Reactive voltage analysis method and system

Abstract

The invention belongs to the technical field of power system monitoring and management, and particularly relates to a reactive voltage analysis method and system.The multi-source heterogeneous data preprocessing module can safely collect and preprocess multi-source heterogeneous power grid data, and the efficiency and precision of follow-up reactive voltage optimization analysis are improved; the reactive voltage analysis module determines an objective function and constraint conditions of the power grid, and performs solving based on an inertial factor weight and learning factor improved particle swarm optimization, so that the solving efficiency and probability can be improved; finally, panoramic display, equipment tree display and analysis result display of data and analysis results can be realized at a display end, related personnel are helped to improve reactive voltage application scenes and analysis ranges, and safe and stable operation of power grid voltage is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system monitoring and management, and more specifically, relates to a reactive voltage analysis method and system. Background Art

[0002] In the new energy power system, reactive voltage can maintain the stable operation of the power system and optimize the power quality, and is an important part of the power system. However, at present, a large number of inductive loads are connected to the power grid, which affects the reactive voltage operation of the power grid. The centralized access of new energy power sources has caused a high penetration rate in some areas. The large-scale random fluctuations in the power of new energy power stations have caused high local voltage fluctuations, approaching the edge of unqualified.

[0003] In addition, the current coverage rate of urban power grid cables is gradually increasing. Compared with overhead lines, cable lines generate more reactive power. The inadequate compensation of inductive reactive power in the power grid is becoming increasingly prominent, and the phenomenon of high operating voltage occurs from time to time. Although there is a reactive voltage display system that can view the reactive operation of the power grid, the power grid data is multi-source and heterogeneous, and it has not been able to achieve a complete and effective reactive voltage data analysis function for the entire main distribution network. The efficiency and accuracy of the comprehensive analysis of reactive operation conditions still need to be improved, and there are deficiencies in the display, which cannot objectively and comprehensively evaluate the reactive voltage operation of the entire network, which is not conducive to the control and adjustment of the power grid voltage under normal and accident modes, and leaves hidden dangers for the safe and stable operation of the reactive voltage of the power grid. Summary of the invention

[0004] The technical problem to be solved by the present invention is: In order to overcome the above technical problems, the present invention provides a reactive voltage analysis method and system.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a reactive voltage analysis method, comprising the following steps:

[0006] Collect multi-source heterogeneous power grid data and perform preprocessing: establish scheduled data collection tasks, set data collection scope and frequency, and extract power grid model data and power grid operation data from the power grid; preprocessing includes outlier and normalization processing of the collected data;

[0007] Analyze and optimize the reactive power operation of the power grid: determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, the potential constraints of the objective function include the active and reactive power balance constraint, the node voltage constraint and the node power constraint; after determining the objective function and constraints, use the particle swarm algorithm improved based on the inertia factor weight and the learning factor to solve.

[0008] The power grid model data and power grid operation data include main transformer data, line data, bus data, distribution transformer data, voltage data, active power data and reactive power data.

[0009] The smart grid in the power grid generates a timestamp for the current period. When data from the power grid is collected, the timestamp is checked first. If it is correct, the data is accepted, and if it is wrong, the data is rejected.

[0010] In the preprocessing step, the Z-score method is used to process the outliers of the collected data. For each group of data samples x, it is converted into the corresponding data z according to the following formula:

[0011]

[0012] Where x is the sample, μ is the mean of the x sample, and σ is the standard deviation of the x sample;

[0013] For normalization, the data is normalized according to the max-min method, that is:

[0014]

[0015] Among them, x o A single data representing a data sample x, x max and x min Respectively represent the minimum and maximum values ​​in sample x, Represents the standardized data, with a value range of [0,1].

[0016] When analyzing and optimizing the reactive power operation of the power grid, the objective function of the reactive power operation is first determined. The objective function includes a network loss target and a voltage deviation target. The network loss target refers to minimizing the effect of distribution network loss. The objective function is as follows:

[0017]

[0018] Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j;

[0019] The voltage deviation objective function is as follows:

[0020]

[0021] Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; Ub is the reference voltage;

[0022] The overall reactive power objective function formula is:

[0023]

[0024] Among them, w 1 and w 2 They represent the weights of the two objectives, which are set according to the actual operation status of the network and the control requirements;

[0025] Active power balance and reactive power balance are shown in the following equations:

[0026]

[0027] Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ij is the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j;

[0028] The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula:

[0029]

[0030] Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits specified for node voltages, respectively;

[0031] Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula:

[0032]

[0033] Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively;

[0034] After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

[0035] In the particle swarm algorithm based on inertia factor weight and learning factor improvement, random inertia weight is used for improvement, that is, a set of random numbers is generated in each cycle, and then the problem of particles falling into local optimality is solved through uncertainty. If the algorithm is close to the optimal value at the beginning of the cycle, a smaller random weight value may be generated to accelerate convergence. If the algorithm cannot find the optimal value at the end of the cycle, a larger random weight value may be generated. The improved inertia factor weight can expand the search range of particles and solve the problem that the existing algorithm cannot find the optimal value. The improved inertia decreasing weight formula is set as follows:

[0036]

[0037] Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively;

[0038] The learning factor includes the self-learning factor c 1 and social learning factor c 2 , the improved learning factor formula is as follows:

[0039]

[0040] Among them, c 1,s and c 2,s The learning factor c 1 and c 2 The upper limit of c 1,e and c 2,e The learning factor c 1 and c 2 The lower limit of t max Refers to the maximum number of iterations set, and t is the current number of iterations.

[0041] It also includes the step of panoramic display: data processing and analysis are performed from the corresponding server based on user needs, and the results are output to the display end for panoramic display, device tree display and analysis result display.

[0042] A reactive voltage analysis system, comprising:

[0043] Multi-source heterogeneous data preprocessing module: used to collect multi-source heterogeneous power grid data and perform preprocessing; the multi-source heterogeneous data preprocessing module includes a data acquisition module, which establishes a scheduled data acquisition task, sets the data acquisition range and data acquisition frequency, and extracts power grid model data and power grid operation data from the power grid; the preprocessing includes outlier and normalization processing of the collected data;

[0044] Reactive voltage analysis module: used to analyze and optimize the reactive operation of the distribution network, determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, the potential constraints of the objective function include active and reactive power balance constraints, node voltage constraints and node power constraints; after determining the objective function and constraints, use the particle swarm algorithm based on inertia factor weight and learning factor improvement to solve.

[0045] In the reactive voltage analysis module, the objective function includes the network loss target and the voltage deviation target; the network loss target refers to minimizing the effect of distribution network loss. The objective function is as follows:

[0046]

[0047] Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j;

[0048] The voltage deviation objective function is as follows:

[0049]

[0050] Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; U b is the reference voltage;

[0051] The overall reactive power objective function formula is:

[0052]

[0053] Among them, w 1 and w 2 They represent the weights of the two objectives, which are set according to the actual operation status of the network and the control requirements;

[0054] Active power balance and reactive power balance are shown in the following equations:

[0055]

[0056] Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ij is the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j;

[0057] The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula:

[0058]

[0059] Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits specified for node voltages, respectively;

[0060] Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula:

[0061]

[0062] Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively;

[0063] After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

[0064] In the particle swarm algorithm based on inertia factor weight and learning factor improvement, random inertia weight is used for improvement, that is, a set of random numbers is generated in each cycle, and then the problem of particles falling into local optimality is solved through uncertainty. If the algorithm is close to the optimal value at the beginning of the cycle, a smaller random weight value may be generated to accelerate convergence. If the algorithm cannot find the optimal value at the end of the cycle, a larger random weight value may be generated. The improved inertia factor weight can expand the search range of particles and solve the problem that the existing algorithm cannot find the optimal value. The improved inertia decreasing weight formula is set as follows:

[0065]

[0066] Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively;

[0067] The learning factor includes the self-learning factor c 1 and social learning factor c 2 , the improved learning factor formula is as follows:

[0068]

[0069] Among them, c 1,s and c 2,s The learning factor c 1 and c 2 The upper limit of c 1,e and c 2,e The learning factor c 1 and c 2 The lower limit of t max Refers to the maximum number of iterations set, and t is the current number of iterations.

[0070] It also includes a panoramic display module, which is used to realize panoramic display, device tree display and analysis result display of data and analysis results on the display end.

[0071] The beneficial effects of the present invention are as follows: a reactive voltage analysis method and system of the present invention can safely collect multi-source heterogeneous power grid data and perform preprocessing, thereby improving the efficiency and accuracy of subsequent reactive voltage optimization analysis; the reactive voltage analysis module determines the objective function and constraint conditions of the power grid, and solves the problem based on the particle swarm algorithm improved by the inertia factor weight and the learning factor, thereby improving the solution efficiency and probability; the inertia factor weight of the present invention can expand the search range of particles and solve the problem that the existing algorithm cannot find the optimal value; the learning factor method of the present invention can enhance the exploration surface and flight speed of the particle swarm during the movement process, shorten the algorithm convergence time, and overcome the problem of slow algorithm convergence speed; the data and analysis results can be displayed on the display end in a panoramic manner, in a device tree, and in an analysis result display, thereby helping relevant personnel to improve the application scenarios and analysis range of reactive voltage and ensure the safe and stable operation of the power grid voltage. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0073] Figure 1 It is an implementation flow chart of the reactive voltage analysis method of the present invention. DETAILED DESCRIPTION

[0074] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0075] like Figure 1 As shown, a reactive voltage analysis method of the present invention comprises the following steps:

[0076] Collect multi-source heterogeneous power grid data and perform preprocessing: establish scheduled data collection tasks, set data collection range and frequency, extract power grid model data and power grid operation data from the power grid; preprocessing includes outlier and normalization processing of the collected data; due to the diversity of sources and data types, data preprocessing can improve the efficiency and accuracy of subsequent reactive voltage optimization analysis.

[0077] Analyze and optimize the reactive power operation of the power grid: determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, the potential constraints of the objective function include the active and reactive power balance constraint, the node voltage constraint and the node power constraint; after determining the objective function and constraints, use the particle swarm algorithm improved based on the inertia factor weight and the learning factor to solve.

[0078] The power grid includes a D5000 system and / or a distribution automation system. The power grid model data and power grid operation data include main transformer data, line data, bus data, distribution transformer data, voltage data, active data and reactive data.

[0079] During the collection process, in order to prevent the collection of malicious and redundant data, the smart grid in the power grid will generate a timestamp for the current period. When the data in the power grid is collected, the timestamp is checked first. If it is correct, the data is accepted; if it is wrong, the data is rejected.

[0080] In the preprocessing step, the Z-score method is used to process the collected data for outliers to reduce the generation of outliers. For each group of data samples x, it is converted into the corresponding data z according to the following formula:

[0081]

[0082] Where x is the sample, μ is the mean of the x sample, and σ is the standard deviation of the x sample. The Z-score method can eliminate the inconvenience caused by the magnitude of power grid data in data analysis, while ensuring the comparability and effectiveness of power grid data.

[0083] For normalization, the data is normalized according to the max-min method, that is:

[0084]

[0085] Among them, x o A single data representing a data sample x, x max and x min Respectively represent the minimum and maximum values ​​in sample x, Represents the standardized data, with a value range of [0,1]. By using the max-min method, the data value range can be limited to improve the accuracy of subsequent data analysis.

[0086] When analyzing and optimizing the reactive power operation of the power grid, the objective function of the reactive power operation is first determined. The objective function includes a network loss target and a voltage deviation target. The network loss target refers to minimizing the effect of distribution network loss. The objective function is as follows:

[0087]

[0088] Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j;

[0089] The reduction of voltage deviation can help improve the voltage of distribution network nodes and avoid voltage exceeding the limit. Therefore, the voltage deviation objective function is as follows:

[0090]

[0091] Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; U b is the reference voltage;

[0092] The overall reactive power objective function formula is:

[0093]

[0094] Among them, w 1 and w 2 Represent the weights of the two objectives respectively, which are set according to the actual operation status of the network and the control requirements. In actual use, w 1 and w 2 Each is set to 0.5;

[0095] At the same time, there are some potential constraints in the above objective function, including active and reactive power balance constraints, node voltage constraints and node power constraints; the active and reactive power balance constraint means that the sum of the power generated by the power source, the power consumed by the load and the power lost in the network loss is 0, which is divided into active power balance and reactive power balance, as shown in the following formula:

[0096]

[0097] Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ij is the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j;

[0098] The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula:

[0099]

[0100] Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits of node voltage respectively; in actual use, U i,min =0.95Un,U i,max =1.05Un, where Un refers to the standard voltage. The standard range of 0.95 and 1.05 is taken into account that in real life, the voltage sometimes fluctuates up and down, but rarely exceeds 0.05.

[0101] Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula:

[0102]

[0103] Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively;

[0104] After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

[0105] The inertia factor ω determines the changing trend of the particle speed and affects the speed and accuracy of the algorithm. The larger the ω is set, the farther the particles fly, the larger the search space is, the easier it is to find the global optimum, and the higher the global search ability is. On the contrary, if ω is set smaller, the particles will not be able to remember the historical speed, and the particles tend to fly shorter, find the local optimum, and have better local optimization ability. Therefore, the value of the inertia weight has a great influence on the particle swarm algorithm. In the process of solving the reactive function target of the distribution network, due to the complexity of the power grid data from multiple sources and the complex relationship between the objective function and the constraint variables, it is also necessary to maintain a balance between the diversity of the solution set and the convergence. Therefore, in the particle swarm algorithm based on the inertia factor weight and the learning factor improvement of the present invention, a random inertia weight is used for improvement, that is, a set of random numbers is generated during each cycle, and then the problem that the particles are prone to fall into the local optimum is solved through uncertainty. If the algorithm is close to the optimal value at the beginning of the cycle, a smaller random weight value may be generated to accelerate convergence. If the algorithm cannot find the optimal value at the end of the cycle, a larger random weight value may be generated. The improved inertia factor weight can expand the search range of the particles and solve the problem that the existing algorithm cannot find the optimal value. The improved inertia decreasing weight formula is set as follows:

[0106]

[0107] Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively. In actual use, μ max =1,μ min =0.4, μ=0.5.

[0108] In addition, in the particle swarm algorithm, the different learning factors have a great impact on the solution efficiency. The learning factor includes the self-learning factor c 1 and social learning factor c 2 , the former controls the individual's self-identification ability, and the latter controls the individual's social cognition ability. In order to improve the efficiency and accuracy of the solution, the present invention uses an improved learning factor, which can enhance the exploration surface and flight speed of the particle swarm during the movement process, shorten the algorithm convergence time, and overcome the problem of slow algorithm convergence speed. The formula is as follows:

[0109]

[0110] Among them, c1,s and c 2,s The learning factor c 1 and c 2 The upper limit of c 1,e and c 2,e The learning factor c 1 and c 2 The lower limit of t max Refers to the maximum number of iterations set, and t is the current number of iterations. In actual use, you can set c 1,s =2.5, c 2,s =0.5, c 1,e =0.5, c 2,e =2.5.

[0111] It also includes the steps of panoramic display: data processing and analysis are performed from the corresponding server based on user needs, and the results are output to the display end for panoramic display, device tree display and analysis result display, to help relevant personnel improve the application scenarios and analysis scope of reactive voltage, and ensure the safe and stable operation of the power grid voltage.

[0112] like Figure 1 As shown, a reactive voltage analysis system includes a multi-source heterogeneous data preprocessing module and a reactive voltage analysis module.

[0113] Multi-source heterogeneous data preprocessing module: used for collecting multi-source heterogeneous power grid data and performing preprocessing; the multi-source heterogeneous data preprocessing module includes a data acquisition module, which establishes a scheduled data acquisition task, sets the data acquisition range and data acquisition frequency, and extracts power grid model data and power grid operation data from the power grid; preprocessing includes performing outlier and normalization processing on the collected data; due to the diversity of sources and data types of the collected data, the multi-source heterogeneous data preprocessing module can safely collect multi-source heterogeneous power grid data and perform preprocessing, thereby improving the efficiency and accuracy of subsequent reactive voltage optimization analysis.

[0114] Reactive voltage analysis module: used to analyze and optimize the reactive operation of the distribution network, determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, and the potential constraints of the objective function include active and reactive power balance constraints, node voltage constraints and node power constraints; after the objective function and constraints are determined, the particle swarm algorithm based on the inertia factor weight and the learning factor is used to solve them. The reactive voltage analysis module determines the objective function and constraints of the power grid, and solves them based on the particle swarm algorithm based on the inertia factor weight and the learning factor, which can improve the efficiency and probability of the solution. The reactive voltage analysis module can comprehensively and objectively optimize the operation of the reactive power grid and help control and adjust the grid voltage.

[0115] In the reactive voltage analysis module, the objective function includes the network loss target and the voltage deviation target; the network loss target refers to minimizing the effect of distribution network loss. The objective function is as follows:

[0116]

[0117] Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j;

[0118] The reduction of voltage deviation can help improve the voltage of distribution network nodes and avoid voltage exceeding the limit. Therefore, the voltage deviation objective function is as follows:

[0119]

[0120] Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; U b is the reference voltage;

[0121] The overall reactive power objective function formula is:

[0122]

[0123] Among them, w 1 and w 2 Represent the weights of the two objectives respectively, which are set according to the actual operation status of the network and the control requirements. In actual use, w 1 and w 2 Each is set to 0.5;

[0124] At the same time, there are some potential constraints in the above objective function, including active and reactive power balance constraints, node voltage constraints and node power constraints; the active and reactive power balance constraint means that the sum of the power generated by the power source, the power consumed by the load and the power lost in the network loss is 0, which is divided into active power balance and reactive power balance, as shown in the following formula:

[0125]

[0126] Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ijis the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j;

[0127] The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula:

[0128]

[0129] Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits of node voltage respectively; in actual use, U i,min =0.95Un,U i,max =1.05Un, where Un refers to the standard voltage. The standard range of 0.95 and 1.05 is taken into account that in real life, the voltage sometimes fluctuates up and down, but rarely exceeds 0.05.

[0130] Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula:

[0131]

[0132] Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively;

[0133] After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

[0134] The inertia factor ω determines the changing trend of the particle speed and affects the speed and accuracy of the algorithm. The larger the ω is set, the farther the particles fly, the larger the search space is, the easier it is to find the global optimum, and the higher the global search ability is. On the contrary, if ω is set smaller, the particles will not be able to remember the historical speed, and the particles tend to fly shorter, find the local optimum, and have better local optimization ability. Therefore, the value of the inertia weight has a great influence on the particle swarm algorithm. In the process of solving the reactive function target of the distribution network, due to the complexity of the power grid data from multiple sources and the complex relationship between the objective function and the constraint variables, it is also necessary to maintain a balance between the diversity of the solution set and the convergence. Therefore, in the particle swarm algorithm based on the inertia factor weight and the learning factor improvement of the present invention, a random inertia weight is used for improvement, that is, a set of random numbers is generated during each cycle, and then the problem that the particles are prone to fall into the local optimum is solved through uncertainty. If the algorithm is close to the optimal value at the beginning of the cycle, a smaller random weight value may be generated to accelerate convergence. If the algorithm cannot find the optimal value at the end of the cycle, a larger random weight value may be generated. The improved inertia factor weight can expand the search range of the particles and solve the problem that the existing algorithm cannot find the optimal value. The improved inertia decreasing weight formula is set as follows:

[0135]

[0136] Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively. In actual use, μ max =1,μ min =0.4, μ=0.5.

[0137] In addition, in the particle swarm algorithm, the different learning factors have a great impact on the solution efficiency. The learning factor includes the self-learning factor c 1 and social learning factor c 2 , the former controls the individual's self-identification ability, and the latter controls the individual's social cognition ability. In order to improve the efficiency and accuracy of the solution, the present invention uses an improved learning factor, which can enhance the exploration surface and flight speed of the particle swarm during the movement process, shorten the algorithm convergence time, and overcome the problem of slow algorithm convergence speed. The formula is as follows:

[0138]

[0139] Among them, c 1,s and c 2,s The learning factor c 1 and c 2 The upper limit of c 1,eand c 2,e The learning factor c 1 and c 2 The lower limit of t max Refers to the maximum number of iterations set, and t is the current number of iterations. In actual use, you can set c 1,s =2.5, c 2,s =0.5, c 1,e =0.5, c 2,e =2.5.

[0140] Finally, the improved inertia factor and learning factor are brought into the particle swarm algorithm for solution.

[0141] Compared with the original particle swarm algorithm and the chaotic particle swarm algorithm, the algorithm of the present invention can search for a very small spatial solution in the early stage of the iteration cycle. At the same time, after 200 iterations, the algorithm of the present invention can search for the optimal value when it first starts to converge, and there is no situation of falling into local convergence. In general, the algorithm of the present invention has good performance in terms of accuracy, stability and convergence speed.

[0142] It also includes a panoramic display module, which is used to realize panoramic display, device tree display and analysis result display of data and analysis results on the display end, helping relevant personnel to improve the application scenarios and analysis scope of reactive voltage and ensure the safe and stable operation of power grid voltage.

[0143] This module allows users to intuitively see data such as reactive voltage analysis, helping to improve the application scenarios and analysis scope of reactive voltage, and ensuring the safe and stable operation of power grid voltage. This module can accept user query requirements, perform data processing and analysis from the corresponding server based on the requirements, and output the results to the display end for display. If the user queries the reactive voltage of the main grid, the system will automatically analyze the distribution transformer voltage limit according to the distribution transformer voltage limit, and expand the bus voltage and reactive compensation of the associated superior substation, conduct analysis and judgment, and transmit the results to the display end for display; if the user queries the distribution voltage, the system will automatically analyze the distribution transformer voltage limit according to the distribution transformer voltage limit, and expand the bus voltage and reactive compensation of the associated superior substation, conduct analysis and judgment, and transmit the results to the display end for display; if the user queries the power grid section, the power grid section analysis module, query and display data processing module and data acquisition module in the server work together to save the power grid reactive operation data section according to the time interval, and compare the data of different time sections. The comparison results are exported and transmitted to the display terminal for display; if the user queries the main distribution network equipment model information, the system will intuitively display the equipment-related information in the form of equipment tree on the display terminal; if the user queries the voltage panoramic display of a certain area, the system will obtain data from the corresponding database, and the display terminal will display the reactive power flow partitions and layers based on the map of the area according to different voltage levels, different regions, and partitions. You can also browse the real-time operation status of the node equipment by clicking on each voltage level node; if the user queries historical data, the system will obtain data from the corresponding database, and the display terminal will display historical reactive operation data, including active power curve, reactive power curve, power factor curve, voltage curve, etc. The panoramic display module comprehensively displays the query results according to the query requirements, provides auxiliary analysis and decision-making for grid voltage adjustment and reactive power balance, and helps relevant personnel improve the application scenarios and analysis scope of reactive voltage, and ensure the safe and stable operation of grid voltage.

[0144] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A reactive voltage analysis method, characterized in that: The following steps are involved: Collect multi-source heterogeneous power grid data and perform preprocessing: establish scheduled data collection tasks, set data collection scope and frequency, and extract power grid model data and power grid operation data from the power grid; preprocessing includes outlier and normalization processing of the collected data; Analyze and optimize the reactive power operation of the power grid: determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, and the potential constraints of the objective function include the active and reactive power balance constraint, the node voltage constraint and the node power constraint; after determining the objective function and constraints, use the particle swarm algorithm improved based on the inertia factor weight and the learning factor to solve.

2. The reactive voltage analysis method according to claim 1, characterized in that: The power grid model data and power grid operation data include main transformer data, line data, bus data, distribution transformer data, voltage data, active power data and reactive power data.

3. The reactive voltage analysis method according to claim 1, characterized in that: The smart grid in the power grid generates a timestamp for the current period. When data from the power grid is collected, the timestamp is checked first. If it is correct, the data is accepted, and if it is wrong, the data is rejected.

4. The reactive voltage analysis method according to claim 1, characterized in that: In the preprocessing step, the Z-score method is used to process the outliers of the collected data. For each group of data samples x, it is converted into the corresponding data z according to the following formula: Where x is the sample, μ is the mean of the x sample, and σ is the standard deviation of the x sample; For normalization, the data is normalized according to the max-min method, that is: Among them, x o A single data representing a data sample x, x max and x min Respectively represent the minimum and maximum values ​​in sample x, Represents the standardized data, with a value range of [0,1].

5. The reactive voltage analysis method according to claim 1, characterized in that: When analyzing and optimizing the reactive power operation of the power grid, the objective function of the reactive power operation is first determined. The objective function includes the network loss target and the voltage deviation target. The network loss target is to minimize the effect of distribution network loss. The network loss objective function is: Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j; The voltage deviation objective function is: Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; U b is the reference voltage; The overall reactive power objective function formula is: Among them, w1 and w2 represent the weights of the two objectives, which are set according to the actual operation status of the network and the control requirements; Active power balance and reactive power balance are shown in the following equations: Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ij is the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j; The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula: Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits specified for node voltages, respectively; Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula: Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively; After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

6. The reactive voltage analysis method according to claim 1, characterized in that: In the particle swarm algorithm based on the improvement of inertia factor weight and learning factor, random inertia weight is used for improvement, and the improved inertia decreasing weight formula is set as follows: Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively; The learning factor includes the self-learning factor c1 and the social learning factor c2. The improved learning factor formula is as follows: Among them, c 1,s and c 2,s Refers to the upper limit of the learning factors c1 and c2, c 1,e and c 2,e refers to the lower limit of the learning factors c1 and c2, t max Refers to the maximum number of iterations set, and t is the current number of iterations.

7. A reactive voltage analysis system, characterized in that: Including multi-source heterogeneous data preprocessing module and reactive voltage analysis module; Multi-source heterogeneous data preprocessing module: used to collect multi-source heterogeneous power grid data and perform preprocessing; the multi-source heterogeneous data preprocessing module includes a data acquisition module, which establishes a scheduled data acquisition task, sets the data acquisition range and data acquisition frequency, and extracts power grid model data and power grid operation data from the power grid; the preprocessing includes outlier and normalization processing of the collected data; Reactive voltage analysis module: used to analyze and optimize the reactive operation of the distribution network, determine the objective function and constraints of the power grid, the objective function includes the network loss target and the voltage deviation target, the potential constraints of the objective function include active and reactive power balance constraints, node voltage constraints and node power constraints; after determining the objective function and constraints, use the particle swarm algorithm based on inertia factor weight and learning factor improvement to solve.

8. The reactive voltage analysis system according to claim 7, characterized in that: In the reactive voltage analysis module, the objective function includes the network loss target and the voltage deviation target; the network loss target refers to minimizing the effect of distribution network loss, and the objective function is: Among them, P loss is the total active power loss of the network, i and j represent the node numbers of the distribution network, N represents the set of distribution network nodes, v(i) represents the node connected to node i, z ij is the line impedance between nodes i and j, I ij is the current flowing between nodes i and j; The voltage deviation objective function is: Among them, U bm is the total voltage deviation of the distribution network, U i is the voltage of node i in the distribution network; U b is the reference voltage; The overall reactive power objective function formula is: Among them, w1 and w2 represent the weights of the two objectives, which are set according to the actual operation status of the network and the control requirements; Active power balance and reactive power balance are shown in the following equations: Among them, P j and Q j The active power and reactive power injected into node j are P ij and Q ij are the active power and reactive power flowing from node i to node j, R ij and X ij is the resistance and reactance between nodes i and j, U j is the voltage of node j, v(j) represents the node connected to node j; The node voltage constraint means that the voltage of each node should be within the specified range, as shown in the following formula: Among them U i is the voltage at node i, U i,min , U i,max They represent the minimum and maximum limits specified for node voltages, respectively; Node power constraint means that the power absorbed by each node should be within the specified range, which is divided into active power constraint and reactive power constraint, as shown in the following formula: Among them, P i , P i,min , P i,max are the active power absorbed by node i and its minimum and maximum amplitudes, Q i , Q i,min , Q i,max are the reactive power absorbed by node i and its minimum and maximum magnitudes respectively; After the reactive power objective function and constraints are determined, the particle swarm algorithm based on inertia factor weight and learning factor is used to solve the problem.

9. The reactive voltage analysis system according to claim 7, characterized in that: In the particle swarm algorithm based on the improvement of inertia factor weight and learning factor, random inertia weight is used for improvement, and the improved inertia decreasing weight formula is set as follows: Among them, n(0,1) represents a random number with a normal distribution, rand(0,1) represents a random number ranging from 0 to 1, σ is the variance of the random weighted average, μ max and μ min are the maximum and minimum values ​​of the random weight averages, respectively; The learning factor includes the self-learning factor c1 and the social learning factor c2. The improved learning factor formula is as follows: Among them, c 1,s and c 2,s refers to the upper limit of the learning factors c1 and c2, c 1,e and c 2,e refers to the lower limit of the learning factors c1 and c2, t max Refers to the maximum number of iterations set, and t is the current number of iterations.

10. The reactive voltage analysis system according to claim 7, characterized in that: It also includes a panoramic display module, which is used to realize panoramic display, device tree display and analysis result display of data and analysis results on the display end.

Images