An adaptive robust state estimation method and system based on PMU and SCADA multi-source measurement

Abstract

The application discloses a kind of adaptive robust state estimation method and system based on PMU and SCADA multi-source measurement, method includes: by section matching degree analysis, sliding time window and power characteristic analysis, generate consistent hybrid measurement section data;By calculating the deviation of measurement data and state evaluation value, combined with historical comprehensive reliability evaluation index and real-time quality evaluation index check, quality analysis is carried out to multi-source measurement data;According to evaluation index, adaptively adjust the measurement source participating in calculation, eliminate abnormal data;By PMU phase angle reference conversion and transformer wiring group correction, the hybrid measurement data screened is substituted into weighted least square model and is fused to solve, obtains system optimal state section.Solve the problem of time asynchronization and quality heterogeneity of hybrid measurement data, improve the precision and robustness of power grid state estimation.

Description

Technical Field

[0001] This invention relates to the field of power system state estimation, specifically to an adaptive robust state estimation method and system based on PMU and SCADA multi-source measurements. Background Technology

[0002] With the large-scale grid connection of new energy sources and the random access of new loads, the new power system exhibits a trend of multi-element coupling of sources, grids, loads, and storage, and complex and ever-changing operating states. This significantly increases the demand for real-time monitoring and accurate state estimation of the power system. Power system state estimation is a core technology for power grid dispatch centers to achieve safety early warning and economic dispatch. Its core objective is to infer the current operating state of the power grid based on multi-source measurement data, providing a reliable basis for subsequent control decisions.

[0003] Currently, the commonly used measurement systems in power systems mainly include two types: one is the Supervisory and Data Acquisition System (SCADA), which collects data such as node voltage amplitude and injected power through remote terminal units (RTUs). Although it has a wide coverage, it has defects such as long measurement update cycle (usually 2-5 seconds), lack of precise time scale for data, and missing phase angle measurements. The other is the synchronous phasor measurement unit (PMU), which achieves millisecond-level time synchronization based on the Global Positioning System (GPS) and can directly collect key dynamic measurement data such as node voltage phase angle and branch current phase angle. However, its deployment cost is high, and it is currently only configured in hub nodes and key branches in the power grid, which is difficult to meet the needs of the entire network state estimation.

[0004] To balance measurement coverage and accuracy, a hybrid measurement mode combining SCADA and PMU is commonly used for state estimation. However, in practical engineering applications, the fusion of multi-source measurement data still faces the following significant challenges: First, data mismatch caused by inconsistent time scales among multi-source measurements. SCADA data update cycles are typically on the order of seconds and lack independent time stamps, while PMU data sampling frequencies are on the order of milliseconds and are synchronized based on satellite time synchronization. This significant difference in time scale leads to time-series asynchrony issues between measurement data from different sources within the same computational cross-section. If precise alignment of measurement data cannot be achieved, information conflicts will occur, introducing large pseudo-measurement errors and severely affecting the accuracy of state estimation. This is especially true in scenarios with rapid power fluctuations, such as UHV lines and new energy collection lines, where local time-series deviations have a more severe impact on the estimation results. Second, reliability bottlenecks caused by heterogeneity in the quality of multi-source measurements. Affected by the accuracy of measurement equipment, communication interference, and operating environment, the quality of data from different sources varies significantly. SCADA data is susceptible to deviations caused by instrument errors and transmission noise; while PMU data has high accuracy, it may also exhibit polarity errors or data anomalies under power grid faults or electromagnetic interference. Existing processing methods often lack multi-dimensional quality assessment indicators, making it difficult to identify and eliminate bad data in real-time computation, resulting in poor robustness of state estimation when facing data fluctuations. Thirdly, traditional algorithms face a trade-off between real-time performance and accuracy when processing massive amounts of data. Existing hybrid measurement state estimation methods are mostly based on weighted least squares (WLS) for nonlinear iterative solutions. For large-scale power grids with tens of thousands of nodes, traditional iterative methods involve enormous computational loads, making it difficult to meet the minute-level latency requirements of online analysis. Furthermore, when processing phase angle measurements, traditional methods often neglect the reference conversion between the PMU absolute phase angle and the state estimation reference phase angle, as well as the phase shift caused by transformer connection groups (such as Y-Δ connections), further limiting the improvement of computational accuracy.

[0005] In summary, existing state estimation methods still have shortcomings in terms of multi-source measurement cross-section alignment, multi-dimensional dynamic quality evaluation, and efficient hybrid solution, making it difficult to meet the requirements of new power systems for high precision, strong robustness, and real-time performance. Summary of the Invention

[0006] Purpose of the invention: The purpose of this invention is to address the shortcomings of the existing technology and provide an adaptive robust state estimation method and system based on PMU and SCADA multi-source measurements, thereby improving the accuracy and robustness of power system state estimation.

[0007] Technical solution: The adaptive robust state estimation method based on PMU and SCADA multi-source measurements described in this invention includes:

[0008] By analyzing the cross-sectional matching degree and the sliding time window and power characteristics, the cross-sections of PMU measurement data, SCADA measurement data and TSCA measurement data are aligned and the equipment-level offset is corrected, generating consistent mixed measurement data.

[0009] Construct a comprehensive quality evaluation system for multi-source measurements. Specifically, this involves calculating the deviation between the measurement data and the state assessment value, and verifying the data by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators to conduct quality analysis on the multi-source measurement data.

[0010] The measurement data used in the calculation are adaptively adjusted according to the evaluation indicators, and abnormal and low-quality data are removed.

[0011] By converting the PMU phase angle reference and correcting the transformer connection group, the screened mixed measurement data is substituted into the weighted least squares model for fusion solution to obtain the optimal state section of the system.

[0012] Furthermore, the cross-sectional matching degree analysis and the sliding time window and power characteristic analysis specifically include:

[0013] Using the time of SCADA measurement data as the reference measurement time, the cross-sectional matching degree between SCADA measurement data and PMU measurement data and the cross-sectional matching degree with TSCA measurement data are calculated respectively. The cross-sectional matching degree is calculated based on the weight of each measurement and the deviation between each measurement data.

[0014] Based on the SCADA baseline measurement time, a sliding time window is constructed. The matching degree between the k-th time scale section of the PMU measurement data and the TSCA measurement data under this time window and the SCADA section is calculated respectively. The matching degree of all sections of the PMU measurement data and the TSCA measurement data is sorted in descending order, and the section corresponding to the maximum matching degree is selected as the optimal time scale measurement section.

[0015] The power characteristic analysis identifies intervals with consistent power trends by constructing a power fluctuation rate index and a power trend consistency analysis index. Within the interval, it matches the characteristic moments of SCADA measurement data and PMU measurement data, calculates the time offset, and corrects the time offset of the PMU measurement.

[0016] Furthermore, the process of calculating the deviation between the measurement data and the state assessment value, and then verifying it by combining historical comprehensive reliability assessment indicators and real-time quality assessment indicators, to perform quality analysis on the multi-source measurement data, specifically involves:

[0017] In the mixed measurement calculation scenario, based on the current and historical state estimation calculation results, the deviations between the PMU measurement, TSCA measurement and SCADA measurement and the state estimation results are calculated respectively.

[0018] For each measurement point, count the number of times each measurement source's deviation is greater than or equal to a set threshold and less than a set threshold within a time window, and calculate the comprehensive reliability index of each measurement source based on the count.

[0019] Within a single calculation section, real-time closed-loop verification is performed on the aligned measurement data. This real-time closed-loop verification includes three types: measurement polarity quality verification, where if the polarity of the timescale measurement is opposite to that of the SCADA measurement, and the SCADA imbalance of the corresponding node is less than a set imbalance threshold, then the timescale measurement is inverted; measurement difference quality verification, where if the absolute difference between the timescale measurement and the SCADA measurement is greater than a set threshold, and the percentage of the relative difference is greater than a percentage threshold, then the measurement point is determined to be an abnormal measurement point and is removed; and measurement balance quality verification, used to perform active power balance verification and reactive power balance verification on the line, transformer windings, and calculation nodes according to physical topology constraints.

[0020] Furthermore, the measurement balance quality verification specifically includes:

[0021] Line balance verification: Determine whether the absolute value of the difference between the active power at the beginning and end of the line is less than the line balance threshold.

[0022] Main transformer balance verification: Determine whether the algebraic sum of active power on each side of the main transformer is less than the main transformer balance threshold;

[0023] Node balance check: Determine whether the algebraic sum of the power of all associated branches on the same computing node is less than the node balance threshold.

[0024] Furthermore, the adaptive adjustment of the measurement data used in the calculation based on the evaluation indicators, and the removal of abnormal and low-quality data, specifically includes:

[0025] Using the comprehensive reliability index of each measurement source within the time window, admission and exit thresholds are set for PMU measurements and TSCA measurements respectively. Measurement sources whose comprehensive reliability index reaches the admission threshold are set to participate in the current calculation and whose reliability meets the calculation requirements; measurement sources whose comprehensive reliability index is lower than the exit threshold are set to exit the current calculation and are judged as having insufficient reliability. Those whose reliability meets the requirements are selected as high-reliability measurement source candidate pool.

[0026] Within the selected measurement candidate pool, when there are multiple measurement sources at the same measurement point, a horizontal comparison is made based on the current cross-sectional calculation deviation during a single state estimation calculation. The measurement source with the smallest real-time deviation value is selected as the input data for that measurement point in the current state estimation calculation.

[0027] Historical comprehensive reliability indicators are used to eliminate measurement sources with long-term unstable performance, and real-time quality assessment indicators are used to deal with power grid changes or short-term communication anomalies.

[0028] Furthermore, the process involves converting the PMU phase angle reference and correcting the transformer connection group, then substituting the filtered mixed measurement data into a weighted least squares model for fusion and solution to obtain the optimal state profile of the system. Specifically:

[0029] After alignment, the phase angle measurements of the PMU are checked for node phase angle deviation and screened for transmission distance phase angle constraints to eliminate abnormal phase angle measurements.

[0030] The node with the PMU installed is selected as the reference node for state estimation. The phase angle offset of the reference node relative to the satellite timing reference system is obtained. Based on the phase angle offset, all PMU phase angle measurements are uniformly corrected, and the transformer wiring is compensated by group.

[0031] Substitute the aligned, filtered and corrected mixed measurement data into the weighted least squares model to construct an objective function with the goal of minimizing the weighted sum of squared measurement errors, and solve for the optimal state estimation vector.

[0032] Active power deviation and measurement pass rate are selected as evaluation indicators to quantitatively evaluate the state estimation results.

[0033] Furthermore, the node phase angle deviation verification specifically involves: calculating the bus phase angle deviation from different sources within the same computing node; if the deviation is less than a set threshold, the average of the phase angles from each source is taken as input; if the deviation exceeds the set threshold, the abnormal measurement with the largest deviation is removed.

[0034] The transmission distance phase angle constraint screening is specifically as follows: a phase angle constraint range is set according to the transmission distance from each node to the reference node. If the relative phase angle measurement of a node exceeds the constraint range, it is determined to be an abnormal measurement and is removed.

[0035] Furthermore, the process of substituting the aligned, filtered, and corrected mixed measurement data into a weighted least squares model to construct an objective function that minimizes the weighted sum of squared measurement errors, and solving for the optimal state estimation vector, is as follows:

[0036] The hybrid measurement equation system is constructed as follows:

[0037] ;

[0038] Where z is a hybrid measurement data vector, including SCADA measurements, TSCA measurements, and corrected PMU phasor measurements; x is the state vector to be solved, including node voltage magnitude and phase angle; It is a nonlinear measurement function; This is the multi-source measurement error vector;

[0039] Using the weighted least squares method to minimize the objective function J(x), the objective function is constructed as follows:

[0040] ;

[0041] Where R is the error covariance matrix dynamically adjusted according to the quality of the measurement source, and its diagonal elements are the variances of each measurement data value. constitute;

[0042] Taking the partial derivative of the objective function J(x) with respect to x and setting the partial derivative to 0, we derive the closed-form solution for the optimal state estimate as follows:

[0043]

[0044] in, H is the optimal state estimation vector; H is the nonlinear measurement function h(x) in the initial reference state vector. The Jacobian matrix calculated at point ; The weight matrix is ​​determined by the real-time quality of multi-source measurements; This is the residual vector between the measured values ​​and the calculated values ​​based on the reference state, with the superscript T indicating transpose.

[0045] Furthermore, the PMU measurement data includes node voltage amplitude, node voltage phase angle, branch current amplitude, branch current phase angle, and node active power and node reactive power calculated from voltage and current; the SCADA measurement data includes node voltage amplitude, node injected active power, node injected reactive power, inter-node active power, and inter-node reactive power.

[0046] The present invention discloses an adaptive robust state estimation system based on PMU and SCADA multi-source measurements, used to implement the method, comprising:

[0047] Measurement section alignment module: It is used to align the sections of PMU measurement data, SCADA measurement data and TSCA measurement data and correct equipment-level offset through section matching degree analysis, sliding time window and power characteristic analysis, and generate consistent mixed measurement data.

[0048] The comprehensive quality evaluation module is used to construct a comprehensive quality evaluation system for multi-source measurements. Specifically, it performs quality analysis on multi-source measurement data by calculating the deviation between the measurement data and the state assessment value, and by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators for verification.

[0049] Adaptive adjustment module: Used to adaptively adjust the measurement data involved in the calculation according to the evaluation index, and remove abnormal and low-quality data;

[0050] Hybrid Measurement State Solution Module: This module uses PMU phase angle reference conversion and transformer connection group correction to substitute the selected hybrid measurement data into a weighted least squares model for fusion solution, thereby obtaining the optimal state profile of the system.

[0051] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0052] (1) This invention effectively solves the asynchronous problem of PMU, SCADA and TSCA measurement data by using cross-section matching degree analysis and sliding time window screening technology, combined with fine time offset correction for devices with rapid power fluctuations, and generates highly consistent multi-source measurement cross-sections, providing a reliable data foundation for subsequent state estimation.

[0053] (2) The present invention constructs a multi-source measurement comprehensive evaluation system that integrates historical reliability evaluation and real-time quality verification, covering measurement data polarity verification, residual difference verification, power flow balance verification, etc., which can effectively identify and isolate bad data with polarity errors, residuals and balance verification failures.

[0054] (3) This invention proposes a dynamic switching strategy for historical and real-time multi-time dimensions, which selects the optimal data source in real time, effectively avoiding the problem of sudden drop in state estimation accuracy or calculation divergence caused by the abnormality of a single data source, and greatly improving the robustness and stability of the entire state estimation process.

[0055] (4) This invention proposes a method for evaluating the state estimation calculation results. Based on the traditional qualification rate index, it adds multi-dimensional power deviation evaluation indexes such as node injection power deviation and branch power flow deviation, which quantifies the fitting accuracy of the state estimation results with the actual operating conditions of the power grid and realizes a comprehensive evaluation of the state estimation calculation results. Attached Figure Description

[0056] Figure 1 This is a flowchart of the method described in this invention.

[0057] Figure 2 This is a diagram showing the cross-sectional matching results of PMU measurement and SCADA measurement in an embodiment of the present invention. Detailed Implementation

[0058] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0059] This embodiment provides an adaptive robust state estimation method based on PMU and SCADA multi-source measurements, including:

[0060] (1) By analyzing the cross-sectional matching degree and the sliding time window and power characteristics, the cross-sectional alignment and equipment-level offset correction of the measurement data of the synchronous phasor measurement unit (PMU), the measurement data of the monitoring and data acquisition system (SCADA) and the time scale steady state (TSCA) measurement data are achieved, and consistent mixed measurement data are generated.

[0061] Specifically, PMU measurement data includes node voltage amplitude. Node voltage phase angle Branch current amplitude Branch current phase angle and the node active power calculated from voltage and current. and node reactive power SCADA measurement data includes node voltage amplitude. Node-injected active power Node-injected reactive power Active power between nodes Inter-node reactive power .

[0062] Measurement section matching degree analysis calculates the spatiotemporal matching index of multi-source measurements. Specifically, it takes the time of SCADA measurement as the reference measurement time and calculates the section matching degree between power system SCADA measurements and PMU measurements, and the section matching degree between SCADA measurements and TSCA measurements. The calculation formula is:

[0063]

[0064]

[0065] in, The weight of the i-th measurement, It is the i-th SCADA measurement. This is the i-th PMU measurement. This is the i-th TSCA measurement. Sectional matching degree. The larger the value, the closer the PMU measurement, TSCA measurement and the reference time SCADA measurement are, that is, the better the time consistency of the two types of measurement data.

[0066] The optimal time-scale hybrid measurement section selection method includes:

[0067] Construct a sliding time window based on the SCADA baseline measurement time:

[0068]

[0069] Where ΔT represents the time range for selecting measurement times. This is the candidate set for PMU and TSCA measurement data. The k-th time-scale section in the data is denoted as PMU. k TSCA k Calculate PMU separately k TSCA k Matching degree with SCADA profile For candidate set All cross-sections are sorted in descending order of matching degree, and the cross-section with the maximum matching degree is selected as the optimal time scale measurement cross-section.

[0070] In the measurement section matching degree analysis, time offset correction is performed on equipment with rapid power fluctuations, such as ultra-high voltage lines and new energy collection lines, as detailed below:

[0071] Constructing quantitative indicators of power change characteristics: including defining the power fluctuation rate Consistency analysis coefficient with power trend The formulas for quantifying the power dynamic characteristics of the equipment are as follows:

[0072]

[0073]

[0074] In the formula: P i Let be the active power measurement value at time i, Δt be the measurement sampling time interval, n be the number of measurement points within the time window, and sign() be the sign function used to characterize the direction of power change within a single time interval.

[0075] Time offset calculation based on feature point matching: Within the identified interval where the power trend is consistent, the time offset of the PMU measurement is obtained by matching the feature moments of SCADA measurement data and PMU measurement data. Specifically: In the SCADA measurement sequence S sca In the process, the starting point of the interval with the current power trend is selected as the reference feature point. ; in PMU measurement sequence In the process, based on the same feature extraction rules, the PMU feature point time corresponding to the benchmark feature point is located. To suppress the impact of random errors on correction accuracy, multiple continuous feature points (such as the start point, midpoint, and inflection point of a trend interval) are selected, their time differences are calculated, and the average is taken. The resulting average time difference is then calculated. The final PMU measurement offset is calculated using the following formula:

[0076]

[0077] in, Let m be the time of the PMU feature point corresponding to the m-th feature point. Let k be the SCADA feature point time corresponding to the m-th feature point, and k be the number of feature points.

[0078] (2) Construct a comprehensive quality evaluation system for multi-source measurement, specifically by calculating the deviation between the measurement data and the state assessment value, and combining historical reliability assessment and real-time quality verification to conduct quality analysis on the multi-source measurement data;

[0079] Specifically, in the mixed measurement calculation scenario, the calculation deviations of PMU measurement, TSCA measurement and SCADA measurement sources are quantitatively analyzed based on the current and historical calculation output results.

[0080] Deviation between PMU measurements and state estimation results for:

[0081]

[0082] Deviation between TSCA measurement and state estimation results for:

[0083]

[0084] Deviation between SCADA measurement and state estimation results for:

[0085]

[0086] In the formula: , , The active power values ​​are obtained from PMU measurement, TSCA measurement, and SCADA measurement, respectively. The calculated value is for the state estimate.

[0087] For each measurement point, by comparing the relative accuracy of different measurement sources, the number of times the difference in deviation between each measurement source within a time window is greater than or equal to a set threshold and less than a set threshold is counted. Based on the counted frequency, a comprehensive reliability index for each measurement source is calculated. When the difference between the deviation of the PMU measurement and the state estimation result and the difference between the deviation of the SCADA measurement and the state estimation result is greater than or equal to a deviation threshold, i.e. , If the deviation threshold is set, the index count corresponding to the measurement source with a large deviation will increase; conversely, the index count corresponding to the measurement source with a small deviation will increase.

[0088] The overall reliability index γ of the computational measurement source:

[0089]

[0090] in, This is the counting index when the deviation is large. This is a counting index for small deviations; this comprehensive reliability index is used to quantify the overall reliability of measurement data within a certain time window.

[0091] The real-time quality verification is as follows: within a single calculation section, real-time closed-loop verification is performed on the aligned measurement data, specifically including:

[0092] Measurement polarity quality verification: If the polarity of the time scale measurement is opposite to that of the SCADA measurement, and the SCADA imbalance of the corresponding node is less than the set imbalance threshold, then the time scale measurement is inverted and corrected.

[0093] Measurement difference quality verification: If the absolute difference between the time-scale measurement and the SCADA measurement is greater than the set threshold value. If the relative difference percentage is greater than the percentage threshold DPT, the measurement point is determined to be an abnormal measurement point and is removed.

[0094] Measurement balance quality verification: Utilizing physical topology constraints, active power balance and reactive power balance verification are performed on the lines, transformer windings, and computational nodes, respectively. Specifically, this includes:

[0095] Line balance check: Determines whether the absolute value of the difference in active power between the beginning and end of the line is less than the line balance threshold. ;

[0096] Main transformer balance verification: Determine whether the algebraic sum of the active power on each side of the main transformer is less than the main transformer balance threshold. ;

[0097] Node balance check: Determines whether the algebraic sum of the power of all associated branches on the same compute node is less than the node balance threshold. .

[0098] (3) Adjust the measurement data used in the calculation adaptively according to the evaluation indicators and remove abnormal and low-quality data;

[0099] (31) Using the comprehensive reliability index of each measurement source within the time window calculated in step (2), set the admission and exit thresholds, and screen the candidate pool of highly reliable measurement sources from the source, specifically as follows:

[0100] PMU measurement admission criteria: When the overall reliability index of PMU measurements... When the PMU measurement is deemed to meet the calculation requirements, it is set to participate in the current calculation; when If its reliability is deemed insufficient, it is set to exit the current calculation. Add a threshold to PMU measurements. Measure the exit threshold for the PMU;

[0101] TSCA Measurement Admission Criteria: When the overall reliability index of TSCA measurements... When its state is stable and its accuracy meets the standard, it is set to participate in the current calculation; when If the error deviates significantly from the true system state, it is set to exit the current calculation. Add a threshold to TSCA measurements. The TSCA measurement exit threshold.

[0102] The above-mentioned entry and exit thresholds are dynamically configured based on the load fluctuation intensity and measurement equipment characteristics under the power grid operation scenario, so as to balance the universality of the calculation and the adaptability of the scenario.

[0103] (32) Within the candidate pool of measurements selected above, for cases where there are multiple measurement sources at the same measurement point, the measurement source is selected in real time by calculating the deviation comparison in real time. This is used to cope with sudden changes in the power grid or short-term communication anomalies, optimize the selection of local measurement sources, and ensure the robustness and continuity of state estimation calculation under abnormal operating conditions.

[0104] In the process of single state estimation calculation, for measurement sources that have passed the basic rule admission, a horizontal comparison is further made based on the current cross-section calculation deviation;

[0105] By comparing the deviations between PMU measurements and state estimation results Deviation between SCADA measurement and state estimation results And the deviation between TSCA measurement and state estimation results The value of the measurement source with the smallest real-time deviation is selected as the core input data for the current state estimation calculation of the measurement point.

[0106] (4) By converting the PMU phase angle reference and correcting the transformer connection group, the screened mixed measurement data are substituted into the weighted least squares model for fusion solution to obtain the optimal state section of the system.

[0107] (41) PMU phase angle measurement data preprocessing: Multiple verifications and filtering are performed on the aligned PMU phase angle measurements to ensure the consistency of phase angle input:

[0108] Node phase angle deviation verification: Calculate the bus phase angle deviation from different sources within the same compute node. , They were respectively in the second Within the nth node The time-marked section and the first The measured PMU voltage phase angle at each time scale section, if the deviation is less than the set threshold If the average of the phase angles from each source is taken as the input, then the abnormal measurement with the largest deviation will be removed if the threshold is exceeded.

[0109] Transmission distance phase angle constraint screening: Based on the positive correlation between phase angle difference and transmission distance in AC power grids, the selection is based on the farthest transmission distance L from node i to the reference node. i Set phase angle constraint range , This is the maximum allowable phase angle difference threshold; if the relative phase angle measurement of a node exceeds this range, it is judged as an abnormal measurement and is removed.

[0110] (42) PMU phase angle measurement reference conversion and correction: Convert the absolute phase angle based on BeiDou / GPS satellite timing into the relative phase angle required for state estimation:

[0111] Reference point phase offset calculation: The node equipped with the PMU is selected as the reference node for conventional state estimation, and the phase offset of the reference point relative to the satellite timing reference system is directly obtained. , The absolute phase angle measurement of the PMU at the reference node;

[0112] Global phase angle correction: based on phase angle offset A unified correction was performed on all PMU phase angle measurements across the entire network. The corrected phase angle value is... ; This is the original value of the absolute phase angle of the PMU.

[0113] Transformer connection group compensation: for star-delta ( Transformers with a delta (or delta) connection method are processed using software algorithms. The phase angle is measured to compensate for a fixed phase deviation (e.g., 30°) to ensure that the measurement data on both the high and low voltage sides of the transformer are consistent.

[0114] (43) Weighted least squares solution based on mixed measurement cross sections: Substitute the aligned, filtered and corrected mixed measurement data into the state estimation model for unified solution:

[0115] The hybrid measurement equation system is constructed as follows:

[0116]

[0117] In the formula: z is a hybrid measurement data vector, covering SCADA measurement, TSCA measurement and corrected PMU phasor measurement; x is the state vector to be solved, including node voltage magnitude and phase angle; It is a nonlinear measurement function; This is the multi-source measurement error vector;

[0118] Construct an objective function, and use the weighted least squares (WLS) method to minimize the objective function J(x). The objective function is as follows:

[0119]

[0120] In the formula: R is the error covariance matrix dynamically adjusted according to the quality of the measurement source, and its diagonal elements are the variances of each measurement. constitute.

[0121] Taking the partial derivative of the objective function J(x) with respect to x and setting the partial derivative to 0, we derive the closed-form solution for the optimal state estimate as follows:

[0122]

[0123] in, H is the optimal state estimation vector; H is the nonlinear measurement function h(x) in the initial reference state vector. The Jacobian matrix calculated at point ; The weight matrix is ​​determined by the real-time quality of multi-source measurements; This is the residual vector between the measured values ​​and the calculated values ​​based on the reference state, with the superscript T indicating transpose.

[0124] (44) Construction of comprehensive evaluation index system: Two core indicators, active power deviation and measurement pass rate, were selected to quantitatively verify the calculation accuracy and multi-source measurement adaptability of the proposed algorithm:

[0125] Active power deviation calculation:

[0126] All active power measurement nodes (including PMU, steady-state timescale, and SCADA active power measurement points) involved in the state estimation calculation were selected in the power grid, and the average active power deviation of the above-mentioned measurement points was statistically analyzed:

[0127]

[0128]

[0129] Where: ΔP i P represents the active power deviation at a single measuring point, and ΔP represents the average active power deviation across all measuring points involved in the calculation. meas,i Let P be the active power measurement value at the i-th measuring point. se,i The calculated state estimate value for the i-th measuring point is... This represents the total number of measurement points.

[0130] Measurement pass rate:

[0131] The evaluation algorithm's adaptability to different quality measurement data is defined and calculated as follows:

[0132]

[0133] Where: N qual N represents the number of qualified measuring points in this calculation. total The total number of measurement points used in the current calculation is represented by μ, which is the measurement pass rate index.

[0134] This embodiment uses power grid data from a certain region for testing, involving 12,724 substations, 27,539 transmission lines, and 30,020 generators, with a total of 62,036 effective calculation nodes. Currently, the platform has 101,000 PMU data points with qualified quality codes, of which 70,000 are used in the algorithm calculation presented in this paper; 21,000 TSCA measurement quality codes are qualified, and 13,000 TSCA measurement points are used in the calculation. A sliding time window W=[t0-ΔT, t0+ΔT] is constructed centered on the baseline measurement time t0, with the sliding time window ΔT=3s and the step size ΔS=0.1s. The matching degree of the candidate PMU timescale set is calculated and sorted. Figure 2 It is known that the average deviation of the optimal mixed measurement section is only 4.59. The system operates stably for a long period of 5 minutes. At each calculation section, the system performs real-time checks on polarity, difference, and balance. Thresholds for incorporating the PMU comprehensive evaluation index are set. Exit threshold Based on this threshold combination, the system can effectively balance the reliability and sufficiency of measurement data, automatically eliminating measurement sources that are unreliable for a long time or have excessive real-time deviations. The filtered and corrected mixed measurement data is input into a weighted least squares module, and the optimal state estimate is directly obtained through a closed-form solution. Table 1 shows a comparison analysis of the traditional state estimation and mixed measurement state estimation algorithms. Under normal operating conditions, the method described in this invention reduces the average active power deviation of the mixed measurement state estimation across all voltage levels by 9.296% compared to conventional state estimation.

[0135] Table 1:

[0136]

[0137] This embodiment also provides an adaptive robust state estimation system based on PMU and SCADA multi-source measurements, including:

[0138] Measurement section alignment module: It is used to align the sections of PMU measurement data, SCADA measurement data and TSCA measurement data and correct equipment-level offset through section matching degree analysis, sliding time window and power characteristic analysis, and generate consistent mixed measurement data.

[0139] The comprehensive quality evaluation module is used to construct a comprehensive quality evaluation system for multi-source measurements. Specifically, it performs quality analysis on multi-source measurement data by calculating the deviation between the measurement data and the state assessment value, and by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators for verification.

[0140] Adaptive adjustment module: Used to adaptively adjust the measurement data involved in the calculation according to the evaluation index, and remove abnormal and low-quality data;

[0141] Hybrid Measurement State Solution Module: This module uses PMU phase angle reference conversion and transformer connection group correction to substitute the selected hybrid measurement data into a weighted least squares model for fusion solution, thereby obtaining the optimal state profile of the system.

[0142] This embodiment also provides a computer device, the device including one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs, when executed by the processors, implement the steps of an adaptive robust state estimation method based on PMU and SCADA multi-source measurements as described above.

[0143] This embodiment also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the steps of an adaptive robust state estimation method based on PMU and SCADA multi-source measurements as described above.

[0144] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus (systems), computer devices, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] This invention is described with reference to a flowchart of a method according to embodiments of the invention. It should be understood that each step in the flowchart and combinations thereof can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 A device for a function specified in one or more processes.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 The function specified in one or more processes.

[0147] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 Steps of a specified function in one or more processes.

Claims

1. An adaptive robust state estimation method based on PMU and SCADA multi-source measurements, characterized in that, include: By analyzing the cross-sectional matching degree and the sliding time window and power characteristics, the cross-sections of PMU measurement data, SCADA measurement data and TSCA measurement data are aligned and the equipment-level offset is corrected, generating consistent mixed measurement data. Construct a comprehensive quality evaluation system for multi-source measurements. Specifically, this involves calculating the deviation between the measurement data and the state assessment value, and verifying the data by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators to conduct quality analysis on the multi-source measurement data. The measurement data used in the calculation are adaptively adjusted according to the evaluation indicators, and abnormal and low-quality data are removed. By converting the PMU phase angle reference and correcting the transformer connection group, the screened mixed measurement data is substituted into the weighted least squares model for fusion solution to obtain the optimal state section of the system.

2. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 1, characterized in that, The cross-sectional matching degree analysis and the sliding time window and power characteristic analysis specifically include: Using the time of SCADA measurement data as the reference measurement time, the cross-sectional matching degree between SCADA measurement data and PMU measurement data and the cross-sectional matching degree with TSCA measurement data are calculated respectively. The cross-sectional matching degree is calculated based on the weight of each measurement and the deviation between each measurement data. Based on the SCADA baseline measurement time, a sliding time window is constructed. The matching degree between the k-th time scale section of the PMU measurement data and the TSCA measurement data under this time window and the SCADA section is calculated respectively. The matching degree of all sections of the PMU measurement data and the TSCA measurement data is sorted in descending order, and the section corresponding to the maximum matching degree is selected as the optimal time scale measurement section. The power characteristic analysis identifies intervals with consistent power trends by constructing a power fluctuation rate index and a power trend consistency analysis index. Within the interval, it matches the characteristic moments of SCADA measurement data and PMU measurement data, calculates the time offset, and corrects the time offset of the PMU measurement.

3. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 1, characterized in that, The process involves calculating the deviation between the measured data and the state assessment value, and then verifying this deviation by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators. Specifically, this process performs quality analysis on the multi-source measured data as follows: In the mixed measurement calculation scenario, based on the current and historical state estimation calculation results, the deviations between the PMU measurement, TSCA measurement and SCADA measurement and the state estimation results are calculated respectively. For each measurement point, count the number of times each measurement source's deviation is greater than or equal to a set threshold and less than a set threshold within a time window, and calculate the comprehensive reliability index of each measurement source based on the count. Within a single calculation section, real-time closed-loop verification is performed on the aligned measurement data. This real-time closed-loop verification includes three types: measurement polarity quality verification, where if the polarity of the timescale measurement is opposite to that of the SCADA measurement, and the SCADA imbalance of the corresponding node is less than a set imbalance threshold, then the timescale measurement is inverted; measurement difference quality verification, where if the absolute difference between the timescale measurement and the SCADA measurement is greater than a set threshold, and the percentage of the relative difference is greater than a percentage threshold, then the measurement point is determined to be an abnormal measurement point and is removed; and measurement balance quality verification, used to perform active power balance verification and reactive power balance verification on the line, transformer windings, and calculation nodes according to physical topology constraints.

4. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 3, characterized in that, The measurement balance quality verification specifically includes: Line balance verification: Determine whether the absolute value of the difference between the active power at the beginning and end of the line is less than the line balance threshold. Main transformer balance verification: Determine whether the algebraic sum of active power on each side of the main transformer is less than the main transformer balance threshold; Node balance check: Determine whether the algebraic sum of the power of all associated branches on the same computing node is less than the node balance threshold.

5. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 1, characterized in that, The process of adaptively adjusting the measurement data used in the calculation based on the evaluation indicators and removing abnormal and low-quality data specifically involves: Using the comprehensive reliability index of each measurement source within the time window, admission and exit thresholds are set for PMU measurements and TSCA measurements respectively. Measurement sources whose comprehensive reliability index reaches the admission threshold are set to participate in the current calculation and whose reliability meets the calculation requirements; measurement sources whose comprehensive reliability index is lower than the exit threshold are set to exit the current calculation and are judged as having insufficient reliability. Those whose reliability meets the requirements are selected as high-reliability measurement source candidate pool. Within the selected measurement candidate pool, when there are multiple measurement sources at the same measurement point, a horizontal comparison is made based on the current cross-sectional calculation deviation during a single state estimation calculation. The measurement source with the smallest real-time deviation value is selected as the input data for that measurement point in the current state estimation calculation. Historical comprehensive reliability indicators are used to eliminate measurement sources with long-term unstable performance, and real-time quality assessment indicators are used to deal with power grid changes or short-term communication anomalies.

6. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 1, characterized in that, The process involves converting the PMU phase angle reference and correcting the transformer connection group, then substituting the selected mixed measurement data into a weighted least squares model for fusion solution to obtain the optimal state profile of the system. Specifically: After alignment, the phase angle measurements of the PMU are checked for node phase angle deviation and screened for transmission distance phase angle constraints to eliminate abnormal phase angle measurements. The node with the PMU installed is selected as the reference node for state estimation. The phase angle offset of the reference node relative to the satellite timing reference system is obtained. Based on the phase angle offset, all PMU phase angle measurements are uniformly corrected, and the transformer wiring is compensated by group. Substitute the aligned, filtered and corrected mixed measurement data into the weighted least squares model to construct an objective function with the goal of minimizing the weighted sum of squared measurement errors, and solve for the optimal state estimation vector. Active power deviation and measurement pass rate are selected as evaluation indicators to quantitatively evaluate the state estimation results.

7. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 6, characterized in that, The node phase angle deviation verification is specifically as follows: calculate the bus phase angle deviation from different sources within the same computing node; if the deviation is less than a set threshold, take the average of the phase angles from each source as input; if the deviation exceeds the set threshold, remove the abnormal measurement with the largest deviation. The transmission distance phase angle constraint screening is specifically as follows: a phase angle constraint range is set according to the transmission distance from each node to the reference node. If the relative phase angle measurement of a node exceeds the constraint range, it is determined to be an abnormal measurement and is removed.

8. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 6, characterized in that, The process involves substituting the aligned, filtered, and corrected mixed measurement data into a weighted least squares model to construct an objective function that minimizes the weighted sum of squared measurement errors. Solving this function yields the optimal state estimation vector. Specifically: The hybrid measurement equation system is constructed as follows: ； Where z is a hybrid measurement data vector, including SCADA measurements, TSCA measurements, and corrected PMU phasor measurements; x is the state vector to be solved, including node voltage magnitude and phase angle; It is a nonlinear measurement function; This is the multi-source measurement error vector; Using the weighted least squares method to minimize the objective function J(x), the objective function is constructed as follows: ； Where R is the error covariance matrix dynamically adjusted according to the quality of the measurement source, and its diagonal elements are the variances of each measurement data value. constitute; Taking the partial derivative of the objective function J(x) with respect to x and setting the partial derivative to 0, we derive the closed-form solution for the optimal state estimate as follows: ； in, H is the optimal state estimation vector; H is the nonlinear measurement function h(x) in the initial reference state vector. The Jacobian matrix calculated at point ; The weight matrix is ​​determined by the real-time quality of multi-source measurements; This is the residual vector between the measured values ​​and the calculated values ​​based on the reference state, with the superscript T indicating transpose.

9. The adaptive robust state estimation method based on PMU and SCADA multi-source measurements according to claim 1, characterized in that, The PMU measurement data includes node voltage amplitude, node voltage phase angle, branch current amplitude, branch current phase angle, and node active power and node reactive power calculated from voltage and current; the SCADA measurement data includes node voltage amplitude, node injected active power, node injected reactive power, inter-node active power, and inter-node reactive power.

10. An adaptive robust state estimation system based on PMU and SCADA multi-source measurements, used to implement the method described in any one of claims 1-9, characterized in that, include: Measurement section alignment module: It is used to align the sections of PMU measurement data, SCADA measurement data and TSCA measurement data and correct equipment-level offset through section matching degree analysis, sliding time window and power characteristic analysis, and generate consistent mixed measurement data. The comprehensive quality evaluation module is used to construct a comprehensive quality evaluation system for multi-source measurements. Specifically, it performs quality analysis on multi-source measurement data by calculating the deviation between the measurement data and the state assessment value, and by combining historical comprehensive reliability assessment indicators with real-time quality assessment indicators for verification. Adaptive adjustment module: Used to adaptively adjust the measurement data involved in the calculation according to the evaluation index, and remove abnormal and low-quality data; Hybrid Measurement State Solution Module: This module uses PMU phase angle reference conversion and transformer connection group correction to substitute the selected hybrid measurement data into a weighted least squares model for fusion solution, thereby obtaining the optimal state profile of the system.

Images