Power grid simulation accuracy index calculation module, used to calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A .
[0121] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for evaluating the accuracy of power grid simulation based on fault inversion, characterized in that: The following steps are involved: Step 1: Based on the online and offline fusion operation mode at the time of the fault, the voltage simulation waveform of each node in the power grid at the time of the fault is obtained after fault inversion; Step 2: According to the fault inversion results, extract the fault node from the voltage simulation waveform of each node in the power grid at the time of the fault. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes; Step 3: Calculate the faulty node based on the Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node j The node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance ; Step 4: Based on Frechet distance Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index ; Step 5: Calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A ; The power grid simulation accuracy index A It is calculated by: ,in, is an intermediate variable, representing the sum of the inverses of the equivalent impedances from the fault node to each primary section node. a For the faulty node i The recorded waveform and simulation waveform Simulation accuracy index The weight of b For all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index The weight of Indicates a faulty node i With the first-level section node j The equivalent impedance between.
2. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: The step 1 comprises: 1.
1. Select the historical fault, obtain the fault time and fault node number, and export the online power flow mode file at the fault time from the full-time and space platform; 1.
2. Use the online and offline fusion technology to splice the node connection relationship, unit startup status, and load level represented by the online power flow mode file at the time of the fault into the typical offline power flow mode file to obtain the spliced typical offline power flow mode file; 1.
3. Using the power grid simulation analysis platform, perform power flow calculation and stability calculation on the spliced typical offline power flow mode file and the typical offline stability mode file to obtain the voltage simulation waveform of each node in the power grid at the time of the fault.
3. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: the fault node in step 3 i The recorded waveform and simulation waveform Frechet distance The calculation process includes: 3.1.
1. The waveform of the node voltage recording is The node voltage simulation waveform is For each sampling point, establish the connection matrix D: (1) 3.1.
2. Calculate the Euclidean distance of each element in the connection matrix D Distance from Frechet : when =1, =1, we have: (2) when From 2 to M, =1, we have: (3) when =1, When it grows from 2 to M, we have: (4) According to formulas (2)-(4), the Frechet distance of each element in the first column and the first row of the connection matrix D is calculated; In the connection matrix D, from =2, =2, and search forward according to formula (5): (5) until =M, =M, get the faulty node i The recorded waveform and simulation waveform Frechet distance , .
4. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: The first-level section node in step 3 j The recorded waveform and simulation waveform Frechet distance The calculation process includes: 3.2.
1. The waveform of node voltage recording is The node voltage simulation waveform is Each sampling point establishes a connection matrix : (7) 3.2.
2. Calculate the connection matrix The Euclidean distance of each element in Distance from Frechet : when =1, =1, we have: (8) when From 2 to M, =1, we have: (9) when =1, When it grows from 2 to M, we have: (10) The connection matrix is calculated according to formulas (8)-(10): The Frechet distance between the first column and the first row of each element; In the connection matrix In, from =2, =2, and search forward according to formula (11): (11) until =M, =M, get the first-level section node j The recorded waveform and simulation waveform Frechet distance , .
5. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: The faulty node in step 4 i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index .
6. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: The first-level section node in step 5 j To the failed node i The electrical distance is expressed based on impedance.
7. The power grid simulation accuracy evaluation method based on fault inversion according to claim 6 is characterized in that: First-level section node j To the failed node i The electrical distance is calculated as: ,in, Indicates a faulty node i With the first-level section node j The equivalent impedance between and Represents the faulty nodes i With the first-level section node j The self-impedance, and Indicates a faulty node i With the first-level section node j mutual impedance.
8. The power grid simulation accuracy evaluation method based on fault inversion according to claim 1 is characterized in that: Faulty Node i The recorded waveform and simulation waveform Simulation accuracy index Weight a The value is 1 / 2, all first-level section nodes j The recorded waveform and simulation waveform Simulation accuracy index Weight b The value is 1 / 2.
9. A power grid simulation accuracy evaluation system based on fault inversion, using the power grid simulation accuracy evaluation method based on fault inversion according to any one of claims 1 to 8, characterized in that: include: The fault inversion module is used for the operation mode at the time of fault based on online and offline fusion. After fault inversion, the voltage simulation waveform of each node in the power grid at the time of fault is obtained; The waveform extraction module is used to extract the fault node from the voltage simulation waveform of each node in the power grid at the time of the fault according to the fault inversion results. i The node voltage recording waveform And node voltage simulation waveform , and the first-level section nodes directly connected to the fault node j The node voltage recording waveform And node voltage simulation waveform , Indicates the sampling length of the node voltage recording waveform, =1,...,M, represents the sampling length of the node voltage simulation waveform, =1,...,M, where M is a natural number, j =1,...,J, J represents the number of first-level section nodes; Frechet distance calculation module, used to calculate fault nodes based on Frechet distance i The node voltage recording waveform And node voltage simulation waveform The waveform difference is calculated to calculate the first-level section node j The node voltage recording waveform And node voltage simulation waveform The waveform difference is obtained to get the fault node i The recorded waveform and simulation waveform Frechet distance , and the first-level section nodes j The recorded waveform and simulation waveform Frechet distance ; Node simulation accuracy index calculation module for Frechet distance-based Calculate the faulty node i The recorded waveform and simulation waveform Simulation accuracy index , based on the Frechet distance Calculate the first-level section node j The recorded waveform and simulation waveform Simulation accuracy index ; Power grid simulation accuracy index calculation module, used to calculate the first-level section nodes j To the failed node i The electrical distance to the fault node i The recorded waveform and simulation waveform Simulation accuracy index , first-level section node j The recorded waveform and simulation waveform Simulation accuracy index Calculate the weighted average to obtain the power grid simulation accuracy index A .
