Site selection and capacity optimization configuration method for distributed phase modifier

Abstract

The invention relates to the technical field of reactive power compensation device configuration of a new energy station, and discloses a distributed phase modifier site selection and capacity optimization configuration method, which comprises the following steps of: acquiring a joint data set based on operation data and simulation data of an AHP (Analytic Hierarchy Process); a wind power abnormal data elimination method combining a GMM clustering algorithm and a quartile method is constructed to analyze and screen wind power data, then a comprehensive weight model based on an improved sensitivity index and a new energy field station short circuit ratio is constructed, and a distributed phase modifier location hierarchical configuration scheme is generated. Meanwhile, a self-adaptive simulated annealing genetic algorithm optimization correction model is constructed, and the capacity of the distributed phase modifier is dynamically corrected; and finally, dynamically adjusting the number of merged nodes of the distributed phase modifier, solving the problem that the distributed phase modifier is frequently merged into and removed from the system, and having great significance for realizing stable operation of a high-proportion new energy sending-end power grid system.

Description

Technical Field

[0001] The present invention relates to the technical field of reactive power compensation device configuration of a new energy station, and in particular to a method for site selection and capacity optimization configuration of a distributed phase regulator. Background Art

[0003] With the growth of wind power installed capacity, the impact of wind power transmission on the power system has become increasingly prominent. On the one hand, large-scale centralized access of new energy to replace thermal power units has led to a decrease in the short-circuit ratio of new energy stations and a weakening of the voltage support capacity of the sending-end nodes; on the other hand, areas with rapid development of new energy are often located at the end of the interconnected power grid, with weak grid structure, small local load, lack of sufficient reactive power support, and difficulty in withstanding the impact of grid disturbances. After a fault, the transient voltage fluctuations at the sending end are severe. Therefore, how to improve the short-circuit ratio of high-proportion wind power systems, reduce transient voltage fluctuations, increase installed capacity, and enhance the transmission capacity of the sending-end system has become one of the urgent issues to be solved. Summary of the invention

[0004] The purpose of the present invention is to solve the above problems and to design a method for site selection and capacity optimization configuration of distributed phase regulators.

[0005] The technical solution of the present invention to achieve the above-mentioned purpose is a method for site selection and capacity optimization configuration of a distributed phase regulator, the method comprising the following steps:

[0006] Step 1: construct a high-voltage direct current transmission network simulation model including multiple renewable energy wind farms at the sending end, and obtain operation data and simulation data of the high-voltage direct current transmission network simulation model;

[0007] Step 2: Process the operation data and simulation data in the HVDC transmission network simulation model based on the AHP hierarchical analysis method, and obtain the joint wind power data set;

[0008] Step 3: Based on the wind power abnormal data elimination model combining the GMM clustering algorithm and the quartile method, the abnormal wind power data is eliminated and a complete comprehensive wind power data set is obtained;

[0009] The abnormal wind power data processing process includes: removing outliers in the data set through the GMM clustering algorithm; processing the clustered and stratified sample points twice through the quartile method;

[0010] Step 4: embed the processed data into the pscad simulation model to generate target data, obtain the simulation operation results at the same time, and determine the fault form of the wind turbine off-grid;

[0011] It should be noted that the processing of step 4 is based on the specific fault scenario of wind turbine trawling, which solves the existing electrical faults and ensures the normal and stable operation of the HVDC transmission system model.

[0012] Step 5: Calculate the short-circuit ratio and improved sensitivity index of each node in the new energy station, and determine the key quantity ratio of the short-circuit ratio and the improved sensitivity index in each node, and then obtain the comprehensive value score of each node, and determine the site selection node of the distributed phase-shifting machine according to the comprehensive value score of the node;

[0013] It should be noted that the method for determining the site selection of distributed phase-shifting units includes: for nodes with a short-circuit ratio greater than 2.5, only the improved sensitivity index is considered; for nodes with an improved sensitivity index less than 0.005, only the short-circuit ratio of the station is considered; and for the remaining nodes, two weights are considered comprehensively. By calculating the comprehensive value score of each node, a distributed phase-shifting unit site selection hierarchical configuration plan is generated based on the node score size;

[0014] Among them, the method for determining the proportion of short-circuit ratio and improved sensitivity index includes: according to the improved sensitivity index of each node and the short-circuit ratio of new energy stations, the entropy weight method is adopted to propose a weight design method for the improved sensitivity index and the short-circuit ratio of new energy stations, and a weight model for the improved sensitivity index and the short-circuit ratio of new energy stations is constructed to calculate the weights of the two typical parameters.

[0015] Step 6: Establish a distributed phase-converter capacity optimization and correction model, use a dynamically adaptive simulated annealing genetic algorithm to optimize the correction model, take the transient voltage fluctuation of the rectifier-side bus as the voltage safety constraint, generate a distributed phase-converter capacity optimization configuration plan and optimize the distributed phase-converter capacity;

[0016] It should be noted that the capacity optimization method of distributed phase-converter includes: based on the actual operation and maintenance topology of the high-proportion wind power sending-end system, construct three operating conditions simulation models of three-phase short circuit fault, two-phase ground short circuit fault, and single-phase ground short circuit fault, establish a distributed phase-converter capacity optimization correction model, and use the dynamic adaptive simulated annealing genetic algorithm to optimize the correction model, take the transient voltage fluctuation of the rectifier-side bus as the voltage safety constraint, generate the distributed phase-converter capacity optimization configuration plan, and further optimize the distributed phase-converter capacity;

[0017] Step 7: Based on the above-mentioned distributed phase-converter capacity optimization configuration scheme, the switch matrix is ​​set according to different operating conditions, and the number of distributed phase-converters incorporated into the node is dynamically adjusted;

[0018] It should be noted that based on the different numbers of nodes connected to distributed phase-shifters under different working conditions, this application sets up five groups of switch matrices to dynamically adjust the number of nodes connected to distributed phase-shifters. The process includes: combining different levels of fault combinations, controlling the switch matrix, and simulating the actual operation and maintenance to connect the number of distributed phase-shifters.

[0019] The operation data and simulation data in step 1 include but are not limited to: wind turbine access location and multiplication model parameter data, conventional thermal power installed capacity data, power grid load parameters, grid-connected line parameter data, distributed phase regulators, SVG and other reactive compensation device configuration data, 35KV, 500KV, 220KV bus voltage level data, DC transmission system network architecture data, wind farm side lines, and transmission side rectifier inverter device operation data.

[0020] The process of processing the operation data and simulation data based on the AHP hierarchical analysis method in step 2 is as follows:

[0021] The synthetic weight of the simulated wind power data and the actual wind power data is determined by the hierarchical analysis method, and the synthetic weight formula is:

[0022]

[0023] Where: P ij represents the weight of the ith factor of the P layer to the total target, D i Represents a hierarchical matrix.

[0024] The process of implementing abnormal wind power data elimination by the wind power abnormal data elimination model in step 3 is as follows:

[0025] Let Φ(X|φ m ) is the power parameter probability density function selected in the GMM algorithm, and its specific calculation formula is as follows:

[0026]

[0027] Where: m represents the mth Gaussian distribution covariance matrix;

[0028] The power parameter probability density function is used to preliminarily process the abnormal wind power data, and then the quartile method is used to perform secondary processing on the abnormal wind power data. The secondary processing process is as follows:

[0029] First, the hybrid wind power data set is processed horizontally based on the quartile method. The wind power of the data set is divided into four equal parts to obtain the upper score Q1, lower score Q3, and median score Q2, and the interquartile range I is calculated. QR and data acceptance interval [F d ,F u ], then the calculation formula is as follows:

[0030] I QR =Q3-Q1 (3)

[0031] [F d ,F u ]=[Q1-1.5I QR ,Q3+1.5IQR ] (4)

[0032] Then, the mixed data are processed longitudinally based on the quartile method, the abnormal data are eliminated, and the abnormal power data are replaced by the average power data to obtain a complete comprehensive wind power data set.

[0033] The wind turbine off-grid failure forms in step 4 include:

[0034] Lack of low voltage ride-through capability, lack of high voltage ride-through capability, and insufficient reactive power regulation capability.

[0035] The calculation process of the short circuit ratio and improved sensitivity index of each node in step 5 is as follows:

[0036] The formula for calculating the node short circuit ratio is:

[0037]

[0038] Where: U Ni The voltage at the grid connection point of the new energy station. is the impedance of the ith row and jth column of the busbar grid-connected point equivalent impedance matrix, P REj Active power injected into the prime mover;

[0039] The calculation model of improved sensitivity index is established, and the formula for calculating the value of node improved sensitivity index is:

[0040]

[0041] Where: I TSj(SC) To configure the distributed phase regulator back node B j Improved trajectory sensitivity index; V i (t k ,Q j0 +ΔQ jSC ) is a typical fault condition B j The node configuration capacity is S SC Distributed phase condenser after t k Node voltage at time V i ; V i (t k ,Q j0 ) is B j The node configuration capacity is S SC Distributed phase condenser current t k Node voltage at time V i ; ΔQ jSC For B j Node distributed phase regulator k Reactive power at the moment.

[0042] The calculation process of the comprehensive value score of each node in step 5 is:

[0043] Construct a weight model of improved sensitivity index and short-circuit ratio of new energy stations, and calculate the weights of the two typical parameters;

[0044] Assume that the short-circuit ratio model of the new energy station after the distributed phase regulator and its system power flow constraints are:

[0045] MRSCR i >MRSCR min (7)

[0046] 0≤k i ≤k max (8)

[0047]

[0048] i=1,2,3....n (11)

[0049] Where: MRSCR i is the short-circuit ratio of renewable energy wind farm i, MRSCR min is the lower limit threshold of the short-circuit ratio of the station, k max Configure the upper limit of the number of distributed phase regulators for the station nodes, P i , Q i are the active power and reactive power of node i, V i 、V j Represent the voltage of the i-th node and the j-th node respectively, G ij With B ij Respectively represent the real and imaginary parts of the node (i, j) admittance matrix, θ ij represents the phase difference between nodes i and j;

[0050] The two parameters are standardized. The larger the value of the improved sensitivity index is, the more ideal it is. The standardized processing is performed using formula (12). The nodes with low short-circuit ratio of the station are nodes with weak voltage support capacity. The standardized processing is performed using formula (13).

[0051]

[0052] Where x is the jth index value, x max is the maximum value of the j-th index, x min is the minimum value of the j-th index, r ij is the standardized value;

[0053] Calculate the weight p of the index value of the i-th scheme under the j-th index ij :

[0054]

[0055] Calculate the entropy weight e of the jth indicator j :

[0056]

[0057] in,

[0058]

[0059] The information entropy redundancy is:

[0060] g j =1-e j (17)

[0061] Calculate the weights of the two influencing factors separately:

[0062]

[0063] The comprehensive score of each station node is calculated based on the improved sensitivity index and the station sensitivity weight:

[0064]

[0065] The distributed phase-shifting capacity optimization correction model established in step 6 is:

[0066]

[0067] Where: n1 is the number of distributed phase-shifting nodes that meet the short-circuit ratio requirements of new energy stations in one stage, K i is the number of distributed phase regulators configured after the capacity correction of wind farm i, k i Configure the number of distributed phase regulators for wind farm i that meets the short-circuit ratio requirements of renewable energy stations in a phase, S i is the capacity of the distributed phase-shifting machine, and γ is the penalty coefficient.

[0068] In step 6, the process of optimizing the modified model by dynamically adaptive simulated annealing genetic algorithm and taking the transient voltage fluctuation of the rectifier-side bus as the voltage safety constraint to generate the optimal configuration scheme of the distributed phase-converter capacity is as follows:

[0069] 1): A fixed-length symbol encoding method is used for chromosome encoding. Each gene in the chromosome corresponds to the number of distributed phase shifters in the station. The two states of "1" and "0" and their superposition states are used to represent the number of distributed phase shifters initially configured in each wind farm;

[0070] 2): Define the initial population as:

[0071] W=[p1,p2,...p n ] (twenty one)

[0072] Where: p1, p2, ... p n Chromosome encoding of the number of distributed phase regulators initially configured for n wind farm nodes in a phase;

[0073] Initialize a population W of size N i , where the number of distributed phase regulators represented by the population chromosome gene cannot be less than the number of phase regulators configured in one stage;

[0074] 3): Initial population W i Substitute into formula (21) to calculate, according to the fitness function and constraints, detect the performance of individuals in the population, reproduce the individuals with good performance to produce offspring, and eliminate the individuals with poor performance;

[0075] 4): The initial population is subjected to genetic algorithm selection, crossover and mutation process to obtain subgroup Q i , merge the parent population with the child population, and calculate the individual objective function value of the new population;

[0076] 5): Introduce the simulated annealing algorithm, use the Metropolis criterion, and compare the fitness function values ​​J2(x2) and J2(x1) of the two generations of individuals;

[0077] If J2(x2)<J2(x1), then accept the new generation; otherwise, according to Probability of accepting a new generation of offspring;

[0078] 6): Iterate step by step until the number of iterations is met, and output the optimal configuration capacity of the camera for the specified node.

[0079] The voltage safety constraint in step 6 is:

[0080]

[0081] i=1,2,3....n (24)

[0082] 0≤k i ≤k max (25)

[0083] V min (t)≤V(t)≤V max (t) (26)

[0084] Where: V min (t) is the lower limit of transient voltage fluctuation, V max (t) is the upper limit of transient voltage fluctuation.

[0085] Compared with the prior art, this application has the following advantages:

[0086] 1. This application ensures the validity and representativeness of the wind power data imported into the model by processing a mixed data set of wind power data, which helps to enhance the accuracy of the site selection and capacity optimization method of distributed phase-converter;

[0087] 2. This application considers the short-circuit ratio and improved sensitivity index of new energy stations, adopts a distributed phase-shifting phase-shifting layered site selection strategy based on the entropy weight method, improves the efficiency of reactive resource allocation, and prolongs the service life of distributed phase-shifting phases;

[0088] 3. This application adopts a dynamically adaptive distributed phase-shifting capacity optimization method to improve the stability of high-voltage power transmission, and at the same time applies a switch matrix to reduce maintenance costs in long-term operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a flow chart of a method for site selection and capacity optimization configuration of a distributed phase regulator according to the present invention;

[0090] Figure 2 It is a topological diagram of a high-proportion wind power transmission-end power grid with multiple voltage levels as described in the present invention;

[0091] Figure 3 It is a topological diagram of the external transmission structure of the ultra-high voltage direct current transmission system of the present invention;

[0092] Figure 4 is a scatter plot after the hybrid wind power data of the present invention is processed;

[0093] Figure 5 It is a comparison diagram of node voltage support effects before and after the site selection of the distributed phase regulator based on various fault conditions described in the present invention;

[0094] Figure 6 This is a comparison diagram of the improvement effect of the rectifier-side bus voltage before and after the distributed phase-converter capacity optimization based on various fault conditions described in the present invention;

[0095] Figure 7 This is a topological diagram of the present invention for controlling the incorporation and removal of distributed phase regulators using a switch matrix. DETAILED DESCRIPTION

[0096] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1-7 As shown;

[0097] First, if Figure 2 and Figure 3As shown: Based on the PSCAD simulation model, a ±800KV high-voltage DC transmission network architecture is constructed. The receiving-end converter station adopts fixed DC voltage and fixed trigger angle control, and the sending-end adopts fixed power control; three sending-end grid node voltage levels of 35kv, 220kv, and 500kv are set; 6 groups of new energy wind turbines are integrated based on traditional thermal power units, and the wind turbines all adopt the x50 aggregation multiplication model, and dual-loop control is adopted on the wind turbine side and the grid side; a single-pole 24-pulse rectifier is used inside the converter valve.

[0098] Secondly, a hierarchical model based on the combination of actual wind power data and simulated wind power data was constructed, and the economic factors, risk factors, and data reference value factors were decomposed into wind turbine operation and maintenance costs, site temporary engineering costs, and simulation platform costs; natural risk factors, platform operation risk factors; actual wind farm operation data, considering the simulated operation data under each operation scenario, and the relative importance of each pair was compared to obtain a comprehensive decision-making result. Among them, economic factors, risk factors, and data reference value factors accounted for 0.17, 0.13, and 0.7, respectively.

[0099] The formula for determining the synthetic weight of simulated wind power data and actual wind power data by analytic hierarchy process is:

[0100]

[0101] Where: P ij represents the weight of the ith factor of the P layer to the total target, D i Represents a hierarchical matrix.

[0102] Substituting the numerical values ​​into the calculation, it can be obtained that the selection ratio of actual wind power data and simulated wind power data is 0.6512 and 0.348.

[0103] Then, the GMM clustering algorithm is used to select power as the variable. The power is affected by multiple factors such as wind speed, wind turbine rotor current, and doubly fed wind turbine parameters. The probability density function of power and its characteristics are used to divide the clusters and eliminate the discrete points in the wind power data set. m ) is the power parameter probability density function selected in the GMM algorithm, and its specific calculation formula is as follows:

[0104]

[0105] Where: m Represents the m-th Gaussian distribution covariance matrix.

[0106] For the data after preliminary processing, the quartile method was used to deal with abnormal data;

[0107] Based on the quartile method, the hybrid wind power data set is processed horizontally, the power parameters are divided into 60MW units, and the wind power of the data set is divided into four equal parts based on the power value size, and the upper score Q1, lower score Q3, and median score Q2 are obtained, and the interquartile range I is calculated. QR and data acceptance interval [F d ,F u ], and use this as the standard to retain wind power data. The calculation formula is as follows:

[0108] I QR =Q3-Q1 (3)

[0109] [F d ,F u ]=[Q1-1.5I QR ,Q3+1.5I QR ] (4)

[0110] The mixed data is processed vertically based on the quartile method. The wind speed data is divided into units of 1m / s, the corresponding wind power data is retained in the wind speed data acceptance interval, and the abnormal data is further proposed.

[0111] At the same time, in order to maintain data integrity, the average power data is used to replace the abnormal power data to improve the comprehensive wind power data set. After removing 19.08% of the abnormal data, the power scatter plot is as follows: Figure 4 shown.

[0112] The processed data is embedded in the pscad simulation model to generate target data, and stable characteristic simulation is performed at the same time to obtain simulation operation results and determine the wind turbine off-grid failure form under various operation scenarios.

[0113] Subsequently, the improved sensitivity value and the short-circuit ratio value of each node are calculated. For nodes with a short-circuit ratio greater than 2.5, only the improved sensitivity index is considered, that is, the improved sensitivity index takes a proportion of 1 and the short-circuit ratio index takes a proportion of 0; for nodes with an improved sensitivity index less than 0.005, only the short-circuit ratio index of the station is considered, that is, the short-circuit ratio index of the station takes a proportion of 1 and the improved sensitivity index takes a proportion of 0; the two weights are combined to calculate the comprehensive value score of each node, and based on the size of the node score, a distributed phase-shifting phase selection hierarchical configuration plan is generated;

[0114] Calculate the short-circuit ratio of the new energy station at the wind turbine access node of the transmission system simulation model. The formula is:

[0115]

[0116] Where: U Ni The voltage at the grid connection point of the new energy station. is the impedance of the ith row and jth column of the busbar grid-connected point equivalent impedance matrix, PREj Active power injected into the prime mover.

[0117] Obtain the node voltage change and reactive power change before and after the target node is configured with distributed phase regulators, establish an improved sensitivity index calculation model, and calculate the improved sensitivity index value of the wind turbine access node within the research range selected by the transmission system simulation model. The formula is:

[0118]

[0119] Where: I TSj(SC) To configure the distributed phase regulator back node B j Improved trajectory sensitivity index; V i (t k ,Q j0 +ΔQ jSC ) is a typical fault condition B j The node configuration capacity is S SC Distributed phase condenser after t k Node voltage at time V i ; V i (t k ,Q j0 ) is B j The node configuration capacity is S SC Distributed phase condenser current t k Node voltage at time V i ; ΔQ jSC For B j Reactive power of the node distributed phase regulator at time tk.

[0120] For the remaining nodes, determine the key weights of the short-circuit ratio of new energy stations and the improved sensitivity index. According to the improved sensitivity index and the short-circuit ratio of new energy stations at each node, based on the entropy weight method model, propose a weight design method for the improved sensitivity index and the short-circuit ratio of new energy stations, construct a weight model for the improved sensitivity index and the short-circuit ratio of new energy stations, and calculate the weights of the two typical parameters.

[0121] The short-circuit ratio model of the new energy station after the distributed phase-converter and its system power flow constraints are as follows:

[0122] MRSCR i >MRSCR min (7)

[0123] 0≤k i ≤k max (8)

[0124]

[0125] i=1,2,3....n(11)

[0126] Where: MRSCR i is the short-circuit ratio of renewable energy wind farm i, MRSCR min k is the lower limit threshold of the short-circuit ratio of the station (the weak system threshold is 2.5). max Configure the upper limit of the number of distributed phase regulators for the station nodes. i , Q i are the active power and reactive power of node i, V i 、V j Represent the voltage of the i-th node and the j-th node respectively, G ij With B ij Respectively represent the real and imaginary parts of the node (i, j) admittance matrix, θ ij Represents the phase difference between nodes i and j.

[0127] The two parameters are standardized. The larger the value of the improved sensitivity index is, the more ideal it is. The standardized processing is performed using formula (12). The nodes with low short-circuit ratio of the station are nodes with weak voltage support capacity. The standardized processing is performed using formula (13).

[0128]

[0129] Where x is the jth index value, x max is the maximum value of the j-th index, x min is the minimum value of the j-th index, r ij is the standardized value.

[0130] Calculate the weight p of the index value of the i-th scheme under the j-th index ij :

[0131]

[0132] Calculate the entropy weight e of the jth indicator j :

[0133]

[0134] The information entropy redundancy is:

[0135] g j =1-e j (17)

[0136] Calculate the weights of the two influencing factors separately:

[0137]

[0138] The comprehensive score of each station node is calculated based on the improved sensitivity index and the station sensitivity weight:

[0139]

[0140] Arrange the comprehensive value scores of the nodes from large to small, taking into account the weak voltage support capability of the nodes and the compensation effect after configuration, determine the access location of the distributed phase-shifting machine, avoid unnecessary waste of reactive resources, and significantly improve the efficiency of reactive resource allocation. Figure 4 After the initial configuration of the site selection for the distributed phase regulator, a comparison chart of the voltage support capacity of the node is configured.

[0141] Based on the above analysis process, the present application can determine the capacity optimization method of distributed phase-shifting devices, including constructing three operating condition simulation models of three-phase short circuit fault, two-phase ground short circuit fault, and single-phase ground short circuit fault based on the actual operation and maintenance topology of the high-proportion wind power sending-end system, establishing a distributed phase-shifting device capacity optimization and correction model, and dynamically adapting the simulated annealing genetic algorithm to optimize the correction model. Taking the transient voltage fluctuation of the rectifier-side bus as a constraint, the distributed phase-shifting device capacity optimization configuration plan is generated to further optimize the distributed phase-shifting device capacity.

[0142] Among them, based on the transient voltage fluctuation characteristics, the design of the transient voltage instability control cost is minimized as the optimization goal, and the distributed phase-converter capacity correction objective function is constructed:

[0143]

[0144] Where: n1 is the number of distributed phase-shifting nodes that meet the short-circuit ratio requirements of new energy stations in one stage, K i is the number of distributed phase regulators configured after the capacity correction of wind farm i, k i Configure the number of distributed phase regulators for wind farm i that meets the short-circuit ratio requirements of renewable energy stations in a phase, S i is the capacity of the distributed phase-shifting machine, and γ is the penalty coefficient.

[0145] Considering the high and low voltage ride-through capabilities of the station wind turbine, in order to further limit transient voltage fluctuations, the voltage safety constraints of the transient response process are added as follows:

[0146]

[0147] i=1,2,3....n (23)

[0148] 0≤k i ≤k max (twenty four)

[0149] V min (t)≤V(t)≤V max (t) (25)

[0150] Where: V min (t) is the lower limit of transient voltage fluctuation, V max(t) is the upper limit of transient voltage fluctuation.

[0151] A short-circuit fault scan is performed on the new energy station that has completed the first-stage configuration of the distributed phase-converter to obtain the transient time-domain simulation characteristics of the wind farm in the sending-end power grid. Based on the bus voltage fluctuation trajectories of different wind farms under various short-circuit faults, a distributed phase-converter capacity optimization model using an improved simulated annealing genetic algorithm is designed to achieve capacity correction of the distributed phase-converter and improve the bus voltage fluctuation on the rectifier side under typical faults. The algorithm steps are as follows:

[0152] Algorithm step 1: Use fixed-length symbol encoding to encode chromosomes. Each gene in the chromosome corresponds to the number of distributed phase regulators in the site. The two states of "1", "0" and their superposition state are used to represent the number of distributed phase regulators initially configured in each wind farm.

[0153] Algorithm step 2: Define the initial population as:

[0154] W=[p1,p2,...p n ] (twenty one)

[0155] Where: p1, p2, ... p n Chromosome encoding of the number of distributed phase regulators initially configured for n wind farm nodes in one phase.

[0156] Initialize a population W of size N i , where the number of distributed phase regulators represented by the population chromosome genes cannot be less than the number of phase regulators configured in one stage.

[0157] Algorithm step 3: Initial population W i Substitute it into equation (21) and calculate it. According to the fitness function and constraints, the performance of individuals in the population is tested, that is, the ability of the wind farm to resist transient overvoltage after the initial configuration of distributed phase regulators is evaluated. Individuals with good performance reproduce to produce offspring, and individuals with poor performance are eliminated.

[0158] Algorithm step 4: The initial population is subjected to genetic algorithm selection, crossover and mutation process to obtain subgroup Q i , merge the parent population with the child population, and calculate the individual objective function value of the new population.

[0159] Algorithm step 5: In order to improve the global search capability, the simulated annealing algorithm is introduced, and the Metropolis criterion is used to compare the fitness function values ​​of the two generations of individuals J2(x2) and J2(x1). If J2(x2)<J2(x1), the new generation is accepted; otherwise, according to the probability Accepting a new generation of offspring.

[0160] Algorithm step 6: Iterate step by step until the number of iterations is met, and output the optimal configuration capacity of the camera for the specified node.

[0161] in, Figure 6 After optimizing the capacity of distributed phase-converter, the transient voltage fluctuation suppression effect of the rectifier-side busbar under different fault conditions is shown.

[0162] Step 7: If Figure 7 As shown, based on the above-mentioned distributed phase-shifting machine configuration and capacity optimization scheme, in order to meet the requirements of different operating conditions, a switch matrix is ​​set to dynamically adjust the number of distributed phase-shifting machines incorporated into the node.

[0163] 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 method for site selection and capacity optimization configuration of distributed phase condensers, characterized in that: The method comprises the following steps: Step 1: construct a high-voltage direct current transmission network simulation model including multiple renewable energy wind farms at the sending end, and obtain operation data and simulation data of the high-voltage direct current transmission network simulation model; Step 2: Process the operation data and simulation data in the HVDC transmission network simulation model based on the AHP hierarchical analysis method, and obtain the joint wind power data set; Step 3: Based on the wind power abnormal data elimination model combining the GMM clustering algorithm and the quartile method, the abnormal wind power data is eliminated and a complete comprehensive wind power data set is obtained; Step 4: embed the processed data into the pscad simulation model to generate target data, obtain the simulation operation results at the same time, and determine the fault form of the wind turbine off-grid; Step 5: Calculate the short-circuit ratio and improved sensitivity index of each node in the new energy station, and determine the key quantity ratio of the short-circuit ratio and the improved sensitivity index in each node, and then obtain the comprehensive value score of each node, and determine the site selection node of the distributed phase-shifting machine according to the comprehensive value score of the node; Step 6: Establish a distributed phase-converter capacity optimization and correction model, use a dynamically adaptive simulated annealing genetic algorithm to optimize the correction model, take the transient voltage fluctuation of the rectifier-side bus as the voltage safety constraint, generate a distributed phase-converter capacity optimization configuration plan and optimize the distributed phase-converter capacity; Step seven, based on the above-mentioned distributed phase-shifting unit capacity optimization configuration scheme, set the switch matrix according to different operating conditions and dynamically adjust the number of distributed phase-shifting units connected to the node.

2. A method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The operation data and simulation data in step 1 include but are not limited to: wind turbine access location and multiplication model parameter data, conventional thermal power installed capacity data, power grid load parameters, grid-connected line parameter data, distributed phase regulators, SVG and other reactive compensation device configuration data, 35KV, 500KV, 220KV bus voltage level data, DC transmission system network architecture data, wind farm side lines, and transmission side rectifier inverter device operation data.

3. The method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The process of processing the operation data and simulation data based on the AHP hierarchical analysis method in step 2 is as follows: The synthetic weight of the simulated wind power data and the actual wind power data is determined by the hierarchical analysis method, and the synthetic weight formula is: Where: P ij represents the weight of the ith factor of the P layer to the total target, D i Represents a hierarchical matrix.

4. The method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The process of implementing abnormal wind power data elimination by the wind power abnormal data elimination model in step 3 is as follows: Let Φ(X|φ m ) is the power parameter probability density function selected in the GMM algorithm, and its specific calculation formula is as follows: Where: m represents the mth Gaussian distribution covariance matrix; The power parameter probability density function is used to preliminarily process the abnormal wind power data, and then the quartile method is used to perform secondary processing on the abnormal wind power data. The secondary processing process is as follows: First, the hybrid wind power data set is processed horizontally based on the quartile method. The wind power of the data set is divided into four equal parts to obtain the upper score Q1, lower score Q3, and median score Q2, and the interquartile range I is calculated. QR and data acceptance interval [F d ,F u ], the calculation formula is as follows: I QR =Q3-Q1(3) <h2 style=";text-align:left;direction:ltr">[F<h2 style=";text-align:left;direction:ltr"> d <h2 style=";text-align:left;direction:ltr"> ,F<h2 style=";text-align:left;direction:ltr"> u <h2 style=";text-align:left;direction:ltr"> ]=[Q1-1.5I<h2 style=";text-align:left;direction:ltr"> QR <h2 style=";text-align:left;direction:ltr"> ,Q3+1.5I<h2 style=";text-align:left;direction:ltr"> QR <h2 style=";text-align:left;direction:ltr"> (4) Then, the mixed data are processed longitudinally based on the quartile method, the abnormal data are eliminated, and the abnormal power data are replaced by the average power data to obtain a complete comprehensive wind power data set.

5. The method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The wind turbine off-grid failure modes in step 4 include: lack of low voltage ride-through capability, lack of high voltage ride-through capability, and insufficient reactive power regulation capability.

6. The method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The calculation process of the short circuit ratio and improved sensitivity index of each node in step 5 is as follows: The formula for calculating the node short circuit ratio is: Where: U Ni The voltage at the grid connection point of the new energy station. is the impedance of the ith row and jth column of the busbar grid-connected point equivalent impedance matrix, P REj Active power injected into the prime mover; The calculation model of improved sensitivity index is established, and the formula for calculating the value of node improved sensitivity index is: Where: I TSj(SC) To configure the distributed phase regulator back node B j Improved trajectory sensitivity index; V i (t k ,Q j0 +ΔQ jSC ) is a typical fault condition B j The node configuration capacity is S SC Distributed phase condenser after t k Node voltage at time V i ; V i (t k ,Q j0 ) is B j The node configuration capacity is S SC Distributed phase condenser current t k Node voltage at time V i ; ΔQ jSC For B j Node distributed phase regulator k Reactive power at the moment.

7. A method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 6, characterized in that: The calculation process of the comprehensive value score of each node in step 5 is: Construct a weight model of improved sensitivity index and short-circuit ratio of new energy stations, and calculate the weights of the two typical parameters; Assume that the short-circuit ratio model of the new energy station after the distributed phase regulator and its system power flow constraints are: MRSCR i >MRSCR min (7) 0≤k i ≤k max (8) i=1,2,3....n (11) Where: MRSCR i is the short-circuit ratio of renewable energy wind farm i, MRSCR min is the lower limit threshold of the short-circuit ratio of the station, k max Configure the upper limit of the number of distributed phase regulators for the station nodes, P i , Q i are the active power and reactive power of node i, V i 、V j Represent the voltage of the i-th node and the j-th node respectively, G ij With B ij Respectively represent the real and imaginary parts of the node (i, j) admittance matrix, θ ij represents the phase difference between nodes i and j; The two parameters are standardized. The larger the value of the improved sensitivity index is, the more ideal it is. The standardized processing is performed using formula (12). The nodes with low short-circuit ratio of the station are nodes with weak voltage support capacity. The standardized processing is performed using formula (13). Where x is the jth index value, x max is the maximum value of the j-th index, x min is the minimum value of the j-th index, r ij is the standardized value; Calculate the weight p of the index value of the i-th scheme under the j-th index ij : Calculate the entropy weight e of the jth indicator j : in, The information entropy redundancy is: g j =1-e j (17) Calculate the weights of the two influencing factors separately: The comprehensive score of each station node is calculated based on the improved sensitivity index and the station sensitivity weight:

8. The method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 1, characterized in that: The distributed phase-shifting capacity optimization correction model established in step 6 is: Where: n1 is the number of distributed phase-shifting nodes that meet the short-circuit ratio requirements of new energy stations in one stage, K i is the number of distributed phase regulators configured after the capacity correction of wind farm i, k i Configure the number of distributed phase regulators for wind farm i that meets the short-circuit ratio requirements of renewable energy stations in a phase, S i is the capacity of the distributed phase-shifting machine, and γ is the penalty coefficient.

9. A method for site selection and capacity optimization configuration of a distributed phase regulator according to claim 8, characterized in that: In step 6, the process of optimizing the modified model by dynamically adaptive simulated annealing genetic algorithm and taking the transient voltage fluctuation of the rectifier-side bus as the voltage safety constraint to generate the optimal configuration scheme of the distributed phase-converter capacity is as follows: 1): A fixed-length symbol encoding method is used for chromosome encoding. Each gene in the chromosome corresponds to the number of distributed phase shifters in the station. The two states of "1" and "0" and their superposition states are used to represent the number of distributed phase shifters initially configured in each wind farm; 2): Define the initial population as: W=[p1,p2,...p n ] (21) Where: p1, p2, ... p n Chromosome encoding of the number of distributed phase regulators initially configured for n wind farm nodes in a phase; Initialize a population W of size N i , where the number of distributed phase regulators represented by the population chromosome gene cannot be less than the number of phase regulators configured in one stage; 3): Initial population W i Substitute into formula (21) to calculate, according to the fitness function and constraints, detect the performance of individuals in the population, reproduce the individuals with good performance to produce offspring, and eliminate the individuals with poor performance; 4): The initial population is subjected to genetic algorithm selection, crossover and mutation process to obtain subgroup Q i , merge the parent population with the child population, and calculate the individual objective function value of the new population; 5): Introduce the simulated annealing algorithm, use the Metropolis criterion, and compare the fitness function values ​​J2(x2) and J2(x1) of the two generations of individuals; If J2(x2)<J2(x1), accept the new generation; otherwise, according to the probability Accepting a new generation of offspring; 6): Iterate step by step until the number of iterations is met, and output the optimal configuration capacity of the camera for the specified node.

10. A method for site selection and capacity optimization configuration of a distributed phase condenser according to claim 9, characterized in that: The voltage safety constraint in step 6 is: i=1,2,3....n (24) 0≤k i ≤k max (25) V min (t)≤V(t)≤V max (t) (26) Where: V min (t) is the lower limit of transient voltage fluctuation, V max (t) is the upper limit of transient voltage fluctuation.

Images