Voltage stability domain analysis method based on network information active sampling

Abstract

The invention relates to the field of direct current network analysis, in particular to a voltage stability domain analysis method based on network information active sampling. The method comprises the following steps: acquiring basic network data of a direct current network, constructing a network power flow model, and analyzing the network power flow model to obtain a voltage stability domain; analyzing the voltage stability domain according to the network power flow model to obtain a stability set; analyzing the stable set to obtain candidate samples; and in combination with the candidate samples, performing numerical simulation test on the direct current network to obtain a test result. According to the invention, the robustness of the network can be improved.

Description

Technical Field

[0001] The present invention relates to the field of DC network analysis, and particularly to a method for analyzing voltage stability region based on active sampling of network information. Background Art

[0002] The stable voltage source rectifier region (SVSR) refers to a concept for the stability analysis and control of voltage source rectifiers in a DC power system (such as a DC transmission system); in a DC power system, the SVSR is mainly used to ensure that power electronic devices (such as rectifiers and inverters) can maintain the stable operation of the system under various operating conditions; especially for high-voltage DC (HVDC) transmission systems, which usually rely on voltage source rectifiers (VSRs) to convert alternating current (AC) into direct current (DC) and ensure the stability of the DC power grid; the SVSR mainly focuses on how to ensure that the voltage source rectifier can still work stably under various dynamics and disturbances in these systems.

[0003] The safety-constrained economic dispatch method for a safety region with the publication number of CN102983573A discloses a safety-constrained economic dispatch method based on a safety region. The method includes: The present invention relates to a power system. To provide an effective way to handle the network security constraints of the system, including branch power flow, static voltage stability and transient stability constraints, and to coordinate the contradiction between economy and security in the traditional optimal operation problem of power system economic dispatch, the technical solution adopted by the present invention is a safety-constrained economic dispatch method based on a safety region, including the following steps: The first step: Calculate the boundary coefficients of the active static safety region, the cut-set voltage stability region and the dynamic safety region of the system respectively; The second step: Establish a safety-constrained economic dispatch model based on the safety region; The third step: Use a cognitive-based social evolution algorithm to solve the unit start-stop state optimization sub-problem; The fourth step: Calculate the generation cost, static voltage stability margin and transient stability margin of each unit within the dispatch period; The fifth step: Obtain a feasible economic dispatch plan; otherwise, return to the third step. The present invention is mainly applied to the optimization of power load allocation.

[0004] In the prior art, the intrinsic information of the sample space and the cyber-physical characteristics have not been effectively utilized, resulting in a significant increase in computational resource consumption when the data volume is large or the sample distribution is complex. In addition, when dealing with large-scale networks, numerical methods face computational bottlenecks and are difficult to efficiently handle high-complexity problems. The lack of flexibility in characterization limits the adaptability and flexibility of analysis methods when facing complex and changing actual systems, making it difficult to adapt to the dynamic changes brought about by the high penetration rate of renewable energy, resulting in insufficient characterization accuracy. Strong dependence on the scale of training data fails to fully utilize the geometric features of the sample space and the useful information in the iterative process, leading to high data preparation and computational costs. In particular, when dealing with large-scale systems, the efficiency and accuracy are limited. The geometric structure information of the sample space is not fully utilized, increasing the computational cost and reducing the efficiency. The inability to perform precise sampling and training according to the characteristics of the sample space is a problem that needs to be solved. Summary of the Invention

[0005] The object of the present invention is to propose a voltage stability region analysis method based on active sampling of network information in view of the problems existing in the background technology.

[0006] The technical solution of the present invention: A voltage stability region analysis method based on active sampling of network information includes the following steps:

[0007] S1. Obtain the basic network data of the DC network, construct a network power flow model, analyze the network power flow model, and obtain the voltage stability region;

[0008] S2. Analyze the voltage stability region according to the network power flow model to obtain a stable set;

[0009] S3. Analyze the stable set to obtain candidate samples;

[0010] S4. Combine the candidate samples and conduct numerical simulation tests on the DC network to obtain test results.

[0011] Preferably, the process of obtaining the basic network data of the DC network and constructing a network power flow model includes:

[0012] The basic network information includes network structure information and network relationship information; the network structure information includes voltage source information and load information; the network relationship information includes the relationship between voltage sources and loads and the relationship between loads and loads;

[0013] Connect the voltage sources and loads through the relationship between voltage sources and loads and the relationship between loads and loads to construct a DC network model.

[0014] Preferably, the process of analyzing the network power flow model to obtain the voltage stability region includes:

[0015] Obtain the operating points of the DC network model, denoted as node 1, node 2, node 3, ……, node i, ……, node j, ……, node u; when there is a power line between node i and node j, denote it as ; when there is no power line between node i and node j, denote it as ; Denote the conductance of the power line between node and node as , being a positive real number, then the Kirchhoff matrix of the DC network model is:

[0016] ;

[0017] By judging whether the node is a load or a voltage source, divide the Kirchhoff matrix and denote it as ; where, is the number of loads; is the number of voltage sources; it is stipulated that the matrix and are positive definite;

[0018] Obtain the load power, load voltage, external power supply voltage, and controllable voltage set point. Then the vectors of the load power, load voltage, external power supply voltage, and controllable voltage set point are , , , and , where ;

[0019] Denote the voltage potential vector at the node as , and stipulate that ; Denote the current injected into the network at the node as ; According to the voltage potential vector at the node and the current injected into the network at the node, obtain the power vector p provided by the node to the power grid;

[0020] ;

[0021] Through Kirchhoff's and Ohm's laws, express as:

[0022] ;

[0023] According to formula (3) and formula (4), obtain the power vector;

[0024] ;

[0025] Denote the steady-state operating point of the system as , denote the load bus voltage at the operating point as , denote the power vector input by the power node to the system as ; According to the system steady-state operating point, the load bus voltage at the operating point, the power vector input by the power node to the system, and Equation (5), perform parameter substitution on each term on the right side of Equation (5) to obtain the static power flow equation, Equation (6); Input the static power flow equation into the DC network model, construct the network power flow model, and analyze the static power flow equation;

[0026] ;

[0027] Wherein, , and are the conductance matrices of the power-source, power-load, and load-load sub-networks in the network power flow model respectively; = , and are sub-matrices of the admittance matrix; For the state variable is quadratic, and for the variable and the state variable is bilinear.

[0028] Preferably, the process of analyzing the static power flow equation is as follows:

[0029] Obtain the solution of the static power flow equation, denoted as the static operating point. If the static power flow equation has no solution, it means that the system static voltage is unstable;

[0030] Specify the voltage stability region as:

[0031] ;

[0032] If for any arbitrarily given controllable voltage setpoint , there exists , satisfying all of the load injection power, and Equation (6) has a solution, then the power set is denoted as the voltage stability region.

[0033] Preferably, the process of analyzing the voltage stability region according to the network power flow model to obtain the stable set includes:

[0034] Specify that under the conditions given by the system, there exists a high load value , and this load value satisfies the voltage stability region condition of the DC network , then when , it means that the voltage stability region condition of the DC network is satisfied;

[0035] When When the static power flow equation has a solution, all have solutions;

[0036] Through the characteristics of the voltage stability region, which refer to monotonicity and convex set characteristics, a stable sub-region is constructed ;

[0037] Suppose there are m elements in the voltage stability region ) , which form a stable set P;

[0038] .

[0039] Preferably, the process of analyzing the stable set to obtain candidate samples includes:

[0040] Perform closed-loop training on the stable set of the DC network. After actively sampling the stable set, sample labels and a data set are obtained. The closed-loop training includes classifier training, sampling, screening, labeling, and SVSR subset identification; the classifier training refers to training the classifier using an extended training set that includes newly selected samples and existing labeled samples; the sampling refers to randomly selecting a batch of samples in the unlabeled sample space; the screening refers to evaluating the posterior probability of the selected samples to determine their uncertainty and selecting the samples with the highest uncertainty as the new candidate label set; the labeling refers to annotating the screened candidate samples through network numerical simulation; the SVSR subset identification refers to identifying the SVSR subset that supports the labeling process using the physical characteristics of the network;

[0041] Set the feasibility posterior probability , which is used to measure the probability of the power flow having a solution; analyze the probability of the static power flow equation having a solution through the feasibility posterior probability.

[0042] Preferably, the process of analyzing the probability of the static power flow equation having a solution through the feasibility posterior probability includes:

[0043] When the value of is close to 1 or 0, it means that the sample is solvable (i.e., belongs to the target category) or unsolvable (i.e., does not belong to the target category); when the value of is close to 0.5, it means that the classification result of the model for this sample is uncertain, and analyze the feasibility posterior probability . If is greater than 0.5, define the quantified uncertainty measure as:

[0044] ;

[0045] If Less than or equal to 0.5, the quantified uncertainty measure is defined as: ;

[0046] Set the training cycle, sort the samples according to the quantified uncertainty, and through the training cycle, mark the samples with the highest uncertainty as candidate samples; the initial sample quantity and the number of samples in each training cycle can be set.

[0047] Preferably, in combination with the candidate samples, the process of numerically simulating and testing the DC network to obtain the test results includes:

[0048] In the numerical simulation test, conduct tests on the IEEE 14-node, 39-node, and 118-node DC networks, perform numerical calculations through IPOPT, conduct simulations through Python and Julia, obtain candidate samples, record the test data, and obtain the test results.

[0049] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects: By constructing a network power flow model and a voltage stability region, analyzing the monotonicity of the DC network power flow equation and the convex set characteristics of the SVSR, an active sampling technique is proposed, which can accurately select the most informative samples, reduce the dependence on large-scale data, significantly reduce the consumption of computing resources and improve the computing efficiency through active sampling, and at the same time improve the accuracy of the SVSR characterization; Through the stable set, using the convex set characteristics of the SVSR and the monotonicity of the power flow equation, improve the accuracy and adaptability of the SVSR characterization, especially performing excellently in dealing with the dynamic changes in the context of high-penetration renewable energy; Through candidate samples, by quantifying the uncertainty and geometric information of the sample space, iteratively optimize the sample selection and labeling process; Through closed-loop screening, reduce the dependence on the scale of training data, improve the robustness and characterization flexibility of the model, and provide a mathematical theoretical guarantee based on the convexity of the SVSR and the monotonicity of the power flow equation, and construct a complete set of algorithms for characterizing the DC network stability region to ensure the high accuracy and wide applicability of the method. Description of the Drawings

[0050] Figure 1 It is a schematic diagram of an embodiment proposed by the present invention. Detailed Embodiments

[0051] Embodiment 1, as Figure 1 shown, a voltage stability region analysis method based on network information active sampling proposed by the present invention includes the following steps:

[0052] S1. Obtain the basic network data of the DC network, construct a network power flow model, analyze the network power flow model, and obtain the voltage stability region;

[0053] S2. Analyze the voltage stability region according to the network power flow model to obtain the stable set;

[0054] S3. Analyze the stable set to obtain candidate samples;

[0055] S4. Combine the candidate samples and conduct numerical simulation tests on the DC network to obtain the test results.

[0056] It should be further noted that in the specific implementation process, the process of obtaining the basic network data of the DC network, constructing the network power flow model, and analyzing the network power flow model to obtain the stability region is as follows:

[0057] The basic network information includes network structure information and network relationship information; the network structure information includes voltage source information and load information; the network relationship information includes the relationship between the voltage source and the load and the relationship between the loads;

[0058] It should be further noted that in the specific implementation process, the relationship between the voltage source and the load refers to the transmission relationship between the voltage source and the load in the DC network; the relationship between the loads refers to the transmission relationship between the loads in the DC network;

[0059] Connect the voltage source and the load through the relationship between the voltage source and the load and the relationship between the loads to construct a DC network model;

[0060] Obtain the operating points of the DC network model, denoted as node 1, node 2, node 3,..., node i,..., node j,..., node u; when there is a power line between node i and node j, it is denoted as ; when there is no power line between node i and node j, it is denoted as ; Denote the conductance of the power line between and node as , is a positive real number, then the Kirchhoff matrix of the DC network model is:

[0061] ;

[0062] By judging whether the node is a load or a voltage source, divide the Kirchhoff matrix and denote it as ; where is the number of loads; is the number of voltage sources; it is stipulated that the matrices and are positive definite;

[0063] Obtain the load power, load voltage, external power supply voltage, and controllable voltage set point. Then the vectors of load power, load voltage, external power supply voltage, and controllable voltage set point are respectively , , , and , where ;

[0064] Denote the voltage potential vector at the node as , and stipulate that ; Denote the current injected into the network at the node as ; According to the voltage potential vector at the node and the current injected into the network at the node, obtain the power vector p provided by the node to the power grid;

[0065] ;

[0066] Through Kirchhoff's and Ohm's laws, express as:

[0067] ;

[0068] According to formula (3) and formula (4), obtain the power vector;

[0069] ;

[0070] Denote the system steady-state operating point as , denote the load bus voltage at the operating point as , and denote the power vector input by the power source node to the system as ; According to the system steady-state operating point, the load bus voltage at the operating point, the power vector input by the power source node to the system, and equation (5), perform parameter substitution on each term on the right side of equation (5) to obtain the static power flow equation, equation (6); Input the static power flow equation into the DC network model to construct the network power flow model;

[0071] ;

[0072] Among them, , and are respectively the conductance matrices of the power source - power source, power source - load, and load - load sub-networks in the network power flow model; = , and are sub-matrices of the admittance matrix; For the state variable it is quadratic, and for the variable and the state variable it is bilinear;

[0073] Obtain the solution of the static power flow equation, denoted as the static operating point. If the static power flow equation has no solution, it indicates that the system's static voltage is unstable;

[0074] Specify the voltage stability region as:

[0075] ;

[0076] If any given controllable voltage setpoint , there exists , satisfying all of the load injection power, and equation (6) has a solution, then the power set , denoted as the voltage stability region;

[0077] It should be further noted that in the specific implementation process, if a certain specific load power , then the load power is solvable.

[0078] It should be further noted that in the specific implementation process, analyzing the voltage stability region according to the network power flow model, the process of obtaining the stable set is as follows:

[0079] Specify that under the given conditions of the system, there exists a high load value , and this load value satisfies the voltage stability region conditions of the DC network , then when , it indicates that the voltage stability region conditions of the DC network are satisfied;

[0080] When for the static power flow equation has a solution, all have solutions;

[0081] Through the characteristics of the voltage stability region, the characteristics of the voltage stability region refer to monotonicity and convex set characteristics, and construct a stable sub-region ;

[0082] Assume that m elements in the voltage stability region ) , constitute the stable set P;

[0083] .

[0084] It should be further noted that in the specific implementation process, the process of analyzing the stable set and obtaining candidate samples is as follows:

[0085] Perform closed-loop training on the stable set of the DC network. After actively sampling the stable set, obtain sample labels and a data set. The closed-loop training includes classifier training, sampling, screening, labeling, and SVSR subset identification. The classifier training refers to training the classifier using an extended training set that includes newly selected samples and existing labeled samples. The sampling refers to randomly selecting a batch of samples from the unlabeled sample space. The screening refers to evaluating the posterior probability of the selected samples to determine their uncertainty and selecting the samples with the highest uncertainty as the new candidate labeling set. The labeling refers to annotating the screened candidate samples through network numerical simulation. The SVSR subset identification refers to identifying the SVSR subset that supports the labeling process using the physical characteristics of the network.

[0086] Set the feasibility posterior probability , and the feasibility posterior probability is used to measure the probability of the power flow having a solution. When 's value is close to 1 or 0, it means the sample is solvable (i.e., belongs to the target category) or unsolvable (i.e., does not belong to the target category). When 's value is close to 0.5, it means the model is uncertain about the classification result of this sample, and analyze the feasibility posterior probability . If is greater than 0.5, define the quantified uncertainty measure as:

[0087] ;

[0088] If is less than or equal to 0.5, define the quantified uncertainty measure as: ;

[0089] Set the training cycle, sort the samples according to the quantified uncertainty, and through the training cycle, record the samples with the highest uncertainty as candidate samples. The initial sample quantity and the number of samples in each training cycle can be set.

[0090] It should be further noted that in the specific implementation process, the process of performing numerical simulation tests on the DC network in combination with candidate samples to obtain test results is as follows:

[0091] In the numerical simulation test, perform tests on the IEEE14-bus, 39-bus, and 118-bus DC networks, perform numerical calculations through IPOPT, and perform simulations through Python and Julia to obtain candidate samples and record the test data to obtain the test results.

Claims

1. A method for analyzing voltage stability region based on active sampling of network information, characterized in that, Including the following steps: S1. Obtain the basic network data of the DC network, construct a network power flow model, analyze the network power flow model, and obtain the voltage stability region; S2. Analyze the voltage stability region according to the network power flow model to obtain the stable set; S3. Analyze the stable set to obtain candidate samples; S4. Combine the candidate samples and conduct numerical simulation tests on the DC network to obtain test results.

2. The voltage stability region analysis method based on active sampling of network information according to claim 1, wherein, The process of obtaining the basic network data of the DC network and constructing a network power flow model includes: The basic network information includes network structure information and network relationship information; the network structure information includes voltage source information and load information; the network relationship information includes the relationship between voltage sources and loads and the relationship between loads; Connect the voltage sources and loads through the relationship between voltage sources and loads and the relationship between loads to construct a DC network model.

3. A method for analyzing a voltage stability region based on active sampling of network information according to claim 2, characterized in that, The process of analyzing the network power flow model to obtain the voltage stability region includes: Obtain the operating points of the DC network model, denoted as node 1, node 2, node 3, ……, node i, ……, node j, ……, node u; when there is a power line between node i and node j, it is denoted as ; when there is no power line between node i and node j, it is denoted as ; denote the conductance of the power line between node and node as , is a positive real number, then the Kirchhoff matrix of the DC network model is: ； By judging whether the node is a load or a voltage source, the Kirchhoff matrix is partitioned and denoted as ; where is the number of loads; is the number of voltage sources; it is stipulated that the matrices and are positive definite; Obtain the load power, load voltage, external power supply voltage, and controllable voltage setpoint. Then, the vectors of the load power, load voltage, external power supply voltage, and controllable voltage setpoint are respectively , , , and , where ; Denote the voltage potential vector at the node as , and it is stipulated that ; Denote the current injected into the network at the node as ; Obtain the power vector p provided by the node to the power grid according to the voltage potential vector at the node and the current injected into the network at the node; ； Through Kirchhoff's and Ohm's laws, is expressed as: ； Obtain the power vector according to formula (3) and formula (4); ； Denote the system steady-state operating point as , denote the load bus voltage at the operating point as , denote the power vector input by the power node to the system as ; According to the system steady-state operating point, the load bus voltage at the operating point, the power vector input by the power node to the system, and Equation (5), perform parameter substitution for each term on the right side of Equation (5) to obtain the static power flow equation, Equation (6); Input the static power flow equation into the DC network model, construct the network power flow model, and analyze the static power flow equation; ； Among them, , and are the conductance matrices of the power-source, power-load, and load-load subnetworks in the network flow model, respectively; = , and are submatrices of the admittance matrix; For the state variable is quadratic, and for the variables and the state variable is bilinear.

4. The voltage stability region analysis method based on active sampling of network information according to claim 3, characterized in that The process of analyzing the static power flow equation is: Obtain the solution of the static power flow equation, denoted as the static operating point. If the static power flow equation has no solution, it means that the system's static voltage is unstable; The voltage stability region is defined as: ； If any given controllable voltage setpoint exists such that for all load injection powers, equation (6) has a solution, then the power set is denoted as the voltage stability region.

5. The voltage stability region analysis method based on active sampling of network information according to claim 4, characterized in that, The process of analyzing the voltage stability region according to the network power flow model to obtain the stable set includes: It is stipulated that under the conditions given by the system, there exists a high load value , and this load value satisfies the voltage stability region conditions of the DC network , then when , it means that the voltage stability region conditions of the DC network are satisfied; When When the static power flow equations have solutions, all have solutions; Construct a stable subdomain based on the characteristics of the voltage stability region, where the characteristics of the voltage stability region refer to monotonicity and convex set properties ; Suppose there are m elements in the voltage stability region ) , which form a stable set P; 。 6. The method for analyzing voltage stability region based on active sampling of network information according to claim 5, wherein The process of analyzing the stable set to obtain candidate samples includes: Perform closed-loop training on the stable set of the DC network. After actively sampling the stable set, obtain sample labels and a data set. The closed-loop training includes classifier training, sampling, screening, labeling, and SVSR subset identification; classifier training refers to training the classifier using an extended training set that includes newly selected samples and existing labeled samples; sampling refers to randomly selecting a batch of samples in the unlabeled sample space; screening refers to evaluating the posterior probability of the selected samples to determine their uncertainty and selecting the samples with the highest uncertainty as the new candidate label set; labeling refers to annotating the candidate samples selected by screening through network numerical simulation; SVSR subset identification refers to identifying the SVSR subset that supports the labeling process using the physical characteristics of the network; Set the posterior probability of feasibility , where the posterior probability of feasibility is used to measure the probability of a solvable power flow; the probability of a solvable static power flow equation is analyzed through the posterior probability of feasibility.

7. A method for analyzing voltage stability region based on active sampling of network information according to claim 6, characterized in that, The process of analyzing the solvability probability of the static power flow equation through the feasibility posterior probability includes: When is close to 1 or 0, it indicates that the sample is solvable or unsolvable; when is close to 0.5, it indicates that the classification result of the model for the sample is uncertain, and the feasibility posterior probability is analyzed. If is greater than 0.5, the quantified uncertainty measure is defined as: ； If is less than or equal to 0.5, the quantified uncertainty measure is defined as: ; Set the training cycle, sort the samples according to the quantified uncertainty, and through the training cycle, record the samples with the highest uncertainty as candidate samples; the initial sample quantity and the number of samples in each training cycle can be set.

8. A method for analyzing voltage stability region based on active sampling of network information according to claim 7, characterized in that, The process of combining the candidate samples and conducting numerical simulation tests on the DC network to obtain test results includes: In the numerical simulation test, conduct tests on the IEEE 14-node, 39-node, and 118-node DC networks, perform numerical calculations through IPOPT, conduct simulations through Python and Julia, obtain candidate samples, record the test data, and obtain the test results.

Images