A method and device for identifying and correcting abnormal parameters in online security analysis of a power grid

Abstract

The application discloses a power grid online security analysis abnormal parameter identification correction method and device, and the method comprises the following steps: acquiring power grid line parameter real-time detection values, transformer operation parameter real-time detection values, SCADA real-time measurement data and power grid operation state estimation data; performing multi-source data fusion based on the SCADA real-time measurement data and the power grid operation state estimation data to obtain power estimation values of power generation devices output in the power grid and power estimation values of power consumption of power grid loads; performing online security analysis based on the power grid line parameter real-time detection values, the transformer operation parameter real-time detection values, the power estimation values of the power generation devices output and the power estimation values of the power consumption of the power grid loads to obtain steady-state power flow operation state data; and identifying whether the power grid line parameter real-time detection values and the transformer operation parameter real-time detection values are abnormal and correcting the same based on an intelligent identification model in combination with the SCADA measurement data, the power grid operation state estimation data and the steady-state power flow operation state data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power grid state detection, and in particular to a power grid online security analysis abnormal parameter identification and correction method and device. BACKGROUND

[0002] With the continuous deepening of energy transformation, the new power system is accelerating the construction, and the power supply composition, power grid form, load characteristics, technical foundation and operation characteristics have undergone profound changes. The safety and stability of the power system are facing new situations, new requirements and new challenges. With the continuous changes in the material and technical foundations of primary energy characteristics, power layout and functions, network size and form, load structure and characteristics, and new power system technologies, the new generation of online security analysis system requires that each link must be completed quickly and accurately within a limited time. However, the online security analysis technology has many links and long business chains, and any deviation in any link or parameter may cause errors in the results of online security analysis, leading to the failure of online security analysis to function normally and effectively. The application of online security analysis of the power system is facing new problems and challenges. Therefore, it is of great significance to study the abnormal parameter identification and correction method of power grid online security analysis for promoting the sustainable development and reliable operation of the power system.

[0003] At present, a few methods for identifying abnormal parameters of power grid equipment have been applied in the power system. Due to the wide range of data sources of various models and measured data in the power system, errors or errors will inevitably occur in the process of maintenance, transmission, storage and processing. The existing researches on line parameter identification are based on different data or improved algorithms, and most of the methods ignore the influence of part of the abnormal data, which may not be covered. Online security analysis only performs simple discrimination and inspection on abnormal data, and lacks effective means for identifying equipment steady-state parameters under the influence of measurement errors and parameter errors. In addition, with the increasing time-varying nonlinearity, partial observability and random uncertainty of the new power system on the source and load side, the power grid operation mode arrangement is complex and variable, and the simulation scale is further expanded. Although the existing researches have studied the reasons for power flow divergence and line parameter identification, they have ignored the huge influence of part of the abnormal data. In the current online security analysis, only simple discrimination and inspection are performed on abnormal data, and there is still a lack of effective means for identifying equipment steady-state parameters under the influence of measurement errors and parameter errors.

[0004] In the fast-changing environment of new power system operation analysis scenarios, the inaccuracy of static and dynamic parameters of equipment models will not only bring large power flow error, but also lead to power flow divergence, abnormal termination of simulation analysis calculation, and other problems, affecting the accuracy and reliability of online safety analysis. It is urgent to develop abnormal diagnosis and adaptive fault-tolerant technology for the whole process of online safety analysis to improve the accuracy and reliability of online safety analysis, and ultimately improve the practical level of online safety analysis calculation in new power systems. SUMMARY

[0005] The purpose of the embodiment of the application is to provide a power grid online safety analysis abnormal parameter identification and correction method and device, which analyzes, identifies and corrects abnormal parameters by combining SCADA measurement, power grid operation state estimation and other multi-source data, realizes rapid tracing and adaptive fault tolerance of key parameters and process abnormalities, and provides accuracy, timeliness and reliability of online safety analysis.

[0006] To solve the above technical problems, the first aspect of the embodiment of the application provides a power grid online safety analysis abnormal parameter identification and correction method, including the following steps:

[0007] Obtain real-time detection values of power grid line parameters, real-time detection values of transformer operating parameters, SCADA real-time measurement data and power grid operation state estimation data;

[0008] Based on the SCADA real-time measurement data and the power grid operation state estimation data, multi-source data fusion is performed to obtain power estimation values of power generation devices output in the power grid and power estimation values of power consumption of power grid loads;

[0009] Based on the real-time detection values of the power grid line parameters, the real-time detection values of the transformer operating parameters, the power estimation values of the power generation devices output and the power estimation values of the power consumption of the power grid loads, online safety analysis steady-state power flow calculation is performed to obtain steady-state power flow operating state data;

[0010] Based on a power grid equipment abnormal parameter intelligent identification model, the SCADA real-time measurement data and its historical measurement data, the power grid operation state estimation data and the steady-state power flow operating state data are combined to identify whether the real-time detection values of the power grid line parameters and the real-time detection values of the transformer operating parameters are abnormal;

[0011] According to the identification result, correction is performed.

[0012] Further, the parameter types of the real-time detection values of the power grid line parameters include the resistance, reactance and shunt susceptance of each line in the power grid;

[0013] The parameter types of the real-time detection values of the transformer operating parameters include the excitation impedance, leakage impedance and tap position of the transformer.

[0014] The parameter type of the power grid operation state estimation data includes: each node voltage, each line power, transformer transmission power, power generation device output power and power grid load consumption power.

[0015] Further, the power grid operation state estimation data includes: voltage amplitude and phase of each node in the power grid, active power and reactive power transmitted by each line, active power and reactive power transmitted by the transformer, active power and reactive power output by the power generation device, and active power and reactive power consumed by the power grid load.

[0016] Further, the correction further includes:

[0017] Obtaining corresponding steady-state power flow operation state data under the abnormal parameter scenario;

[0018] Calculating the Euclidean distance value of the steady-state power flow operation state data and a plurality of groups of steady-state power flow operation state data in the intelligent identification model training set of the power grid equipment abnormal parameter;

[0019] Selecting a group of steady-state power flow operation state data with the smallest Euclidean distance value to correct the abnormal parameter.

[0020] Further, the calculation formula of the Euclidean distance value is:

[0021]

[0022] Wherein, the vector is all steady-state power flow operation state data in the training set; is the Euclidean distance of the power flow operation state vector x under the current abnormal parameter and any , i is the serial number of the node, and N is the number of power grid nodes.

[0023] Further, the power grid online security analysis abnormal parameter identification correction method further includes:

[0024] Performing a credibility evaluation of the parameter abnormality, and the credibility evaluation formula is:

[0025]

[0026] Wherein, θ is the credibility of the parameter abnormality, 1 represents credibility, and 0 represents uncredibility; d0 is the minimum Euclidean distance threshold in .

[0027] Further, the fusion method of the multi-source data fusion is:

[0028]

[0029] wherein, S i is the active and reactive power variable of the power generation device and power grid load of node i required for online security analysis steady-state power flow calculation, i = 1, 2,..., N, N is the total number of nodes of the power grid; is the power estimation value of the power generation device output and the power estimation value of the power grid load consumption, and the number of node power measurement values is usually less than N, wherein i [1, 2,..., M], M is the total number of measurement nodes of the power grid, and usually M < N; is the power estimation value of the power generation device output and the power estimation value of the power grid load consumption, and the number of node power measurement values is usually less than N, wherein i [1, 2,..., M], M is the total number of measurement nodes of the power grid, and usually M < N;

[0030] Further, the judging formula whether the real-time detection value of the power grid line parameter and the real-time detection value of the transformer operating parameter are abnormal or not is:

[0031]

[0032] wherein, κ ij is the parameter of the branch ij where the power grid line or the transformer is located, ij is the node i and the node j connected by the branch respectively, ij = 1, 2,..., L, and i≠j, L is the total number of branches of the power grid; is the parameter of the branch ij where the real line and the transformer are located; and alpha is an abnormal coefficient of data;

[0033] When the value of the abnormal coefficient alpha is greater than 3, it is determined that the parameter is abnormal.

[0034] Further, the intelligent identification model of the abnormal parameter of the power grid equipment is a DNN deep neural network.

[0035] Correspondingly, a second aspect of the embodiment of the application also provides a power grid online security analysis abnormal parameter identification correction method and device, comprising:

[0036] A detection data acquisition module is configured to acquire real-time detection values of power grid line parameters, real-time detection values of transformer operating parameters, SCADA real-time measurement data and power grid operation state estimation data;

[0037] A data fusion processing module is configured to perform multi-source data fusion based on the SCADA real-time measurement data and the power grid operation state estimation data to obtain power estimation values of power generation device outputs and power estimation values of power grid load consumptions;

[0038] A steady-state power flow calculation module is configured to perform online security analysis steady-state power flow calculation based on the real-time detection values of the power grid line parameters, the real-time detection values of the transformer operating parameters, the power estimation values of the power generation device outputs and the power estimation values of the power grid load consumptions to obtain steady-state power flow operation state data.

[0039] a data anomaly judgment module, configured to identify whether the real-time detection values of the power grid line parameters and the real-time detection values of the transformer operation parameters are abnormal based on the power grid equipment anomaly parameter intelligent identification model in combination with the SCADA real-time measurement data and historical measurement data, the power grid operation state estimation data and the steady-state power flow operation state data;

[0040] a data anomaly correction module, configured to correct according to the identification result.

[0041] Correspondingly, the third aspect of the embodiment of the present application also provides an electronic device, comprising: at least one processor; and a memory connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the power grid online safety analysis anomaly parameter identification and correction method.

[0042] Correspondingly, the fourth aspect of the embodiment of the present application also provides a computer readable storage medium having computer instructions stored thereon, and the instructions are executed by a processor to implement the power grid online safety analysis anomaly parameter identification and correction method.

[0043] The above technical solutions of the embodiment of the present application have the following beneficial technical effects:

[0044] 1. By obtaining real-time detection values of power grid lines, transformers and the like, as well as SCADA real-time measurement data and power grid operation state estimation data, and performing multi-source data fusion, the various types of scattered data can be effectively integrated, which can more comprehensively and accurately estimate the power state of the power generation device output and the power state of the power grid load consumption, provide a reliable data basis for subsequent analysis and decision-making, and help accurately grasp the operation situation of the power grid;

[0045] 2. Based on the integrated multiple data, the steady-state power flow calculation of online safety analysis is performed to obtain the steady-state power flow operation state data. This enables accurate analysis of the power flow distribution and the like of the power grid under the current operation state, and early insight into possible safety hazards, such as line overload, voltage out-of-limit and the like, thereby providing strong support for ensuring the safe and stable operation of the power grid, and timely taking corresponding measures for adjustment and optimization;

[0046] 3. Utilize the grid equipment abnormal parameter intelligent identification model, combine rich real-time and historical data for comprehensive judgment, can intelligently identify whether the grid line parameter real-time detection value and the transformer operation parameter real-time detection value exist abnormally, once finding abnormal, can correct in time, effectively avoid the grid operation failure or inaccurate analysis result caused by parameter abnormality, improve the reliability of grid operation and the accuracy of analysis result, reduce the operation and maintenance cost and potential risk. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 It is the flow chart of the grid online safety analysis abnormal parameter identification correction method provided by the embodiment of the application.

[0048] Figure 2 It is the IEEE-39 node standard topology provided by the embodiment of the application.

[0049] Figure 3a It is the model identification result schematic diagram of the typical line 1 parameter abnormality provided by the embodiment of the application.

[0050] Figure 3b It is the model identification result schematic diagram of the typical line 5 parameter abnormality provided by the embodiment of the application.

[0051] Figure 3c It is the model identification result schematic diagram of the typical line 9 parameter abnormality provided by the embodiment of the application.

[0052] Figure 3d It is the model identification result schematic diagram of the typical line 13 parameter abnormality provided by the embodiment of the application.

[0053] Figure 4 It is the correction result schematic diagram of the abnormal parameter provided by the embodiment of the application.

[0054] Figure 5 It is the module block diagram of the grid online safety analysis abnormal parameter identification correction device provided by the embodiment of the application.

[0055] Reference signs:

[0056] 1, detection data acquisition module, 2, data fusion processing module, 3, steady-state flow calculation module, 4, data abnormality judgment module, 5, data abnormality correction module. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical scheme and advantages of the present application more clear and obvious, the present application is further described in detail below in combination with specific embodiments and with reference to the drawings. It should be understood that these descriptions are only exemplary and do not limit the scope of the present application. In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present application.

[0058] Please refer to Figure 1 The first aspect of the embodiment of the present application provides a power grid online security analysis abnormal parameter identification correction method, comprising the following steps:

[0059] Step S100, acquiring power grid line parameter real-time detection value, transformer operation parameter real-time detection value, SCADA real-time measurement data and power grid operation state estimation data.

[0060] Specifically, the parameter types of the power grid line parameter real-time detection value include the resistance, reactance and shunt susceptance of each line in the power grid, the parameter types of the transformer operation parameter real-time detection value include the excitation impedance, leakage impedance and tap position of the transformer, and the parameter types of the power grid operation state estimation data include the voltage of each node, the power of each line, the transmission power of the transformer, the output power of the power generation device and the power consumption power of the power grid.

[0061] Further, the power grid operation state estimation data refers to the operation state data of the entire power grid obtained by mathematical methods such as least squares, including the voltage amplitude and phase of each node in the power grid, the active power and reactive power transmitted by each line, the active power and reactive power transmitted by the transformer, the active power and reactive power output by the power generation device, and the active power and reactive power consumed by the power grid load.

[0062] The SCADA (Supervisory Control and Data Acquisition) system can collect data of numerous devices and nodes in the power grid in real time, including the current, voltage, switch state and other information of each line, which directly reflects the actual operation of each monitoring point in the power grid at the current time; the above information has high real-time performance and can quickly feedback various changes in the power grid. For example, when the load of a line in the power grid suddenly increases, the SCADA system can almost instantly detect the change of the line current and transmit the related data.

[0063] Step S200, performing multi-source data fusion based on the SCADA real-time measurement data and the power grid operation state estimation data to obtain power estimation values of the power generation device output and the power grid load consumption.

[0064] Specifically, the fusion method of multi-source data fusion is as follows:

[0065]

[0066] Wherein, S iThe power generation device and power grid load active and reactive power variables of node i required for online security analysis steady-state power flow calculation, i = 1, 2, …, N, N is the total number of nodes of the power grid; The power generation device output power estimate value and power grid load consumed power estimate value of node i, usually the number of node power measurement values is less than N, here i ∈ [1, 2, …, M], M is the total number of measurement nodes of the power grid, usually M < N; The power generation device output power estimate value and power grid load consumed power estimate value of node i, here i ∈ [M+1, M+2, …, N].

[0067] Step S300, based on the real-time detection value of the power grid line parameter, the real-time detection value of the transformer operating parameter, the power generation device output power estimate value and the power grid load consumed power estimate value, online security analysis steady-state power flow calculation is carried out, and steady-state power flow operating state data is obtained.

[0068] In addition, when the above online security analysis steady-state power flow calculation is carried out, the number of groups of steady-state power flow operating state data G is usually selected to be ≥1000 groups.

[0069] Step S400, based on the power grid equipment abnormal parameter intelligent identification model, combining the SCADA real-time measurement data and its historical measurement data, the power grid operating state estimate data and the steady-state power flow operating state data, whether the power grid line parameter real-time detection value and the transformer operating parameter real-time detection value are abnormal is identified, and if abnormal, correction is carried out.

[0070] Specifically, the judgment formula of whether the power grid line parameter real-time detection value and the transformer operating parameter real-time detection value are abnormal is:

[0071]

[0072] Wherein, κ ij The parameter of the line or transformer branch ij, ij is the node i and node j connected by the transformer branch, ij = 1, 2, …, L, and i≠j, L is the total number of branches of the power grid; The true parameter of the line and transformer branch ij; α is the abnormal coefficient of the data.

[0073] The training set of the model contains a sufficient number of normal parameter groups and abnormal parameter groups to meet the statistical requirement for the amount of data, thereby ensuring the effectiveness and reliability of the model training. The normal parameters refer to the parameters of the lines and transformer devices that have been recorded in the SCADA system. The abnormal parameters are composed of two parts. The first part is the abnormal parameters identified based on the deep experience of field experts and engineers and industry standards. These parameters directly reflect various abnormal situations that may occur in actual operation. The second part is obtained by secondary screening of the data set to address the possible lack of abnormal data. These abnormal parameters refer to the parameters of a certain power grid device that deviate greatly due to human negligence or other reasons.

[0074] Generally, the abnormal parameters usually deviate more than 3 times the actual device parameter value, i.e., when the value of the abnormal coefficient a is greater than 3, it is determined that the parameter is abnormal.

[0075] For a plurality of groups of normal or abnormal parameters of power grid devices κ ij , combined with the active and reactive power variables S i of the power generation devices and power grid loads of each node obtained as described above, DSA steady-state flow calculation is performed to obtain steady-state flow operating state data x i under each group of parameters, i = 1, 2, …, N, including the voltage amplitude and phase of all nodes of the system, branch transmission active and reactive power; wherein the specific calculation method of the steady-state flow calculation is not limited.

[0076] A suitable threshold is selected to label the abnormal parameters of the power grid devices, i.e., whether the parameters κ ij of the line or transformer branch ij are abnormal is determined according to the 3σ rule, and are labeled as 1 or 0, respectively. The specific labeling method is as follows:

[0077]

[0078] In the formula: represents that the parameters κ ij of the line or transformer branch ij are normal. represents that the parameters κ ij of the line or transformer branch ij are abnormal; ∩ represents and, σ i , σ j are the mean and variance of the flow operating state data x i and x j of the nodes on both sides of any power grid device.

[0079] In addition, the above threshold can be further combined with data checked by artificial experience and industry standards to improve the accuracy of the model.

[0080] The power grid equipment abnormal parameter intelligent identification model is specifically a model taking SCADA measurement data state estimation data and DSA online power flow calculation data x i as input, taking all branch parameters κ ij as output to obtain an artificial intelligence model f AI as follows:

[0081]

[0082] In the formula, the output is whether the branch parameter κ ij is abnormal, 1 representing that κ ij is normal, and 0 representing that κ ij is abnormal.

[0083] Optionally, the SCADA system historical data include complete record of historical measurement data for not less than 1 year.

[0084] Step S500, correction is performed according to the identification result.

[0085] Further, after the correction in step S500, the following is further included:

[0086] Step S610, corresponding steady-state power flow operating state data under the abnormal parameter scenario are obtained.

[0087] Step S620, the Euclidean distance value of the steady-state power flow operating state data and a plurality of groups of steady-state power flow operating state data in the training set of the power grid equipment abnormal parameter intelligent identification model is calculated.

[0088] Step S630, the steady-state power flow operating state data with the minimum Euclidean distance value is selected to correct the abnormal parameter.

[0089] For the abnormal parameters of the identified real-time detection values of the power grid line parameters and the real-time detection values of the transformer operating parameters, the Euclidean distance of the power flow operating state data under the abnormal parameter scenario and all G groups of data is calculated, then the credibility of the parameter abnormality is evaluated, and the power grid equipment parameters of the data sample with the minimum distance are selected to correct the abnormal parameters that are not credible.

[0090] Specifically, the calculation formula of the Euclidean distance value is as follows:

[0091]

[0092] In the formula, the vector is all steady-state power flow operating state data in the training set; is the power flow operating state vector x under the current abnormal parameter, and is any Euclidean distance of the i-th node, i is the serial number of the node, and N is the number of nodes in the power grid.

[0093] Further, the power grid online security analysis abnormal parameter identification correction method further comprises:

[0094] The credibility of the parameter abnormality is evaluated, and the credibility evaluation formula is:

[0095]

[0096] wherein, θ is the credibility of the parameter abnormality, 1 represents credibility, and 0 represents uncredibility; d0 is the minimum Euclidean distance threshold in the formula (1), which can be selected as 0.0001.

[0097] The power grid equipment parameter of the data sample with the minimum distance is selected, and the uncredible abnormal parameter is corrected:

[0098]

[0099] In the formula, κ i ′ j is the correction value of the branch abnormal parameter κ ij . is the original power grid equipment parameter of the sample corresponding to the minimum Euclidean distance in the formula (1).

[0100] The above identification correction method can accurately obtain the power estimation value of the power generation device output and the power grid load consumption in the power grid by comprehensively integrating the real-time detection values of the power grid lines and transformers, the SCADA real-time measurement data and the power grid operation state estimation data and other multi-source information, and through multi-source data fusion, thereby providing a reliable basis for subsequent analysis. Based on the rich data, the steady-state power flow calculation can accurately obtain the steady-state power flow operation state data, thereby realizing in-depth understanding of the power grid operation condition. In addition, by means of the power grid equipment abnormal parameter intelligent identification model, in combination with the SCADA real-time and historical measurement data, the power grid operation state estimation data and the steady-state power flow operation state data, the abnormal condition of the real-time detection values of the power grid line and transformer operation parameters can be intelligently and accurately identified, and timely correction can be performed, thereby improving the safety and reliability of the power grid operation, effectively reducing the power grid failure risk caused by parameter abnormality, and improving the efficiency and accuracy of power grid operation analysis and maintenance, thereby providing a powerful guarantee for stable and efficient operation of the power grid.

[0101] Optionally, the power grid equipment abnormal parameter intelligent identification model is a DNN deep neural network.

[0102] Hereinafter, the above identification correction process will be described as follows by taking one specific embodiment of the present application as an example: ​​

[0103] Step One: Data Collection and Multi-source Fusion

[0104] The power grid system data comes from the resistance, reactance and shunt susceptance of each line in the power grid, the excitation impedance, leakage impedance and tap position of the transformer, and the voltage of each node, the power of each line, the transmission power of the transformer, the output power of the power generation device and the power consumption of the power grid load, which come from different data sources such as sensors, SCADA systems, historical databases, etc., which constitute the basis for power grid abnormal parameter identification and correction. These data sources may use different coding standards. In order to process and analyze the data later, first of all, the multi-source data needs to be converted into a unified format. The specific operation is to convert all data to the same coding standard (UTF-8), for each data source, establish a data mapping table, map the original data field to the corresponding field of the unified data model, and save the converted data.

[0105] Step Two: Data Verification and Formatting

[0106] In view of the data from diversified information sources, especially for text data such as model identification and name details of lines and transformers in the power grid system, there are often challenges of inconsistent formats in data storage. In order to ensure the excellence of data quality, this step introduces a unified formatting mechanism, aiming to unify and standardize the multiple components in the device model information, such as transformer name and voltage level, to the standardized format of "NN-XX / YY". This measure not only promotes the clarity and consistency of the internal structure of the data, but also lays a solid foundation for subsequent data processing and analysis.

[0107] Then, a strict data verification process is performed to ensure that all information meets the preset format standard. Next, in order to optimize data characteristics and improve the learning efficiency and stability of the neural network model, a data normalization processing strategy is implemented. This strategy scales all feature dimensions to the same order of magnitude through precise mathematical transformation, effectively alleviating the gradient vanishing or explosion problem caused by feature scale differences, thereby accelerating the convergence speed of the training process and significantly enhancing the robustness and prediction accuracy of the model. In addition, data normalization further optimizes the distribution characteristics of the data, creating more favorable conditions for subsequent data mining and pattern recognition tasks.

[0108] Step Three: State Estimation

[0109] For an IEEE-39 node system, the power grid device parameters and operating condition data preprocessed through step one are read into the program, and using these data, the state estimation of the power grid is carried out through the least square method to obtain the operating state data including the amplitude and phase of each node voltage, the transmission active and reactive power of the line and transformer. The state estimation result is combined with the measurement data of the power system data acquisition and monitoring control system to obtain the original data for subsequent power flow calculation of the embodiment.

[0110] Step four: selection of data and DSA power flow calculation

[0111] Combined with the parameters of the power grid lines and the power data of the power generation devices and power grid loads, online security analysis steady-state power flow calculation is carried out to obtain the voltage amplitude and phase of all nodes of the system and the active and reactive power data transmitted by the branch.

[0112] In order to ensure the statistical law and the subsequent training effect, more than 1000 groups of data are selected in the embodiment, including normal and abnormal state data. The normal data refers to the line and transformer device parameters set according to the common range in engineering practice and the typical values recorded by the SCADA system, while the abnormal data mainly comes from the data identified by experts and engineers according to deep experience and industry standards, supplemented by secondary screening by abnormal parameter judgment formula. The combination of the two parts of abnormal data can ensure that the model can learn the characteristics of extreme abnormal conditions.

[0113] Step five: establishment of artificial intelligence model for abnormal parameter identification

[0114] Using the selected data set in step four, an artificial intelligence power grid device abnormal parameter intelligent identification model is established, a DNN deep neural network is used to construct a prediction model, and the SCADA measurement data State estimation data and DSA online power flow calculation data x i are input, and whether all branch parameters are abnormal is output to obtain the artificial intelligence model.

[0115] Combined Figure 2 , Figure 3a , Figure 3b , Figure 3c and Figure 3d show the abnormal parameter identification results of some lines, in which typical lines 1, 5, 9 and 13 respectively represent the lines between nodes 0-38, 2-17, 4-7 and 7-8 in Figure 2 In this data set, there are a total of 25316 samples, only 11 samples are misjudged, and the correct identification rate of parameter abnormality is more than 99%, so it can be seen that the artificial intelligence model used in the embodiment can realize the identification of line abnormality, and the identification accuracy is high.

[0116] To characterize the accuracy of the AI ​​model's recognition, this implementation method selected a portion of the data that had already been labeled with parameter anomalies as a test set for training. The training results are shown in Table 1 below.

[0117] Table 1. Results of Artificial Intelligence Model Parameter Recognition

[0118]

[0119]

[0120] For dataset 2, which differs significantly from the training set, despite the increased difficulty in recognition, it still maintains extremely high recognition accuracy. Out of 25,100 total samples, the number of misclassified samples was 426, which means it achieved a correct recognition rate of 98.30%, verifying that the model still has high stability and effectiveness when facing datasets with significant differences.

[0121] Step Six: Correcting Abnormal Parameters.

[0122] For the identified abnormal parameters of power grid equipment, the Euclidean distance between the power flow operation status data and all data under the abnormal parameter scenario is calculated using the Euclidean distance calculation formula. Then, the credibility of the parameter anomaly is evaluated according to the credibility evaluation formula, and the power grid equipment parameters of the data sample with the smallest distance are selected to correct unreliable abnormal parameters.

[0123] The identification and correction methods described in the above embodiments are data-driven and can make full use of a large amount of historical and online real-time data to identify abnormal parameters in the power grid, which helps to improve the accuracy of power system models and make theoretical analysis closer to actual operation. Furthermore, through automated and intelligent identification and correction of abnormal parameters, the reliance on manual monitoring can be reduced, labor costs can be lowered, and operation and maintenance efficiency can be improved. Finally, the proposal and application of the above methods promote the development of power systems towards intelligence and automation, and provide strong technical support for the construction of new power systems.

[0124] In addition, such as Figure 4 As shown, in a specific application scenario, the initial resistance value of a line was measured to be 10.4742Ω. This resistance value significantly deviated from the normal or expected range and was considered abnormally high. Based on the abnormal parameter identification and correction method for online power grid safety analysis in this invention, the resistance value of this line was precisely corrected. After correction, the resistance value of the line was significantly reduced to 1.792176Ω. This process not only effectively eliminated the abnormally high resistance value but also brought the resistance value back to a reasonable and stable range, thereby ensuring the accuracy and reliability of power grid operation.

[0125] Correspondingly, please refer to Figure 5 The second aspect of the embodiment of the present application also provides an online security analysis abnormal parameter identification correction device of a power grid, comprising:

[0126] A detection data acquisition module 1 is configured to acquire real-time detection values of power grid line parameters, real-time detection values of transformer operation parameters, SCADA real-time measurement data and power grid operation state estimation data;

[0127] A data fusion processing module 2 is configured to perform multi-source data fusion based on the SCADA real-time measurement data and the power grid operation state estimation data to obtain power estimation values of power generation devices output in the power grid and power estimation values of power consumption of loads in the power grid;

[0128] A steady-state power flow calculation module 3 is configured to perform online security analysis steady-state power flow calculation based on the real-time detection values of power grid line parameters, the real-time detection values of transformer operation parameters, the power estimation values of power generation devices output and the power estimation values of power consumption of loads in the power grid to obtain steady-state power flow operation state data;

[0129] A data anomaly judgment module 4 is configured to identify whether the real-time detection values of power grid line parameters and the real-time detection values of transformer operation parameters are abnormal based on a power grid equipment abnormal parameter intelligent identification model in combination with the SCADA real-time measurement data and historical measurement data thereof, the power grid operation state estimation data and the steady-state power flow operation state data;

[0130] A data anomaly correction module 5 is configured to perform correction according to the identification result.

[0131] Correspondingly, the third aspect of the embodiment of the present application also provides an electronic device, comprising at least one processor and a memory connected with the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the online security analysis abnormal parameter identification correction method of the power grid.

[0132] Correspondingly, the fourth aspect of the embodiment of the present application also provides a computer readable storage medium having computer instructions stored thereon, the instructions being executed by a processor to implement the online security analysis abnormal parameter identification correction method of the power grid.

[0133] The embodiment of the application aims to protect a power grid online security analysis abnormal parameter identification correction method, comprising the following steps: acquiring power grid line parameter real-time detection values, transformer operation parameter real-time detection values, SCADA real-time measurement data and power grid operation state estimation data power load; based on the SCADA real-time measurement data and the power grid operation state estimation data, multi-source data fusion is carried out to obtain power estimation values of power generation devices in the power grid and power estimation values of power consumption of the power load; based on the power grid line parameter real-time detection values, the transformer operation parameter real-time detection values, the power estimation values of the power generation devices in the power grid and the power estimation values of the power consumption of the power load, online security analysis steady-state power flow calculation is carried out to obtain steady-state power flow operation state data; based on a power grid equipment abnormal parameter intelligent identification model, the SCADA real-time measurement data and its historical measurement data, the power grid operation state estimation data and the steady-state power flow operation state data are combined to identify whether the power grid line parameter real-time detection values and the transformer operation parameter real-time detection values are abnormal; and correction is carried out according to the identification result. The above technical scheme has the following effects:

[0134] 1. By acquiring real-time detection values of power grid lines, transformers and the like, SCADA real-time measurement data and power grid operation state estimation data, and carrying out multi-source data fusion, the various types of scattered data can be effectively integrated, so that the power state of the power generation devices and the power state of the power load consumption can be more comprehensively and accurately estimated, providing a reliable data basis for subsequent analysis and decision-making, and helping to accurately grasp the operation situation of the power grid;

[0135] 2. Based on the integrated multiple data, steady-state power flow calculation is carried out for online security analysis to obtain steady-state power flow operation state data. This enables accurate analysis of the power flow distribution and the like of the power grid under the current operation state, and early insight into possible security risks, such as line overload and voltage out-of-limit, thereby providing strong support for the safe and stable operation of the power grid, and timely taking corresponding measures for adjustment and optimization;

[0136] 3. The power grid equipment abnormal parameter intelligent identification model is used to comprehensively judge in combination with rich real-time and historical data, which can intelligently identify whether the power grid line parameter real-time detection values and the transformer operation parameter real-time detection values are abnormal. Once an abnormality is found, correction can be carried out in time, effectively avoiding power grid operation faults or inaccurate analysis results caused by abnormal parameters, improving the reliability of power grid operation and the accuracy of analysis results, and reducing operation and maintenance costs and potential risks.

[0137] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0138] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, 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 apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0139] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0140] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0141] Finally, it should be noted that the above-mentioned embodiments are merely intended for describing the technical solutions of the present application, but not for limiting it. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the technical solutions of the present application can still be modified or equivalent replaced without departing from the spirit and scope of the present application, and any modification or equivalent replacement should be covered in the protection scope of the claims of the present application.

Claims

1. A method for online security analysis of an electric power grid, characterized in that, The method comprises the following steps: obtaining real-time detection values of power grid line parameters, real-time detection values of transformer operating parameters, SCADA real-time measurement data and power grid operating state estimation data; performing multi-source data fusion based on the SCADA real-time measurement data and the power grid operating state estimation data to obtain power estimation values of power generation devices in the power grid and power estimation values of power consumption of power grid loads; performing online security analysis steady-state power flow calculation based on the real-time detection values of power grid line parameters, the real-time detection values of transformer operating parameters, the power estimation values of power generation devices and the power estimation values of power consumption of power grid loads to obtain steady-state power flow operating state data; identifying whether the real-time detection values of power grid line parameters and the real-time detection values of transformer operating parameters are abnormal based on a power grid equipment abnormal parameter intelligent identification model in combination with the SCADA real-time measurement data and historical measurement data thereof, the power grid operating state estimation data and the steady-state power flow operating state data; performing correction according to the identification result, and the correction further comprises: obtaining corresponding steady-state power flow operating state data under an abnormal parameter scenario; calculating Euclidean distance values of the steady-state power flow operating state data and a plurality of sets of steady-state power flow operating state data in a training set of the power grid equipment abnormal parameter intelligent identification model; selecting a set of steady-state power flow operating state data with the smallest Euclidean distance value to correct the abnormal parameters; a judgment formula for whether the real-time detection values of power grid line parameters and the real-time detection values of transformer operating parameters are abnormal is: ； wherein, is the parameter of the branch where the line or transformer is located, are the nodes to which the branch where the transformer is located is connected, respectively, , , and , is the total number of branches of the power grid; is the parameter of the branch where the real line and transformer are located; is the anomaly coefficient of the data;​​​ When the abnormality coefficient is greater than 3, it is determined that the parameter is abnormal.

2. The power grid online security analysis abnormal parameter identification correction method according to claim 1, characterized in that: the parameter types of the real-time detection values of power grid line parameters include resistances, reactances and shunt admittances of each line in the power grid; the parameter types of the real-time detection values of transformer operating parameters include excitation impedances, leakage impedances and tap positions of transformers; the parameter types of the power grid operating state estimation data include node voltages, line powers, transformer transmission powers, power generation device output powers and power grid load consumption powers.

3. The power grid online security analysis abnormal parameter identification correction method according to claim 2, characterized in that: the power grid operating state estimation data include voltage amplitudes and phases of each node in the power grid, active powers and reactive powers transmitted by each line, active powers and reactive powers transmitted by transformers, active powers and reactive powers output by power generation devices and active powers and reactive powers consumed by power grid loads.

4. The power grid online security analysis abnormal parameter identification correction method according to claim 1, characterized in that: a calculation formula of the Euclidean distance values is: ； where the vector is the steady-state power flow operating state data in the training set; is the power flow operating state vector under the current abnormal parameter is the Euclidean distance between any , i is the serial number of the node, and N is the number of nodes in the power grid.

5. The method of claim 4, wherein, and further comprising: performing a credibility evaluation of the abnormal parameters, and a credibility evaluation formula is: ； wherein, is the confidence of the parameter anomaly, 1 represents trust, and 0 represents untrust; is the minimum Euclidean distance threshold in .

6. The method of grid on-line security analysis abnormal parameter identification correction according to claim 1, characterized in that, a fusion method of the multi-source data fusion is: ； in, Nodes required for online security analysis and steady-state power flow calculation The active and reactive power variables of power generation devices and grid loads , This represents the total number of nodes in the power grid. For nodes The estimated power output of the generating units and the estimated power consumption of the grid load are typically insufficient due to a lack of node power measurements. One, here , The total number of measurement nodes in the power grid is usually... ; For nodes The estimated output power of the generating units and the estimated power consumption of the grid load, here .

7. The power grid online security analysis abnormal parameter identification correction method according to any one of claims 1-6, characterized in that: the power grid equipment abnormal parameter intelligent identification model is a DNN deep neural network.

8. An apparatus for identifying and correcting abnormal parameters in online security analysis of a power grid, characterized in that, The abnormal parameter identification correction method for online security analysis of the power grid based on any one of claims 1-7 comprises: a detection data acquisition module for acquiring real-time detection values of power grid line parameters, real-time detection values of transformer operating parameters, SCADA real-time measurement data, and power grid operating state estimation data; a data fusion processing module for performing multi-source data fusion based on the SCADA real-time measurement data and the power grid operating state estimation data to obtain power estimation values of power generation devices within the power grid and power estimation values of power consumed by loads of the power grid; a steady-state power flow calculation module for performing online security analysis steady-state power flow calculation based on the real-time detection values of power grid line parameters, the real-time detection values of transformer operating parameters, the power estimation values of power generation devices, and the power estimation values of power consumed by loads of the power grid to obtain steady-state power flow operating state data; a data anomaly judgment module for identifying whether the real-time detection values of power grid line parameters and the real-time detection values of transformer operating parameters are abnormal based on a power grid equipment abnormal parameter intelligent identification model in combination with the SCADA real-time measurement data and historical measurement data thereof, the power grid operating state estimation data, and the steady-state power flow operating state data; a data anomaly correction module for performing correction according to the identification result.

Images