A power distribution network topology real-time checking and early warning method based on topology-measurement consistency analysis

Abstract

The application discloses a kind of based on topological-quantity measurement consistency analysis distribution network topology real-time checking and early warning method, comprising the following steps: first, with high-precision synchronous measurement data as time reference, time alignment is carried out to multi-source measurement data and SCADA topological data;Second, construct the quantification analysis model based on weighted least square method, calculate the consistency index of whole network to determine the topological correctness;When inconsistent, the branch state suspicious degree is calculated by fusing state estimation residual and voltage space correlation, abnormal positioning and early warning are realized;Finally, generate backup topology set and evaluate its consistency, select the optimal topology and output for correction.The application solves the problem of distribution network partition basic distortion caused by uneven quality of distribution network measurement data, realizes the quantification analysis and real-time topology checking of the internal consistency of topology and real-time measurement data under the online operation state of distribution network, and provides a reliable foundation for subsequent application of distribution network.

Description

Technical Field

[0001] This invention relates to a real-time sensing and state estimation method, and more particularly to a real-time verification and early warning method for distribution network topology based on topology-measurement consistency analysis. Background Technology

[0002] Currently, the high proportion of distributed renewable energy access has led to strong randomness and high volatility in the source-load characteristics of the distribution network. The plug-and-play nature of distributed power sources, load switching, and network reconfiguration requirements have also caused frequent changes in the topology of the distribution network, exhibiting significant time-varying characteristics.

[0003] The existing technical solutions for topology models mainly include the following: 1. Direct application of SCADA remote signaling data: Currently, the dispatch master station generally uses the remote signaling signals of switches and disconnectors uploaded by the SCADA system to directly generate the network topology. This method takes the correctness of the remote signaling data as a default premise and is the foundation of the entire monitoring system. 2. Manual on-site verification: When operators find abnormal measurement data or receive fault alarms, they rely on dispatching maintenance personnel to the site to verify the actual position of the switches and manually correct the master station topology. 3. Offline topology analysis based on historical data: Using historically archived SCADA measurement data or AMI data at a specific time, post-event analysis is performed through state estimation algorithms. This infers possible topology errors with a large time delay and cannot support online applications.

[0004] The aforementioned existing technologies are ill-suited to the high real-time requirements of new active distribution networks. Their shortcomings are mainly reflected in the following aspects: 1. In practical engineering, due to communication delays, device failures, and other reasons, the remote signaling data recorded by the SCADA system often desynchronizes or contains errors with the actual switch status, leading to inconsistencies between topology and measurement. 2. Manual inspection and offline analysis are time-consuming, resulting in a gap of several hours or even days between the occurrence and correction of topology errors, making it unable to cope with rapidly changing operating conditions. 3. Existing technologies fail to effectively integrate the spatiotemporal redundancy information provided by the increasingly abundant multi-source measurement data such as AMI and PMU in the distribution network, lacking a real-time online closed-loop verification mechanism. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide a method for real-time verification and early warning of distribution network topology that features online real-time verification, high data reliability, and a high degree of automation.

[0006] Technical solution: The method for real-time verification and early warning of distribution network topology based on topology-measurement consistency analysis described in this invention includes the following steps:

[0007] (1) Collect multi-source measurement data and SCADA remote signaling topology data of the distribution network, and use high-precision synchronous measurement data as the time reference to complete the data alignment of the time section by interpolation method;

[0008] (2) Based on the preprocessed data, a topology-measurement consistency quantitative analysis model is constructed, and the topology-measurement consistency index of the entire network is calculated to determine whether the topology and measurement data are consistent.

[0009] (3) When the judgment is inconsistent, the spatiotemporal distribution characteristics of the state estimation residual are used to calculate the branch state suspicion degree based on the correlation between the standardized residual and the voltage, locate the topological anomaly and generate an early warning signal.

[0010] (4) For the suspected abnormal branch, the switch state is flipped to generate a set containing one or more candidate topologies; for each topology in the candidate topology set, the topology-measurement consistency index is recalculated, and the topology with the smallest recalculated topology-measurement consistency index and the electrical connection constraint is selected as the corrected real-time topology output.

[0011] Preferably, the multi-source measurement data in step (1) includes SCADA telemetry data, advanced measurement system data, and synchronous phasor measurement unit data.

[0012] Preferably, the high-precision synchronous measurement data in step (1) is synchronous phasor measurement unit data.

[0013] Preferably, the construction of the topology-measurement consistency quantitative analysis model in step (2) includes the following steps:

[0014] (21) Establish the objective function based on weighted least squares method :

[0015]

[0016] in, For measurement vectors; For nonlinear measurement functions determined by the current topology; This is a state estimate; Here is the measurement error covariance matrix; T is the transpose operation;

[0017] (22) Define the topology-measurement consistency index based on the objective function :

[0018]

[0019] in, For the total number of measurements; For the first Residuals of individual measurements; is the standard deviation of the residuals.

[0020] Preferably, when the topology-measurement consistency index calculated in step (2) is greater than a preset threshold, it is determined that the current topology is inconsistent with the measurement data; the preset threshold is determined based on the statistical characteristics of the measurement data.

[0021] Preferably, the formula for calculating the branch status suspicion degree in step (3) is:

[0022]

[0023] in, branch road Suspiciousness of the status The ratio of the standardized residual of the branch-related measurements to the maximum residual of the entire network; These are the weighting coefficients; For the nodes at both ends of the branch road and The Pearson correlation coefficient of the voltage amplitude sequence.

[0024] Preferably, the formula for calculating the Pearson correlation coefficient is as follows:

[0025]

[0026] in, This represents the number of sampling points; For a moment node The voltage amplitude; This represents the average voltage.

[0027] Preferably, when locating topology anomalies, a preset state suspicion warning value is set. If the state suspicion value of a certain branch is greater than the preset state suspicion warning value, the branch is marked as a suspected abnormal topology, and the system issues a topology anomaly warning signal. The state suspicion warning value is determined based on historical operating data statistics.

[0028] Preferably, the compliance with electrical connection constraints in step (4) specifically refers to:

[0029] For a branch whose state is flipped to "closed", the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≥0.85; for a branch whose state is flipped to "open", the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≤0.3.

[0030] Meanwhile, the modified topology satisfies the radial operation constraints of the distribution network.

[0031] Preferably, the data alignment in step (1) adopts a hierarchical interpolation alignment strategy; wherein, the synchronous phasor measurement unit data directly adopts the original synchronous sampling value that is closest to the target time section; the SCADA telemetry data adopts the point-by-point linear interpolation method; and the advanced measurement system data adopts the piecewise linear interpolation fitting method.

[0032] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) It can process topology verification online in real time, which greatly shortens the time for topology error detection and correction; (2) It significantly improves the objectivity and accuracy of anomaly identification and accurately locates suspected abnormal branches, reducing the false alarm rate; (3) It has the ability to automatically generate alternative topologies and select the optimal solution, realizing the self-correction of topology data and reducing manual intervention; (4) It outputs high-reliability real-time topology data, providing a solid data foundation for subsequent applications. Attached Figure Description

[0033] Figure 1 This is a flowchart of the present invention;

[0034] Figure 2 This is a topology diagram of a verification example of the present invention;

[0035] Figure 3 This is a comparison chart of the residual distribution of state estimation under different operating conditions according to the present invention;

[0036] Figure 4 This is a heat map of the correlation analysis of voltage amplitude in the fault area according to the present invention. Detailed Implementation

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0038] As shown in the attached figure, the real-time verification and early warning method for distribution network topology based on topology-measurement consistency analysis disclosed in this invention includes four steps: data alignment and preprocessing, model construction and consistency verification, anomaly location and early warning, and topology correction and output.

[0039] Step 1: Data Alignment and Preprocessing

[0040] Collect multi-source heterogeneous measurement data of the distribution network and remote signaling topology data reported by the SCADA system. The multi-source heterogeneous measurement data includes SCADA telemetry data, Advanced Measurement System (AMI) data, and synchronous phasor measurement unit (PMU) data, covering electrical quantities such as node voltage amplitude, phase angle, and branch power.

[0041] To eliminate data asynchrony issues caused by communication delays and sampling frequencies, a hierarchical interpolation alignment strategy using high-precision synchronized PMU data as a unified time reference is adopted. Specifically, the GPS synchronization timescale of the PMU is used as the unified time section reference for the entire network, and different processing methods are applied to data from different sources:

[0042] PMU data: This is high-frequency data, which directly uses the original synchronous sampling value that is closest to the target time segment.

[0043] SCADA telemetry data: its sampling period is on the order of seconds, and interpolation alignment is performed using point-by-point linear interpolation.

[0044] AMI data: This is low-frequency data, with a collection period of 15 minutes, and interpolation alignment is performed using a piecewise linear interpolation fitting method.

[0045] Through the above processing, a standardized measurement section with strict consistency in time and spatiotemporal matching is formed, providing a synchronous data foundation for subsequent analysis.

[0046] Step 2: Model Building and Consistency Verification

[0047] This step aims to quantitatively assess the consistency between the current SCADA topology and real-time measurement data.

[0048] 1. Establish a state estimation model

[0049] Based on the remote signaling topology provided by the current SCADA system, a state estimation objective function based on the weighted least squares method is established:

[0050]

[0051] in, For measurement vectors; For nonlinear measurement functions determined by the current topology; Here is the state estimate; T is the transpose operation; Let be the measurement error covariance matrix, a positive definite diagonal matrix. Assume that all measurement errors are independent, with their off-diagonal elements being zero, and the diagonal elements representing the variance of the errors for each measurement type, assigned values ​​based on the measurement type and equipment accuracy:

[0052] (1) Variance of voltage measurement : ;

[0053] (2) Variance of active power measurement : ;

[0054] (3) Variance of reactive power measurement : ;

[0055] (4) Variance of current measurement : ;

[0056] in, , , , These are the rated or reference values ​​for voltage, active power, reactive power, and current, respectively. Multi-source measurements are assigned values ​​based on accuracy: PMU measurements have the smallest variance, followed by SCADA measurements, and AMI measurements have a slightly larger variance. These are arranged sequentially by measurement sequence number to form a diagonal matrix.

[0057] The Newton-Raphson iterative algorithm is used to solve the above objective function to obtain the state estimate that minimizes the objective function.

[0058] 2. Calculate the consistency index

[0059] Based on the state estimation results, calculate the network-wide topology-measurement consistency index. The calculation formula is:

[0060]

[0061] in, For the total number of measurements; For the first Residuals of individual measurements; is the standard deviation of the residuals.

[0062] 3. Consistency judgment

[0063] A consistency index threshold is set, which is based on the standardized residual statistical characteristics of distribution network state estimation and can be adjusted according to the specific network and measurement configuration. In this embodiment, the consistency index threshold is set to 20. When the calculated consistency index... If the consistency index threshold is exceeded, it is determined that the topology recorded by the current SCADA system is inconsistent with the real-time measurement data, indicating a topology anomaly, and the process proceeds to step 3; otherwise, the topology is considered reliable, and the process ends.

[0064] Step 3: Anomaly Location and Early Warning

[0065] Once a topology anomaly is detected, this step is used to locate the suspicious abnormal branch.

[0066] 1. Calculate the suspicion level of branch status

[0067] The state suspicion level of each branch is calculated using the following formula:

[0068]

[0069] in, branch road Suspiciousness of the status The ratio of the standardized residual of the branch-related measurements to the maximum residual of the entire network; For the nodes at both ends of the branch road and The Pearson correlation coefficient of the voltage amplitude sequence. The formula for calculating the Pearson correlation coefficient is:

[0070]

[0071] in, This represents the number of sampling points; For a moment node The voltage amplitude; This represents the average voltage.

[0072] The weighting coefficient is set to 0.5. Based on multi-scenario simulation tests and statistical analysis of historical operating data of the improved IEEE 33-bus system, the weighting coefficient is set to balance the contributions of standardized residual characteristics and Pearson voltage correlation characteristics to the state suspicion level, ensuring that the anomaly detection sensitivity and anti-interference capability are optimal.

[0073] 2. Issue an early warning

[0074] The branch status suspicion threshold is set to 0.7. This value is derived from a statistical comparison of the status suspicion distribution under historical normal and abnormal operating conditions. All branches are iterated over; if a branch meets the condition suspicion threshold of 0.7, it is marked as a "suspected abnormal topology," and the system issues a topology anomaly warning signal, clearly indicating the switch location suspected of having an error.

[0075] Step 4: Topology Correction and Output

[0076] This step generates and outputs a corrected topology for suspected abnormal branches identified in the early warning.

[0077] 1. Generate and evaluate candidate topology sets

[0078] For all marked suspected abnormal branches, a set of alternative topologies is generated by "flipping" their switch states (changing their state from "closed" to "open", or vice versa).

[0079] For each candidate topology in the set, re-execute the complete process of step 2 to calculate its corresponding topology-measurement consistency index and the voltage Pearson correlation coefficient of the nodes at both ends of the relevant branch;

[0080] 2. Select the optimal topology

[0081] From the set of candidate topologies, the optimal topology is selected as the corrected result. The selection criteria must simultaneously satisfy the following:

[0082] (1) Optimal objective function: minimize the network topology-measurement consistency index.

[0083] (2) Meets the physical characteristics of electrical connection:

[0084] (21) For a branch whose state is flipped to “closed”, the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≥0.85, indicating strong electrical coupling;

[0085] (22) For a branch whose state is flipped to "disconnected", the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≤0.3, indicating weak electrical coupling or no coupling;

[0086] (23) The modified topology satisfies the radial operation constraints of the distribution network, has no electrical ring network and no isolated nodes, and maintains the legal topology connectivity.

[0087] 3. Output a reliable topology

[0088] The optimal topology that meets the above conditions will be output as the corrected and reliable real-time topology data to support subsequent advanced applications such as dynamic zoning and optimized control of the distribution network.

[0089] Verification example:

[0090] This verification example constructs an improved IEEE 33-bus distribution system incorporating high-penetration distributed power sources for simulation testing. The reference voltage is 12.66 kV, and the reference power is 10 MVA. Its topology is shown in the attached figure. Figure 2 As shown.

[0091] Considering the dual uncertainties of the power supply and load in the new distribution network, the following modifications were made to the original network:

[0092] (1) High proportion of new energy access: Distributed photovoltaic (PV) and wind power (WT) are connected at nodes 18, 22, 25 and 33 respectively, with the total installed capacity accounting for 65% of the system peak load, forming a typical active distribution network.

[0093] (2) Heterogeneous measurement configuration: Simulate the actual power distribution Internet of Things environment, configure synchronous phasor measurement units (PMU) at the substation outlet and key interconnection switches, and configure smart meters (AMI) and traditional SCADA measurement systems at other nodes.

[0094] (3) Noise environment simulation: The measurement data superimposed according to the following Distributed Gaussian white noise, in which voltage amplitude error Power injection error To test the robustness of the algorithm under non-ideal measurement environments.

[0095] Detailed distributed resource configuration parameters are shown in Table 1. The simulation platform is built based on MATLAB R2024b, and Matpower 7.1 is used to perform power flow calculations and generate measurement reference data.

[0096] Table 1. Improved Distributed Resource Configuration Parameters for the IEEE 33-Node System

[0097]

[0098] Scene settings: Configure the system in At that moment, the tie switch on branch 6-26 was mistakenly closed from an open state, but due to a communication failure, the remote signaling status displayed by the SCADA system remained open. At this time, the topology model built based on the SCADA data... With actual physical power grid A deviation has occurred.

[0099] The topology-measurement consistency index proposed in this invention Real-time monitoring of system status. (Attached) Figure 3 The comparison of the weighted normalized residual (WNR) distribution under normal operation and topology error conditions is shown.

[0100] From the appendix Figure 3 It is evident that, under normal operating conditions, the state estimation residuals conform to the standard... Distribution, Consistency Indicators Maintain at a low level ( However, when a topology error occurs in branch 6-26, the severe mismatch between the physical constraint equations and the measurement data leads to a significant distortion in the residual distribution across the entire network. The surge far exceeded the warning threshold. Simulation results show that the consistency index can effectively shield against the interference of random measurement noise, exhibits extremely high sensitivity to topological structural variations, and can accurately trigger the anomaly warning mechanism.

[0101] After the warning is triggered, the voltage correlation coefficient is further utilized. The abnormal region was located. Node 6 and its neighboring nodes were selected, and the Pearson correlation coefficient matrix of their voltage amplitude time series was calculated, as shown in the attached figure. Figure 4 As shown.

[0102] Under the assumption of SCADA topology false alarm disconnection, node 6 and node 26 are electrically far apart, and the theoretical correlation coefficient should be low. However, (attached) Figure 4 Data analysis shows that the voltage correlation coefficient between node 6 and node 26 is as high as This exhibits strong coupling characteristics, which aligns with the actual state of physical closure. Simultaneously, the state suspicion index... The accuracy reached 0.89. By combining residual distribution and correlation characteristics, the algorithm successfully pinpointed the topology error to branch 6-26 and automatically corrected the real-time topology model, providing reliable data for dynamic zoning of the distribution network.

Claims

1. A method for real-time verification and early warning of distribution network topology based on topology-measurement consistency analysis, characterized in that, Includes the following steps: (1) Collect multi-source measurement data and SCADA remote signaling topology data of the distribution network, and use high-precision synchronous measurement data as the time reference to complete the data alignment of the time section by interpolation method; (2) Based on the preprocessed data, a topology-measurement consistency quantitative analysis model is constructed, and the topology-measurement consistency index of the entire network is calculated to determine whether the topology and measurement data are consistent. (3) When the judgment is inconsistent, the spatiotemporal distribution characteristics of the state estimation residual are used to calculate the branch state suspicion degree based on the correlation between the standardized residual and the voltage, locate the topological anomaly and generate an early warning signal. (4) For the suspected abnormal branch, the switch state is flipped to generate a set containing one or more candidate topologies; for each topology in the candidate topology set, the topology-measurement consistency index is recalculated, and the topology with the smallest recalculated topology-measurement consistency index and the electrical connection constraint is selected as the corrected real-time topology output.

2. The real-time verification and early warning method according to claim 1, characterized in that, The multi-source measurement data mentioned in step (1) includes SCADA telemetry data, advanced measurement system data, and synchronous phasor measurement unit data.

3. The real-time verification and early warning method according to claim 2, characterized in that, The high-precision synchronous measurement data mentioned in step (1) is synchronous phasor measurement unit data.

4. The real-time verification and early warning method according to claim 1, characterized in that, The construction of the topology-measurement consistency quantitative analysis model in step (2) includes the following steps: (21) Establish the objective function based on weighted least squares method : in, For measurement vectors; For nonlinear measurement functions determined by the current topology; This is a state estimate; Here is the measurement error covariance matrix; T is the transpose operation; (22) Define the topology-measurement consistency index based on the objective function : in, For the total number of measurements; For the first Residuals of individual measurements; is the standard deviation of the residuals.

5. The real-time verification and early warning method according to claim 1, characterized in that, When the topology-measurement consistency index calculated in step (2) is greater than the preset threshold, it is determined that the current topology is inconsistent with the measurement data; the preset threshold is determined based on the statistical characteristics of the measurement data.

6. The real-time verification and early warning method according to claim 1, characterized in that, The formula for calculating the branch status suspicion level in step (3) is as follows: in, branch road Suspiciousness of the status The ratio of the standardized residual of the branch-related measurements to the maximum residual of the entire network; These are the weighting coefficients; For the nodes at both ends of the branch road and The Pearson correlation coefficient of the voltage amplitude sequence.

7. The real-time verification and early warning method according to claim 6, characterized in that, The formula for calculating the Pearson correlation coefficient is as follows: in, This represents the number of sampling points; For a moment node The voltage amplitude; This represents the average voltage.

8. The real-time verification and early warning method according to claim 6, characterized in that, The specific steps of locating the topology anomaly and generating the early warning signal in step (3) are as follows: a preset state suspicion warning value is set. If the state suspicion of a branch is greater than the state suspicion warning value, the branch is marked as a suspected abnormal topology, and the system issues a topology anomaly early warning signal. The state suspicion warning value is determined based on historical operational data statistics.

9. The real-time verification and early warning method according to claim 7, characterized in that, The compliance with electrical connection constraints mentioned in step (4) specifically refers to: For a branch whose state is flipped to "closed", the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≥0.85; for a branch whose state is flipped to "open", the Pearson correlation coefficient of the voltage amplitude at both ends of the branch is ≤0.

3. Meanwhile, the modified topology satisfies the radial operation constraints of the distribution network.

10. The real-time verification and early warning method according to claim 2, characterized in that, The data alignment described in step (1) adopts a hierarchical interpolation alignment strategy; wherein, the synchronous phasor measurement unit data directly adopts the original synchronous sampling value that is closest to the target time section; the SCADA telemetry data adopts the point-by-point linear interpolation method; and the advanced measurement system data adopts the piecewise linear interpolation fitting method.

Images