A key indicator-based power grid safe operation feature extraction method and system

Abstract

A power grid safe operation feature extraction method and system based on key indicators, characterized by real-time collection of power grid operation data to calculate power grid indicators, including line overload safety margin, main transformer overload safety margin, bus voltage safety level, section safety margin, short-circuit current, power grid power supply capacity, load prediction error, new energy prediction error and average frequency error; track the calculation results of the indicators, determine the outliers in the calculation results, and extract the abnormal stage of the indicators according to the time corresponding to the outliers; taking the power grid load data, power grid maintenance plan and power grid generation plan as independent variables, and each indicator as dependent variable, based on the relationship between independent variables and dependent variables in the abnormal stage, the sensitivity function of each indicator is constructed respectively; solve the sensitivity function to obtain the decisive independent variable in the current abnormal stage, and generate the power grid safe operation feature based on the decisive independent variable.

Description

Technical Field

[0001] This invention relates to the field of power system automated dispatching, and more specifically, to a method and system for extracting power grid safety operation features based on key indicators. Background Technology

[0002] As the proportion of renewable energy sources in the power grid gradually increases, the traditional power system is transforming and upgrading into a new type of power system dominated by renewable energy. During this process, the material and technological foundations related to primary energy characteristics, power source layout and function, network scale and form, load structure and characteristics, grid balance mode, and power system technology are undergoing profound changes, posing new problems and challenges to the safe and stable operation of the power system.

[0003] Compared to the relatively predictable uncertainties in traditional power systems, the uncertainties in power systems under energy transition have increased significantly. Safety and stability issues are intertwined with balance regulation and renewable energy consumption; frequency, voltage, and power angle, among other safety and stability factors, influence each other; and the contradiction between grid construction and excessive short-circuit currents has become more prominent. These factors pose significant challenges to grid operation and control, and the complexity and difficulty of grid dispatch and operation are increasing daily. There is an urgent need to break through the limitations of single, fragmented evaluation indicators and to comprehensively consider the planning and operation of the power grid from multiple levels and perspectives. Further in-depth analysis is needed to uncover the grid defects, operational patterns, and evolutionary trends hidden behind the indicators, assisting grid dispatch and operation personnel in quickly and accurately grasping the overall operating status of the grid, identifying weak links, optimizing control measures, and improving the safety operation and risk prevention and control level of the large power grid.

[0004] Mining and extracting power grid safety operation characteristics by combining historical indicator data with power grid operation data is one of the effective means to improve the level of power grid operation safety and risk prevention and control. However, among the current mainstream technologies, there is no specific research on techniques that use historical indicator data to reflect power grid characteristics. This makes the technology for mining and extracting power grid safety operation characteristics still limited. For example, the analysis of power grid anomalies is not sufficient, and the exploration of the ways in which various indicators affect power grid operation characteristics is not thorough enough. These factors ultimately lead to inaccurate results in the mining and extraction of power grid safety operation characteristics, making it difficult to effectively improve the level of power grid operation safety and risk prevention and control.

[0005] To address the aforementioned issues, there is an urgent need for a method and system for extracting power grid safety operation features based on key indicators. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for extracting power grid safety operation characteristics based on key indicators. By extracting and analyzing the sensitivity influencing factors of power grid anomaly indicators, the safe operation characteristics of the power grid are extracted and visualized. In addition, based on the characteristics of the sensitivity influencing factors, a report with recommendations for power grid operation safety and risk prevention and control is provided.

[0007] The present invention adopts the following technical solution.

[0008] The first aspect of this invention relates to a method for extracting power grid safety operation features based on key indicators. The method includes the following steps: real-time acquisition of power grid operation data to calculate power grid indicators, including line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, short-circuit current, power grid supply capacity, load forecasting error, new energy forecasting error, and average frequency error; tracking the calculation results of the indicators and identifying outliers in the calculation results, and extracting the abnormal stage of the indicators based on the time corresponding to the outliers; using power grid load data, power grid maintenance plan, and power grid generation plan as independent variables, and each indicator as a dependent variable, constructing a sensitivity function for each indicator based on the relationship between the independent and dependent variables under the abnormal stage; solving the sensitivity function to obtain the independent variable that plays a decisive role under the current abnormal stage, and extracting power grid safety operation features based on the independent variable that plays a decisive role.

[0009] Preferred power supply capacity indicators for

[0010] in, For the real-time load of the power grid, The real-time power supply to the power grid; Load forecasting error for

[0011] in, This represents the predicted load value of the power grid; New energy prediction error is

[0012] in, To provide real-time power output for new energy sources in the power grid. Contribute to the prediction of new energy sources in the power grid; Average frequency error for

[0013] in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

[0014] Preferably, the calculation results of the tracking indicators and the determination of outliers in the calculation results also include: calculating the real-time value of each indicator at a preset time interval, and constructing a curve of each indicator changing over time based on the real-time value and the time interval; obtaining the real-time values ​​of continuous indicators under the test period, and extracting outliers in the real-time values ​​according to the distribution of the real-time values.

[0015] Preferably, the extraction of outliers from real-time values ​​based on the distribution of real-time values ​​further includes: arranging real-time values ​​in order according to their magnitude to obtain an indicator array; extracting the middle portion of the indicators in the indicator array according to the extraction rate to obtain the first and last indicators among the partial indicators; obtaining the normal value range of the real-time values ​​based on the weight difference between the first and last indicators; and determining any indicator in the indicator array that exceeds the normal value range as an outlier.

[0016] Preferably, the abnormal stage of extracting indicators based on the time corresponding to the abnormal value also includes: for a single abnormal value or multiple consecutive abnormal values, extracting all real-time values ​​within a preset time range before and after the abnormal value, thereby constituting an abnormal stage.

[0017] Preferably, using grid load data, grid maintenance plans, and grid generation plans as independent variables further includes: the grid load data at least includes the real-time load of the grid corresponding to the current abnormal phase. The power grid maintenance plan should at least include the maintenance time, scope, and outage capacity corresponding to the current abnormal phase of the power grid. The power grid generation plan should include at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which should consist at least of the power grid's predicted load values. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply to the power grid. Provide evidence; extract the values ​​of each independent variable at different times under the current abnormal stage, and construct a bar chart.

[0018] Preferably, constructing the sensitivity function for each indicator based on the relationship between the independent and dependent variables under the abnormal stage further includes: constructing a fitting formula between the current indicator and the current independent variable based on the calculation formula of the current indicator, so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage; and taking the derivative of the current independent variable based on the fitting formula to obtain the sensitivity function based on the current independent variable and the current dependent variable.

[0019] Preferably, solving the sensitivity function to obtain the independent variable that plays a decisive role in the current abnormal stage further includes: extracting the highest-order coefficient of the current independent variable in the sensitivity function as the determination coefficient of the current independent variable in the current abnormal stage; comparing the determination coefficients of each independent variable of the current dependent variable, and selecting one or more independent variables whose determination coefficients exceed a preset value as the independent variables that play a decisive role in the current abnormal stage.

[0020] Preferably, generating a power grid safe operation characteristic report based on the decisive independent variable further includes: if the decisive independent variable in the current abnormal phase is the power grid maintenance plan, then providing improvement suggestions for the power grid maintenance plan based on the outlier; if the decisive independent variable in the current abnormal phase is the power grid load data or the power grid generation plan, then proposing improvement suggestions for the power grid equipment stability limit; and summarizing the improvement suggestions into power grid safe operation characteristics.

[0021] Preferably, a power grid safety operation feature report is generated based on the extracted power grid safety operation features.

[0022] The second aspect of this invention relates to a power grid safety operation feature extraction system based on key indicators. The system includes an indicator acquisition module, an anomaly extraction module, a function construction module, and a report generation module. The indicator acquisition module is used to collect power grid operation data in real time to calculate power grid indicators, including line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, short-circuit current, power supply capacity, load forecasting error, new energy forecasting error, and average frequency error. The anomaly extraction module is used to track the calculation results of the indicators, identify outliers in the calculation results, and extract the abnormal stage of the indicators based on the time corresponding to the outliers. The function construction module is used to construct a sensitivity function for each indicator based on the relationship between the independent and dependent variables under the abnormal stage, using power grid load data, power grid maintenance plans, and power grid generation plans as independent variables and each indicator as the dependent variable. The report generation module is used to solve the sensitivity function to obtain the independent variables that play a decisive role in the current abnormal stage, and extract power grid safety operation features based on the independent variables that play a decisive role.

[0023] Preferably, the power grid indicators calculated by the indicator acquisition module are: Power supply capacity indicators for

[0024] in, For the real-time load of the power grid, The real-time power supply to the power grid; Load forecasting error for

[0025] in, This represents the predicted load value of the power grid; New energy prediction error is

[0026] in, To provide real-time power output for new energy sources in the power grid. Contribute to the prediction of new energy sources in the power grid; Average frequency error for

[0027] in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

[0028] Preferably, the anomaly extraction module calculates the real-time value of each indicator at a preset time interval, and constructs a curve of each indicator changing over time based on the real-time value and the time interval; furthermore, the anomaly extraction module obtains the real-time values ​​of continuous indicators during the test period, and extracts outliers in the real-time values ​​according to the distribution of the real-time values.

[0029] Preferably, the anomaly extraction module is further configured to: arrange real-time values ​​in order according to their magnitude and obtain an indicator array; extract the middle portion of the indicators in the indicator array according to the extraction rate to obtain the first and last indicators among the partial indicators; obtain the normal value range of the real-time values ​​based on the weight difference between the first and last indicators; and determine the abnormal value of any indicator in the indicator array if it exceeds the normal value range.

[0030] Preferably, for a single outlier or multiple consecutive outliers, the anomaly extraction module extracts all real-time values ​​within a preset time range before and after the outlier, thus constituting an outlier phase.

[0031] Preferably, the function construction module is also used to process grid load data, grid maintenance plans, and grid generation plans; and the grid load data at least includes the real-time load of the grid corresponding to the current abnormal phase. The power grid maintenance plan should at least include the maintenance time, scope, and outage capacity corresponding to the current abnormal phase of the power grid. The power grid generation plan should include at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which should consist at least of the power grid's predicted load values. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply to the power grid. Provide evidence; the function building module extracts the values ​​of each independent variable at different times under the current abnormal stage and constructs a bar chart.

[0032] Preferably, the function construction module is also used to: construct a fitting formula between the current indicator and the current independent variable based on the calculation formula of the current indicator, so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage; and differentiate the current independent variable based on the fitting formula to obtain a sensitivity function based on the current independent variable and the current dependent variable.

[0033] Preferably, the report generation module is used to: extract the highest-order coefficient of the current independent variable in the sensitivity function as the determinant coefficient of the current independent variable under the current abnormal stage; compare the determinant coefficients of each independent variable of the current dependent variable, and select one or more independent variables whose determinant coefficients exceed a preset value as the independent variables that play a decisive role under the current abnormal stage.

[0034] Preferably, the report generation module is also used to determine: if the independent variable that plays a decisive role in the current abnormal phase is the power grid maintenance plan, then provide improvement suggestions for the power grid maintenance plan based on the outlier; if the independent variables that play a decisive role in the current abnormal phase are power grid load data and power grid generation plan, then propose improvement suggestions for the stability limit of power grid equipment; and the report generation module summarizes the improvement suggestions into power grid safe operation characteristics.

[0035] Preferably, the report generation module generates a power grid safety operation feature report based on the extracted power grid safety operation features.

[0036] The beneficial effects of this invention are that, compared with the prior art, the method and system for extracting power grid safety operation characteristics based on key indicators in this invention extracts and analyzes the sensitivity influencing factors of power grid anomaly indicators, extracts the safety operation characteristics of the power grid and achieves visualization display, and provides a report with suggestions for power grid operation safety and risk prevention and control based on the characteristics of the sensitivity influencing factors. This invention is effective and reliable, and by analyzing power grid anomaly data, it fully explores the operating characteristics of the power grid under abnormal conditions and provides improvement suggestions.

[0037] The beneficial effects of this application also include: 1. This invention is based on key indicators to realize the technology of mining and extracting historical data power grid operation characteristics, which is used to assist power grid dispatchers in mastering the laws of safe and stable power grid operation, guiding the arrangement of operation modes of complex power systems with strong uncertainty, guiding the adjustment of steady-state operation modes of the power grid, forming a closed-loop dispatch mode from situational awareness to intelligent control, assisting dispatchers in deeply perceiving the power grid operation status, identifying weak links, optimizing control measures, and further improving the level of safe operation and risk prevention and control of large power grids.

[0038] 2. The method of this invention uses index ranking to statistically analyze the distribution of continuous real-time indicators and analyzes outliers in the power grid based on this method. It not only constructs new indicators but also establishes a method for judging outliers, improving the ability to analyze and extract anomalies in the power grid and ensuring that potential faults and hidden safety risks in the power grid can be effectively identified through various means.

[0039] 3. The method of this invention targets only abnormal states of the power grid and conducts specific sensitivity analyses of various factors. The analysis method accurately obtains the degree of influence of multiple factors on abnormal power grid operation and proposes improvement suggestions for the decisive influence of different factors. This assists power grid personnel in operation and maintenance, dispatching, and other related work in automatically obtaining methods to improve the safety performance of the power grid, and to a certain extent provides guidance for dispatching, operation and maintenance, and other related work. Attached Figure Description

[0040] Figure 1 This is a schematic diagram illustrating the steps of a method for extracting features of power grid safe operation based on key indicators according to the present invention. Figure 2 This is a schematic diagram of the module architecture of a power grid safety operation feature extraction system based on key indicators according to the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this invention are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments not described in this invention obtained by those skilled in the art based on the embodiments described in this invention without creative effort should fall within the protection scope of this invention.

[0042] Figure 1 This is a schematic diagram illustrating the steps of a method for extracting features for safe power grid operation based on key indicators, as described in this invention. Figure 1 As shown, the first aspect of the present invention relates to a method for extracting features of power grid safe operation based on key indicators, the method comprising steps 1 to 4.

[0043] Step 1: Collect real-time power grid operation data to calculate power grid indicators, including line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, short-circuit current, power supply capacity of the power grid, load forecasting error, new energy forecasting error, and average frequency error.

[0044] In one embodiment of the present invention, based on massive historical data of key indicators, all data of a certain indicator within a certain time period are obtained. Taking the cross-sectional safety margin indicator as an example, the method can obtain the indicator result data for a certain time period and analyze the distribution of its cross-sectional safety margin, power flow value, and cross-sectional limit data. Therefore, in this embodiment, the indicator can be cross-sectional safety margin, power flow value, and cross-sectional limit data.

[0045] In another embodiment, considering the continuously increasing pressure on power supply, the new installed capacity of conventional power sources has consistently lagged behind the load growth rate, and the overall supply and demand situation of the power grid has reversed from "supply exceeding demand" to "tight balance across the entire grid with localized gaps." Furthermore, the random fluctuations in renewable energy output further increase the pressure on the power grid to ensure supply. Therefore, it is necessary to strengthen the centralized and effective monitoring of relevant indicators.

[0046] In this context, the indicators proposed in this invention may include various indicators such as main transformer overload safety margin, line overload safety margin, N-1 main transformer overload safety margin, N-1 line overload safety margin, spinning reserve, etc.

[0047] In fact, this invention does not limit the content of power grid indicators, but rather selects important indicators as the basis for extracting power grid safety operation characteristics based on actual conditions. Indicators such as line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, and short-circuit current can be calculated using existing technologies or obtained from real-time calculation results through connection to existing business platforms in the power system. These indicators are not necessarily overall indicators of a local power grid; they can also be corresponding data from a specific line or device within the power grid. Furthermore, in the subsequent calculation of sensitivity indicators, the sensitivity of the impact of different factors on these indicators under different abnormal periods can be calculated based on the specific power grid location corresponding to each indicator.

[0048] In one embodiment of the present invention, several new indicators are proposed to meet the requirements of power grid security and supply. These indicators are the power supply capacity of the power grid, load forecasting error, new energy forecasting error, and average frequency error.

[0049] Preferred power supply capacity indicators for

[0050] in, For the real-time load of the power grid, This refers to the real-time power supply to the power grid.

[0051] Based on the power supply capacity index values, the method can classify the power grid's power supply capacity into different levels. For example, a value greater than 2% indicates that the current power supply capacity is normal, and this index can be represented in green on the visualization interface. If the value is between 0 and 2%, it is an alarm state, represented in orange; if the value is less than 0, it is an emergency state, represented in red.

[0052] Load forecasting error and renewable energy forecasting error are used to assess the degree of deviation between day-ahead forecasts and actual operating values. Given insufficient grid balance margins, these two indicators effectively reflect the impact of forecasting errors on spinning reserves. Larger forecasting errors will relatively reduce spinning reserves, further impacting the safe operation of the power grid.

[0053] Load forecasting error for

[0054] in, This represents the predicted load value of the power grid.

[0055] Similarly, specific value ranges for indicators can be set to use different colors to represent normal, alarm, or emergency states. When the load forecast error is greater than 0, it is considered normal; when the value is between -3 and 0, it is considered an alarm; and when the value is less than -3, it is considered an emergency.

[0056] New energy prediction error is

[0057] in, To provide real-time power output for new energy sources in the power grid. It contributes to the prediction of new energy sources in the power grid.

[0058] When the new energy prediction error is greater than zero, the indicator is normal; when the value is between -15 and 0, the indicator is alarming; and when it is less than -15, it is an emergency.

[0059] Average frequency error for

[0060] in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

[0061] In one embodiment of the present invention, the time interval between two adjacent sampling times can be 5 seconds, and the value of N can be 60. Therefore, each average frequency offset index can be used to calculate the data content of consecutive frequency sampling times with a period of 5 minutes, and the calculation result can be obtained.

[0062] In this invention, the average frequency error is considered normal when it is greater than -0.033Hz, alarm when it is between -0.1 and -0.033Hz, and emergency when it is less than -0.1Hz.

[0063] After obtaining the calculation method of the above power grid indicators, the present invention can repeatedly and in real time calculate the values ​​of the power grid indicators at preset time intervals.

[0064] Step 2: Track the calculation results of the indicators, identify outliers in the calculation results, and extract the abnormal phase of the indicators based on the time corresponding to the outliers.

[0065] Preferably, tracking the calculation results of the indicators and determining outliers in the calculation results further includes: calculating the real-time value of each indicator at a preset time interval, and constructing a curve of each indicator changing over time based on the real-time value and the time interval; obtaining the real-time values ​​of continuous indicators during the test period, and extracting outliers in the real-time values ​​according to the distribution of the real-time values.

[0066] It is understood that in this invention, a large amount of historical data of the indicators under the test period can be obtained first. For each preset time interval under the test period, the real-time value of the indicator can be calculated, which is the repeated, real-time calculation process of the power grid indicator value in step 1.

[0067] After obtaining all the real-time values, the method can construct a curve chart of these indicators in chronological order. The conversion of the data into a graph can be achieved using existing technologies, and the curve chart can be displayed in a visualization interface in a way that corresponds to the real-time data.

[0068] The method of the present invention also supports the analysis of the distribution of real-time values, and based on the analysis of the distribution, the normal value range and the abnormal value range of real-time values ​​are obtained, thereby further obtaining the abnormal values ​​in the real-time values.

[0069] Preferably, the extraction of outliers from real-time values ​​based on the distribution of real-time values ​​further includes: arranging real-time values ​​in order according to their magnitude to obtain an indicator array; extracting the middle portion of the indicators in the indicator array according to the extraction rate to obtain the first and last indicators among the partial indicators; obtaining the normal value range of the real-time values ​​based on the weight difference between the first and last indicators; and determining any indicator in the indicator array that exceeds the normal value range as an outlier.

[0070] It is understandable that, before outliers in the real-time data in this invention, multiple real-time values ​​need to be arranged sequentially according to their magnitude; both sequential and reverse order are allowed. The method can pre-set an extraction rate, such as 50% as the core indicator value range, then extract 50% of the data points located at the center of the sequence. At this point, the first 25% and the last 25% of indicators can be excluded, and the remaining indicators are those within the core value range. The first and last indicators in this portion can then be defined as the value range of the core indicator.

[0071] After obtaining the value range of the core indicators, this invention also supports appropriate expansion of the core indicators to obtain the maximum normal value range for real-time values. In one embodiment of this invention, the normal value range is obtained based on the weight difference between the first and last indicators.

[0072] Specifically, the first and last indicators can be assigned weights. If a weighted average is used, subtracting the last indicator from the first gives the limit value in the forward direction of the normal range. Conversely, subtracting the first indicator from the last gives the limit value in the backward direction of the normal range. By determining the limits at both ends of the normal range, the method can obtain an accurate result for the normal range. If an indicator's value exceeds this range, it can be identified as an outlier.

[0073] In one embodiment of the present invention, the weights can be 1.5 and 2.5. If the first indicator is subtracted from the last indicator, the weight of the first indicator is 2.5 and the weight of the last indicator is 1.5. If the last indicator is subtracted from the first indicator, the weight of the first indicator is 1.5 and the weight of the last indicator is 2.5.

[0074] Using the above method, this invention calculates and obtains data on all abnormal values ​​during the period to be detected. Considering the shortcomings of existing technologies, this invention can determine the root cause of actual faults or safety risks based on this abnormal data, and based on this fundamental cause, attempts to fundamentally solve, overcome, or mitigate the safety risks of the power grid.

[0075] Therefore, the subsequent analysis methods of this invention are based on outlier data, or are centered around outlier data.

[0076] Preferably, the abnormal stage of extracting indicators based on the time corresponding to the abnormal value also includes: for a single abnormal value or multiple consecutive abnormal values, extracting all real-time values ​​within a preset time range before and after the abnormal value, thereby constituting an abnormal stage.

[0077] In this invention, a single outlier or multiple consecutive outliers can be identified. For these outliers, the method can employ a certain approach to extract the outlier data, as well as the adjacent data in the time dimension surrounding the outlier.

[0078] Therefore, the method of the present invention can consider first setting a time range and extracting data within the time range before and after the outlier. Alternatively, the method can also consider starting with the magnitude of the outlier; if the difference between the surrounding data values ​​and the outlier is relatively small within a unit time interval, then the outlier can be included in the adjacent data.

[0079] In summary, the method aims to obtain outliers and sufficient data around the outliers through various means, thereby fitting a curve segment of a certain indicator changing over time, and fitting a formula for the change of the indicator based on the changing pattern of the curve segment.

[0080] Step 3: Using grid load data, grid maintenance plan, and grid generation plan as independent variables, and each indicator as a dependent variable, construct the sensitivity function for each indicator based on the relationship between the independent and dependent variables during the abnormal phase.

[0081] Preferably, using grid load data, grid maintenance plans, and grid generation plans as independent variables further includes: the grid load data at least includes the real-time load of the grid corresponding to the current abnormal phase. The power grid maintenance plan should at least include the maintenance time, scope, and outage capacity corresponding to the current abnormal phase of the power grid. The power grid generation plan should include at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which should consist at least of the power grid's predicted load values. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply to the power grid. Provide evidence; extract the values ​​of each independent variable at different times under the current abnormal stage, and construct a bar chart.

[0082] In this invention, power grid load data, power grid maintenance plans, and power grid generation plans are used as independent variables to construct functions for each indicator. Specifically, multiple indicators actually contain the content of the independent variables mentioned here, or are indirectly influenced by the independent variables mentioned here.

[0083] For example, the real-time load of the power grid Forecast load values ​​of the power grid Forecasted power output of new energy sources in the power grid These exist in the three indicators mentioned above. Power grid maintenance plans can actually cause a partial outage of power grid capacity, which leads to changes in the grid's output capacity. Therefore, the outage capacity can also be considered as... This is equivalent to the change in power grid output capacity at the moment the maintenance plan occurs.

[0084] In addition, the average frequency shift is actually caused by the imbalance of active power. Each independent variable in this invention indirectly affects the balance of active power in a specific way. Therefore, this invention can also indirectly construct a correlation formula between the independent variables and the average frequency shift.

[0085] Preferably, constructing the sensitivity function for each indicator based on the relationship between the independent and dependent variables under the abnormal stage further includes: constructing a fitting formula between the current indicator and the current independent variable based on the calculation formula of the current indicator, so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage; and taking the derivative of the current independent variable based on the fitting formula to obtain the sensitivity function based on the current independent variable and the current dependent variable.

[0086] It is readily apparent that this invention can substitute the various independent variables into the calculation formula of the current indicator, and adjust the formula using methods found in existing technologies. For example, if one wants to use... As the independent variable, If the dependent variable is the real-time power supply, then the method could consider using the real-time power supply. Set it to a constant. Alternatively, consider the real-time power supply. and The inherent connection will Transformed into based on The function, and Substitute power supply capacity indicators In the calculation formula, thus obtaining and The relationship between them. After this transformation, Possibly A multidimensional function. However, regardless of the method, it can obtain the fitting relationship between the independent and dependent variables.

[0087] Based on the steps described above, the method can extract the values ​​of independent and dependent variables during the abnormal phase and construct a polynomial based on the highest dimension obtained from the above method. Then, according to the values ​​of the fitted curve, the coefficients of each polynomial are fitted, thus obtaining a definite fitting formula. At this point, the fitting formula and the fitted curve should correspond to each other.

[0088] Similarly, other independent or dependent variables can also be reasonably constructed into fitting formulas using the methods mentioned above, thereby achieving the sensitivity analysis of the above formulas.

[0089] In another embodiment, if the average frequency error If the calculation formula does not directly include a certain independent variable, then the method can suggest a relationship between a certain independent variable and the real-time frequency. The relationship between them. When load fluctuations, maintenance plan adjustments, or power generation plan adjustments occur in the power grid, corresponding adjustments to the real-time frequency will follow. Therefore, the method can adopt a reasonable approach to first construct the relationship between the independent variables and the real-time frequency. The correlation between them is then substituted into the formula for calculating the average frequency error.

[0090] After obtaining the fitting formula, this invention can differentiate the fitting formula to obtain the characteristics of how the dependent variable changes with the unit change of the independent variable. Therefore, the method can solve for the coefficient before the independent variable term in the differentiated formula; this coefficient is the output result of the sensitivity analysis.

[0091] Preferably, the sensitivity function is solved to obtain the independent variable that plays a decisive role in the current abnormal stage, and the highest-order coefficient of the current independent variable in the sensitivity function is extracted as the determination coefficient of the current independent variable in the current abnormal stage; the determination coefficients of each independent variable of the current dependent variable are compared, and one or more independent variables whose determination coefficients exceed a preset value are selected as the independent variables that play a decisive role in the current abnormal stage.

[0092] In one embodiment of the present invention, the method can obtain the coefficient of determination after solving the sensitivity function. The larger the coefficient of determination, the more significant the influence of the independent variable on the dependent variable is, and the more decisive the value of the dependent variable is.

[0093] Step 4: Solve the sensitivity function to obtain the independent variables that play a decisive role in the current abnormal stage, and extract the characteristics of power grid safe operation based on the independent variables that play a decisive role.

[0094] Preferably, generating power grid safe operation characteristics based on the decisive independent variables further includes: if the decisive independent variable in the current abnormal phase is the power grid maintenance plan, then providing improvement suggestions for the power grid maintenance plan based on the outlier; if the decisive independent variables in the current abnormal phase are power grid load data and power grid generation plan, then proposing improvement suggestions for the stability limits of power grid equipment; and summarizing the improvement suggestions into power grid safe operation characteristics.

[0095] Preferably, a power grid safety operation feature report is generated based on the extracted power grid safety operation features.

[0096] In this invention, one or more independent variables can be identified as having a decisive role. This identification can be based on expert experience or a pre-established method of determination. After obtaining the decisive independent variable, the method can derive corresponding improvement suggestions based on the content of that independent variable and its inherent properties.

[0097] The improvement suggestions here can be directed at power grid equipment stability limits or power grid maintenance plans. Specifically, the method can pre-set calculation formulas based on the type of the affected dependent variable, the type of the determining independent variable, the values ​​of the independent and dependent variables, and the extent of outliers, and provide risk warnings and improvement strategies at different levels.

[0098] For example, if the independent variable is the power grid generation plan and the dependent variable is the power grid supply capacity, and the outlier exceeds the limit significantly, then the proposed improvement suggestions for the stability limits of power grid equipment can be relatively larger. In practical implementation, the method can set a parameter for the goal of improving the stability limits of power grid equipment, and this parameter can be divided into multiple value ranges. By considering the types of independent and dependent variables and the range of outliers, the parameter value under the current situation can be determined. When the parameter value falls into a lower level, the improvement for the stability limits of power grid equipment can be relatively smaller, while if the parameter value falls into a higher level, the improvement can be relatively larger. The method can pre-provide template fields for reports at multiple different levels, similar to a health check report for power equipment. The corresponding template fields are automatically loaded based on the parameter value to generate the report.

[0099] Figure 2 This is a schematic diagram of the module architecture of a power grid safety operation feature extraction system based on key indicators according to the present invention. Figure 2 As shown, in a second aspect, this invention relates to a power grid safety operation feature extraction system based on key indicators. The system includes an indicator acquisition module, an anomaly extraction module, a function construction module, and a report generation module. The indicator acquisition module is used to collect power grid operation data in real time to calculate power grid indicators, including power supply capacity, load forecasting error, new energy forecasting error, and average frequency error. The anomaly extraction module is used to track the calculation results of the indicators, identify outliers in the calculation results, and extract the abnormal phase of the indicator based on the time corresponding to the outlier. The function construction module is used to construct a sensitivity function for each indicator based on the relationship between the independent and dependent variables under the abnormal phase, using power grid load data, power grid maintenance plans, and power grid generation plans as independent variables and each indicator as a dependent variable. The report generation module is used to solve the sensitivity function to obtain the independent variable that plays a decisive role in the current abnormal phase, and extract power grid safety operation features based on the independent variable that plays a decisive role.

[0100] Preferably, the power grid indicators calculated by the indicator acquisition module are: Power supply capacity indicators for

[0101] in, For the real-time load of the power grid, The real-time power supply to the power grid; Load forecasting error for

[0102] in, This represents the predicted load value of the power grid; New energy prediction error is

[0103] in, To provide real-time power output for new energy sources in the power grid. Contribute to the prediction of new energy sources in the power grid; Average frequency error for

[0104] in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

[0105] Preferably, the anomaly extraction module calculates the real-time value of each indicator at a preset time interval, and constructs a curve of each indicator changing over time based on the real-time value and the time interval; furthermore, the anomaly extraction module obtains the real-time values ​​of continuous indicators during the test period, and extracts outliers in the real-time values ​​according to the distribution of the real-time values.

[0106] Preferably, the anomaly extraction module is further configured to: arrange real-time values ​​in order according to their magnitude and obtain an indicator array; extract the middle portion of the indicators in the indicator array according to the extraction rate to obtain the first and last indicators among the partial indicators; obtain the normal value range of the real-time values ​​based on the weight difference between the first and last indicators; and determine the abnormal value of any indicator in the indicator array if it exceeds the normal value range.

[0107] Preferably, for a single outlier or multiple consecutive outliers, the anomaly extraction module extracts all real-time values ​​within a preset time range before and after the outlier, thus constituting an outlier phase.

[0108] Preferably, the function construction module is also used to process grid load data, grid maintenance plans, and grid generation plans; and the grid load data at least includes the real-time load of the grid corresponding to the current abnormal phase. The power grid maintenance plan should at least include the maintenance time, scope, and outage capacity corresponding to the current abnormal phase of the power grid. The power grid generation plan should include at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which should consist at least of the power grid's predicted load values. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply to the power grid. Provide evidence; the function building module extracts the values ​​of each independent variable at different times under the current abnormal stage and constructs a bar chart.

[0109] Preferably, the function construction module is also used to: construct a fitting formula between the current indicator and the current independent variable based on the calculation formula of the current indicator, so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage; and differentiate the current independent variable based on the fitting formula to obtain a sensitivity function based on the current independent variable and the current dependent variable.

[0110] Preferably, the report generation module is used to: extract the highest-order coefficient of the current independent variable in the sensitivity function as the determinant coefficient of the current independent variable under the current abnormal stage; compare the determinant coefficients of each independent variable of the current dependent variable, and select one or more independent variables whose determinant coefficients exceed a preset value as the independent variables that play a decisive role under the current abnormal stage.

[0111] Preferably, the report generation module is also used to determine: if the independent variable that plays a decisive role in the current abnormal phase is the power grid maintenance plan, then provide improvement suggestions for the power grid maintenance plan based on the outlier; if the independent variables that play a decisive role in the current abnormal phase are power grid load data and power grid generation plan, then propose improvement suggestions for the stability limit of power grid equipment; and the report generation module summarizes the improvement suggestions into power grid safe operation characteristics.

[0112] Preferably, the report generation module generates a power grid safety operation feature report based on the extracted power grid safety operation features.

[0113] It is understood that the extraction system, in order to implement the various functions in the methods provided in the embodiments of this application, includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the algorithmic steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0114] This application embodiment can divide the system into functional modules based on the above method example. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0115] The system is implemented through communication connections of one or more devices. Each device includes at least one processor, a bus system, and at least one communication interface. The processor may consist of a central processing unit (CPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or other hardware. The memory may consist of read-only memory (ROM), random access memory (RAM), etc. The memory may exist independently and be connected to the processor via a bus. Alternatively, the memory may be integrated with the processor. The hard disk may be a mechanical hard disk (HDD) or a solid-state drive (SSD), etc. This embodiment of the invention does not limit the specific implementation. The above embodiments are typically implemented using software and hardware. When implemented using software programs, it can be implemented in the form of a computer program product. This computer program product includes one or more computer instructions.

[0116] When computer program instructions are loaded and executed on a computer, the corresponding functions are implemented according to the process provided in the embodiments of this invention. The computer program instructions involved may be assembly instructions, machine instructions, or code written in a programming language, etc.

[0117] The beneficial effects of this invention are that, compared with the prior art, the method and system for extracting power grid safety operation characteristics based on key indicators in this invention extracts and analyzes the sensitivity influencing factors of power grid anomaly indicators, extracts the safety operation characteristics of the power grid and achieves visualization display, and provides a report with suggestions for power grid operation safety and risk prevention and control based on the characteristics of the sensitivity influencing factors. This invention is effective and reliable, and by analyzing power grid anomaly data, it fully explores the operating characteristics of the power grid under abnormal conditions and provides improvement suggestions.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for extracting power grid safety operation features based on key indicators, characterized in that, The method includes the following steps: Data on power grid operation is collected to calculate power grid indicators, which include the power grid's line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, short-circuit current, power grid supply capacity, load forecasting error, new energy forecasting error, and average frequency error. Track the calculation results of the indicator, identify outliers in the calculation results, and extract the abnormal phase of the indicator based on the time corresponding to the outlier. The step of tracking the calculation results of the indicators and determining outliers in the calculation results includes: calculating the real-time value of each indicator at a preset time interval, and constructing a curve of each indicator changing over time based on the real-time value and the time interval; obtaining the real-time values ​​of the indicators continuously during the test period, and extracting outliers from the real-time values ​​according to the distribution of the real-time values; the step of extracting the abnormal stage of the indicator based on the time corresponding to the outlier includes: for a single outlier or multiple consecutive outliers, extracting all the real-time values ​​within a preset time range before and after the outlier, thereby constituting an abnormal stage; Using grid load data, grid maintenance plan, and grid power generation plan as independent variables, and each of the aforementioned indicators as dependent variables, sensitivity functions for each of the aforementioned indicators are constructed based on the relationship between the independent and dependent variables under the aforementioned abnormal phases. Solve the sensitivity function to obtain the independent variable that plays a decisive role in the current abnormal stage, and extract the characteristics of power grid safe operation based on the independent variable that plays a decisive role; The step of solving the sensitivity function to obtain the independent variable that plays a decisive role in the current abnormal stage includes: extracting the highest-order coefficient of the current independent variable in the sensitivity function as the determination coefficient of the current independent variable in the current abnormal stage; comparing the determination coefficients of each independent variable of the current dependent variable, and selecting one or more independent variables whose determination coefficients exceed a preset value as the independent variables that play a decisive role in the current abnormal stage.

2. The method for extracting power grid safety operation features based on key indicators as described in claim 1, characterized in that: The power supply capacity index for in, For the real-time load of the power grid, The real-time power supply to the power grid; The load forecasting error for in, This represents the predicted load value of the power grid; The new energy prediction error is: in, To provide real-time power output for new energy sources in the power grid. Contribute to the prediction of new energy sources in the power grid; The average frequency error for in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

3. The method for extracting power grid safety operation features based on key indicators as described in claim 1, characterized in that: The step of extracting outliers from the real-time values ​​based on their distribution also includes: Arrange the real-time values ​​in order according to their magnitude to obtain an indicator array; Based on the extraction rate, extract the middle portion of the indicators in the indicator array to obtain the first and last indicators in the middle portion of the indicators; Based on the weight difference between the first and last indicators, the normal value range of the real-time values ​​is obtained respectively. If any indicator in the indicator array exceeds the normal value range, it is determined to be an outlier.

4. The method for extracting power grid safety operation features based on key indicators as described in claim 1, characterized in that: The method of using grid load data, grid maintenance plans, and grid generation plans as independent variables also includes: The power grid load data includes at least the real-time load of the power grid corresponding to the current abnormal phase. ; The power grid maintenance plan shall at least include the maintenance time, maintenance scope, and outage capacity corresponding to the current abnormal phase of the power grid. ; The power grid generation plan includes at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which is based at least on the predicted load value of the power grid. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply for the power grid. Provide evidence; Extract the values ​​of each independent variable at different times during the current abnormal phase and construct a bar chart.

5. The method for extracting power grid safety operation features based on key indicators as described in claim 1, characterized in that: The method of constructing sensitivity functions for each indicator based on the relationship between independent and dependent variables during the abnormal phase further includes: Based on the calculation formula of the current indicator, a fitting formula between it and the current independent variable is constructed so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage. Based on the fitting formula, the derivative of the current independent variable is calculated to obtain the sensitivity function based on the current independent variable and the current dependent variable.

6. The method for extracting power grid safety operation features based on key indicators as described in claim 1, characterized in that: The generation of power grid safe operation characteristics based on the decisive independent variables also includes: If the power grid maintenance plan is the decisive independent variable during the current abnormal phase, then suggestions for improving the power grid maintenance plan are provided based on the abnormal value. If the decisive independent variables during the current abnormal phase are grid load data and grid generation plan, then suggestions for improving the grid equipment stability limit should be proposed. The improvement suggestions are summarized into the power grid safe operation characteristics.

7. The method for extracting power grid safety operation features based on key indicators as described in claim 6, characterized in that: Based on the extracted power grid safety operation characteristics, a power grid safety operation characteristic report is generated.

8. A power grid safety operation feature extraction system based on key indicators, characterized in that: The system includes an indicator acquisition module, an anomaly extraction module, a function construction module, and a report generation module; wherein, The indicator acquisition module is used to collect power grid operation data in real time to calculate power grid indicators, which include the power grid's line overload safety margin, main transformer overload safety margin, bus voltage safety level, cross-sectional safety margin, short-circuit current, power supply capacity, load prediction error, new energy prediction error, and average frequency error. The anomaly extraction module is used to track the calculation results of the indicator, determine the abnormal values ​​in the calculation results, and extract the abnormal stage of the indicator based on the time corresponding to the abnormal value. The anomaly extraction module calculates the real-time value of each indicator at preset time intervals and constructs a curve of each indicator changing over time based on the real-time value and the time interval. Furthermore, the anomaly extraction module acquires the real-time values ​​of the indicators consecutively during the test period and extracts outliers from the real-time values ​​based on their distribution. For a single outlier or multiple consecutive outliers, the anomaly extraction module extracts all real-time values ​​within a preset time range before and after the outlier, thus constituting an anomaly stage. The function construction module is used to construct a sensitivity function for each of the above indicators, based on the relationship between the independent and dependent variables under the abnormal stage, using grid load data, grid maintenance plan and grid power generation plan as independent variables and each of the above indicators as dependent variables. The report generation module is used to solve the sensitivity function to obtain the independent variable that plays a decisive role in the current abnormal stage, and extract the characteristics of power grid safe operation based on the independent variable that plays a decisive role. The report generation module is used to extract the highest-order coefficient of the current independent variable in the sensitivity function as the determinant coefficient of the current independent variable under the current abnormal stage; compare the determinant coefficients of each independent variable of the current dependent variable, and select one or more independent variables whose determinant coefficients exceed a preset value as the independent variables that play a decisive role under the current abnormal stage.

9. A power grid safety operation feature extraction system based on key indicators as described in claim 8, characterized in that: The power grid indicators calculated by the indicator acquisition module are: Power supply capacity indicators for in, For the real-time load of the power grid, The real-time power supply to the power grid; Load forecasting error for in, This represents the predicted load value of the power grid; New energy prediction error is in, To provide real-time power output for new energy sources in the power grid. Contribute to the prediction of new energy sources in the power grid; Average frequency error for in, and These are the numbers of the first and last frequency sampling moments in multiple consecutive frequency sampling data sets, respectively, used in the calculation of the average frequency error. The number of samples taken at multiple consecutive frequencies. For the first The real-time frequency collected at each frequency sampling time. This is the standard frequency for the power grid.

10. A power grid safety operation feature extraction system based on key indicators as described in claim 8, characterized in that: The anomaly extraction module is also used for: Arrange the real-time values ​​in order according to their magnitude to obtain an indicator array; Based on the extraction rate, extract the middle portion of the indicators in the indicator array to obtain the first and last indicators in the middle portion of the indicators; Based on the weight difference between the first and last indicators, the normal value range of the real-time values ​​is obtained respectively. If any indicator in the indicator array exceeds the normal value range, it is determined to be an outlier.

11. A power grid safety operation feature extraction system based on key indicators as described in claim 8, characterized in that: The function construction module is also used to process grid load data, grid maintenance plans, and grid generation plans; and... The power grid load data includes at least the real-time load of the power grid corresponding to the current abnormal phase. ; The power grid maintenance plan shall at least include the maintenance time, maintenance scope, and outage capacity corresponding to the current abnormal phase of the power grid. ; The power grid generation plan includes at least a real-time generation plan corresponding to the power grid during the current abnormal phase, which is based at least on the predicted load value of the power grid. Forecasted power output of new energy sources in the power grid It constitutes and provides real-time power supply for the power grid. Provide evidence; The function construction module extracts the values ​​of each independent variable at different times under the current abnormal stage and constructs a bar chart.

12. A power grid safety operation feature extraction system based on key indicators as described in claim 8, characterized in that: The function building module is also used for: Based on the calculation formula of the current indicator, a fitting formula between it and the current independent variable is constructed so that the fitting formula conforms to the fitting curve generated by the current independent variable and the current dependent variable under the current abnormal stage. Based on the fitting formula, the derivative of the current independent variable is calculated to obtain the sensitivity function based on the current independent variable and the current dependent variable.

13. A power grid safety operation feature extraction system based on key indicators as described in claim 8, characterized in that: The report generation module is also used to determine: If the power grid maintenance plan is the decisive independent variable during the current abnormal phase, then suggestions for improving the power grid maintenance plan are provided based on the abnormal value. If the decisive independent variables during the current abnormal phase are grid load data and grid generation plan, then suggestions for improving the grid equipment stability limit should be proposed. Furthermore, the report generation module summarizes the improvement suggestions into the power grid safe operation characteristics.

14. A power grid safety operation feature extraction system based on key indicators as described in claim 13, characterized in that: The report generation module generates a power grid safety operation feature report based on the extracted power grid safety operation features.

Images