Electric power monitoring system alarm method and platform based on risk control rule engine

Abstract

The invention discloses an electric power monitoring system alarm method and platform based on a risk control rule engine, and relates to the related technical field of power grid monitoring, and the method comprises the steps: carrying out the multi-element risk control rule mining and lightweight decision packaging of a target electric power system, taking the multi-element risk control rule mining and lightweight decision packaging as a reference rule base, and constructing the risk control rule engine; through interaction between the FPGA module and the front-end power monitoring assembly, power flow data are determined, data preprocessing and data environment reconstruction are carried out, and a virtual risk space is updated; and according to a risk control rule engine, performing autonomous perception and determination based on rule element nodes on the virtual risk space, generating risk control alarm information, and performing power feedback management based on a risk control rule. The technical problems that in the prior art, complex and changeable electric power risk situations are difficult to perceive and respond in real time, and consequently an electric power monitoring system is insufficient in alarm precision and poor in risk prevention and control capability and operation reliability are solved, and the technical effect of improving the alarm precision, the risk prevention and control capability and the operation reliability of the electric power monitoring system is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid monitoring, and particularly relates to a power monitoring system alarm method and platform based on a risk control rule engine. BACKGROUND

[0002] A power monitoring system is a key to ensuring stable operation of a power grid and preventing faults. A traditional power monitoring system relies on threshold judgment and simple logic rules for alarm. In the face of complex and variable power operation environments, the system often shows insufficient flexibility and poor adaptability. Moreover, the power monitoring system generates a large amount of data of various types. How to efficiently and timely process the various data and extract valuable risk control information from them becomes an important part of power monitoring. With the application of the risk control rule engine technology in the field of network security, the technology can analyze and make decisions on complex events in real time through a predefined rule set. The introduction of the risk control rule engine into the power monitoring system can realize multi-dimensional and intelligent analysis of power operation data, thereby improving the accuracy and timeliness of alarm. However, the data environment in the power monitoring system is complex, and the power flow data has the characteristics of strong time sequence, much noise and high correlation, which is difficult to effectively preprocess and reconstruct and cannot achieve accurate risk control, thereby affecting the risk prevention and control capability and operation reliability of the power monitoring system.

[0003] Therefore, in the related art, there is a technical problem that it is difficult to realize real-time perception and response to complex and variable power risk situations, resulting in insufficient alarm precision, poor risk prevention and control capability and operation reliability of the power monitoring system. SUMMARY

[0004] The present application provides a power monitoring system alarm method and platform based on a risk control rule engine, which solves the technical problem that it is difficult to realize real-time perception and response to complex and variable power risk situations in the prior art, resulting in insufficient alarm precision, poor risk prevention and control capability and operation reliability of the power monitoring system, and achieves the technical effect of improving the alarm precision, risk prevention and control capability and operation reliability of the power monitoring system.

[0005] The present application provides a power monitoring system alarm method based on a risk control rule engine, which comprises: performing multi-element risk control rule mining and lightweight decision encapsulation on a target power system as a benchmark rule library and constructing a risk control rule engine; determining power flow data through interaction between an FPGA module and a front-end power monitoring component, performing data preprocessing and data environment reconstruction, and updating a virtual risk space, wherein the virtual risk space is updated synchronously according to the collection of the power flow data; performing autonomous perception and judgment on the virtual risk space based on a rule element node according to the risk control rule engine, generating risk control alarm information, and performing power feedback management based on the risk control rule; and wherein the risk control rule engine and the FPGA module are embedded plug-ins of the power monitoring system.

[0006] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: mining a first rule, wherein the first rule is any risk control rule of the target power system; performing training encapsulation on the first rule based on autonomous perception and lightweight rule judgment, as a first rule meta node; completing the construction of the Nth rule meta node, as a rule meta node cluster, wherein N is the total number of rules of the target power system.

[0007] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: determining rule grouping conditions, grouping the rule meta node cluster according to the rule grouping conditions, generating the benchmark rule base; formulating risk control and management logic based on the benchmark rule base and supervising training until convergence, and constructing the risk control rule engine.

[0008] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: using risk control probability rules as the first group of classes, causal paradox rules as the second group of classes, probability collapse rules as the third group of classes, and feedback adjustment rules as the fourth group of classes, wherein the risk control probability is defined based on risk confidence, and probability collapse is based on the directional probability of risk spatiotemporal evolution; writing the first group of classes into the first rule area of ​​the imbalance detection rule sub-library, and writing the second and third group of classes into the second rule area of ​​the imbalance detection rule sub-library; writing the fourth group of classes into the rule area of ​​the balance feedback rule sub-library; and constructing the benchmark rule library based on the imbalance detection rule sub-library and the balance feedback rule sub-library.

[0009] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine also performs the following processing: scanning the power flow data and performing multi-source data fusion and transformation to determine the power panorama state vector; using the power panorama state vector, performing parallel data cleaning and high-frequency feature extraction to determine the risk feature tensor; and mapping the risk feature tensor to a virtual risk space.

[0010] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: by activating the risk control rule engine, a virtual risk space inspection based on rule meta nodes is performed according to the first rule area in the imbalance detection rule sub-library; if there is a rule meta node response, a risk control probability value based on the responding rule meta node is determined; and risk control management of the target power system is performed according to the risk control probability value, wherein the risk control management dimension includes risk control alarm and feedback adjustment.

[0011] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: if the risk control probability value meets the preset probability value, rule matching is performed in the balanced feedback rule sub-library to determine the target feedback adjustment rule, wherein the preset probability value is a critical value based on the risk confidence level; according to the target feedback adjustment rule, the targeted risk control feedback adjustment of the target power system is performed.

[0012] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: if the risk control probability value does not meet the preset probability value, according to the second rule area of ​​the imbalance detection rule sub-library, perform association rule matching and judgment based on causal paradox rule and probability collapse rule to determine the joint judgment result; and perform rule matching and feedback adjustment based on the balance feedback rule sub-library on the joint judgment result.

[0013] In a possible implementation, the power monitoring system alarm method based on the risk control rule engine further performs the following processing: storing rule record data in a temporary database, wherein each rule record data includes an imbalance detection rule, a balance feedback rule, and a rule response result; periodically retrieving rule record data based on the temporary database to identify rule blind spots; and updating and learning the risk control rule engine based on the rule blind spots.

[0014] This application also provides an alarm platform for a power monitoring system based on a risk control rule engine. The platform includes: a risk control rule engine construction unit, used for multi-dimensional risk control rule mining and lightweight decision encapsulation of the target power system, serving as a benchmark rule library and constructing a risk control rule engine; a virtual risk space update unit, used for interaction between the FPGA module and the front-end power monitoring components to determine power flow data, perform data preprocessing and data environment reconstruction, and update the virtual risk space, wherein the virtual risk space is updated synchronously according to the collection of power flow data; and a risk control alarm information generation unit, used for autonomous perception and judgment of the virtual risk space based on rule meta nodes according to the risk control rule engine, generating risk control alarm information and performing power feedback management based on risk control rules; wherein the risk control rule engine and the FPGA module are embedded plug-ins of the power monitoring system.

[0015] This application proposes a power monitoring system alarm method and platform based on a risk control rule engine. This method involves mining multi-dimensional risk control rules and encapsulating lightweight decision-making for the target power system, using these rules as a baseline rule base to construct a risk control rule engine. Through interaction between the FPGA module and front-end power monitoring components, power flow data is determined, and data preprocessing and data environment reconstruction are performed to update the virtual risk space. Based on the risk control rule engine, the virtual risk space is autonomously perceived and judged using rule-based meta-nodes, generating risk control alarm information and implementing power feedback management based on the risk control rules. This addresses the technical problems in existing technologies where real-time perception and response to complex and changing power risk situations are difficult, leading to insufficient alarm accuracy, poor risk prevention and control capabilities, and low operational reliability in power monitoring systems. The method achieves the technical effect of improving the alarm accuracy, risk prevention and control capabilities, and operational reliability of power monitoring systems. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the platform according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic diagram of the alarm method for a power monitoring system based on a risk control rule engine, provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the alarm platform structure of the power monitoring system based on the risk control rule engine provided in the embodiments of this application.

[0019] Figure labeling: Risk control rule engine construction unit 10, virtual risk space update unit 20, risk control alarm information generation unit 30. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides an alarm method for a power monitoring system based on a risk control rule engine, such as... Figure 1 As shown, the method includes: Step S100: Perform multi-dimensional risk control rule mining and lightweight decision encapsulation on the target power system to serve as a benchmark rule library and build a risk control rule engine.

[0022] Preferably, multi-dimensional risk control rules for risk management are mined from historical data, operating mechanisms, and expert experience of the target power system. Specifically, this involves analyzing electrical quantities such as voltage, current, frequency, and power, as well as non-electrical quantities such as equipment temperature and ambient humidity, before and after historical short circuits, overloads, and fluctuations, to obtain massive amounts of historical alarm and fault data. Based on the power grid topology and physical laws such as Kirchhoff's laws, various normal, abnormal, and fault conditions are simulated, and combined with power industry safety operation procedures and expert experience to generate multiple risk control rules of different types. Then, lightweight decision encapsulation is performed on the multi-dimensional risk control rules, that is, the judgment logic, required input data, calculation algorithm, and output results of each risk control rule are packaged into an independent functional module, called a rule meta-node. For example, a line overload risk rule meta-node may encapsulate the current value, line rated capacity, ambient temperature, and overload risk probability value. The encapsulated risk control rules are then grouped and stored according to preset rule grouping conditions to form a structured benchmark rule library. This leads to the construction of a risk control rule engine, which is then integrated into the power monitoring system as an embedded plugin. This engine can dynamically activate relevant rule meta nodes based on real-time data and organize multiple rule meta nodes to work collaboratively according to preset control logic to complete risk perception and decision output. Finally, by integrating the outputs of each rule meta node, the final risk control alarm information or feedback management instructions are generated.

[0023] Furthermore, step S100 also includes step S110, mining a first rule, wherein the first rule is any risk control rule of the target power system; step S120, performing training encapsulation on the first rule based on autonomous perception and lightweight rule judgment, as a first rule meta node; step S130, completing the construction of the Nth rule meta node, as a rule meta node cluster, wherein N is the total number of rules of the target power system.

[0024] Preferably, one of the multiple risk control rules mined from the target power system is randomly selected as the first rule, which includes a specific risk scenario and its triggering conditions. This scenario is then transformed into a structured logical statement. For example, when the real-time current value of a transmission line continuously exceeds its dynamic safe current carrying capacity threshold, there is a risk of overload burnout. An independent functional module is created based on the first rule, which has a data interface and specific judgment logic. Specifically, the unique identifier, input parameters, judgment algorithm, and risk flag or risk probability value of the rule are encapsulated as attributes to obtain the first rule meta-node. A standardized data interface is configured for the rule meta-node, so that when the external data environment provides real-time data, the rule meta-node automatically senses and obtains the input parameters from it. At the same time, it is ensured that the judgment logic encapsulated inside the node does not undergo complex iterative optimization or large-scale data search, but only performs fast comparisons or logical operations. The node is trained until convergence, including debugging and calibrating the internal parameters of the rule meta-node using historical data. The node is also run on a test dataset to ensure that the accuracy, recall, and other indicators of its judgment results meet the predetermined requirements and remain stable. The training and encapsulation of all risk control rules are repeated to complete the construction of the Nth rule meta node, where N is a positive integer greater than 1, representing the total number of rules in the target power system. All rule meta nodes are then combined to generate a rule execution network covering all risk scenarios, i.e., a rule meta node cluster, which contains multiple standardized and executable functional modules. Risk control rules can dynamically and in parallel activate one or more rule meta nodes to work collaboratively as needed, and ensure the stability and reliability of judgment quality.

[0025] Furthermore, step S100 also includes step S140, determining rule grouping conditions, grouping the rule meta node cluster according to the rule grouping conditions, and generating the benchmark rule library; step S150, formulating risk control and management logic based on the benchmark rule library and supervising training until convergence, and constructing the risk control rule engine.

[0026] Preferably, the rule grouping conditions are determined based on the rule's technical type and logical attributes. The rule's technical type refers to a verification rule used for risk identification or a feedback rule used for execution. Logical attributes may include risk control probability rules based on probability statistics, causal paradox rules based on causal reasoning, probability collapse rules based on dynamic evolution, or feedback adjustment rules based on control actions. Then, the rule meta-node cluster is traversed and grouped according to the rule grouping conditions. This involves performing logical judgments on each rule meta-node and assigning it to its corresponding category. The grouping results are then transformed into a structured rule set to generate a baseline rule library. For example, the top layer is divided into an imbalance verification rule sub-library and a balance feedback rule sub-library. Each sub-library is further subdivided into different rule areas, such as the first rule area storing risk control probability rules and the second rule area storing causal paradox rules. The process involves: first, defining probability collapse rules; then, defining a decision-making process for coordinating, scheduling, and executing rules in the baseline rule base; constructing a risk control rule engine; and achieving a stable and reliable state through simulation learning, which may specify the rule execution order, triggering conditions, rule conflict resolution, and output aggregation logic; then, debugging and optimizing the decision-making process of the risk control rule engine. Specifically, in a training environment containing a large number of historical cases or high-fidelity simulations, the initial baseline rule base and risk control logic are deployed, and the output of the risk control rule engine is compared with the standard case solutions. If the risk control rule engine makes a decision error or misses, falsely reports, or responds slowly, its risk control logic is adjusted in reverse, such as modifying the rule execution order, adjusting the rule triggering threshold, and optimizing the conflict resolution strategy, ultimately obtaining a stable and reliable risk control rule engine.

[0027] Furthermore, step S140 also includes step S141, using risk control probability rules as the first group of classes, causal paradox rules as the second group of classes, probability collapse rules as the third group of classes, and feedback adjustment rules as the fourth group of classes, wherein risk control probability is defined based on risk confidence, and probability collapse is the directional probability based on the spatiotemporal evolution of risk; step S142, writing the first group of classes into the first rule area of ​​the imbalance detection rule sub-library, and writing the second and third group of classes into the second rule area of ​​the imbalance detection rule sub-library; step S143, writing the fourth group of classes into the rule area of ​​the balance feedback rule sub-library; step S144, constructing the benchmark rule library based on the imbalance detection rule sub-library and the balance feedback rule sub-library.

[0028] Preferably, based on the inherent logic of the rules, they are divided into four groups: risk control probability rules as the first group, causal paradox rules as the second group, probability collapse rules as the third group, and feedback adjustment rules as the fourth group. Specifically, risk control probability rules are defined based on risk confidence levels and are rules that output quantifiable risk occurrence confidence levels based on historical statistical data or real-time data analysis. Causal paradox rules are used to detect and identify contradictory, unreasonable, or causal-logically violated operating states in the system, and are used to identify and determine hidden or complex faults. Probability collapse rules are rules that dynamically track the trend of risk confidence levels over time and space, and are used to determine and assess the directional trend of risk evolution. Probability collapse is based on the directional probability of risk spatiotemporal evolution. Feedback adjustment rules are rules that directly map to specific control actions or target power system adjustment strategies.

[0029] Preferably, the four categories are categorized into two higher-level sub-libraries according to their functional roles in the risk control process: an imbalance detection rule sub-library and a balance feedback rule sub-library. The imbalance detection rule sub-library is responsible for monitoring, identifying, and diagnosing risks and imbalances in the system. Specifically, the first category is written into the first rule area of ​​the imbalance detection rule sub-library as a rapid screening channel for handling relatively clear and directly quantifiable risks; the second and third categories are written into the second rule area of ​​the imbalance detection rule sub-library as a complex analysis channel for handling latent faults and dynamic risks. The balance feedback rule sub-library is responsible for formulating and implementing corrective measures to restore system balance. The fourth category is written into the rule area of ​​the balance feedback rule sub-library for generating control, adjustment, and protection instructions. Finally, the imbalance detection rule sub-library and the balance feedback rule sub-library are integrated to form a baseline rule library, which highly optimizes the workflow of the risk control rule engine and greatly improves the efficiency and accuracy of risk processing.

[0030] Step S200: Through the interaction between the FPGA module and the front-end power monitoring component, the power flow data is determined and data preprocessing and data environment reconstruction are performed to update the virtual risk space, wherein the virtual risk space is updated synchronously according to the acquisition of power flow data.

[0031] Preferably, the FPGA module is used to directly collect raw, rapidly changing operating data from the power system's monitoring equipment. The FPGA module is a field-programmable gate array, integrated as an embedded plug-in into the power monitoring system. Specifically, the FPGA module communicates with the output interface of the front-end power monitoring components through its configurable I / O pins, capturing and extracting effective power operating parameters in real time and continuously, forming time-series power flow data. The front-end power monitoring components refer to various sensor devices installed in the power system, which may include instrument transformers, smart meters, phasor measurement units (PMUs), remote terminal units (RTUs), etc., and can output continuous high-frequency measurements of voltage, current, power, and frequency. Next, the power flow data is preprocessed within the FPGA module. This includes using digital filters to eliminate high-frequency noise and interference, calibrating the measured values ​​based on sensor turns ratios and coefficients, quickly identifying and removing obviously unreasonable data points based on hardware logic, and utilizing the parallel computing capabilities of the FPGA module to simultaneously perform real-time calculations on multiple types of acquired signals to obtain preliminary features, such as calculating harmonics using fast Fourier transform, calculating the effective values ​​of voltage / current, and calculating power and power factor. Then, the data environment is reconstructed according to the physical and logical relationships of the power system, assembling and determining the power panorama state vector and risk feature tensor as a structured data environment. Finally, this is mapped to the memory space used for risk analysis, updating the virtual risk space. The contents of the virtual risk space are refreshed immediately with each new data acquisition from the front-end power monitoring components, thus forming a risk mirror that changes almost synchronously with the physical power grid, achieving extremely high real-time performance and accuracy in risk identification.

[0032] Furthermore, step S200 also includes step S210, scanning the power flow data and performing multi-source data fusion and transformation to determine the power panorama state vector; step S220, using the power panorama state vector, performing parallel data cleaning and high-frequency feature extraction to determine the risk feature tensor; and step S230, mapping the risk feature tensor to a virtual risk space.

[0033] Preferably, power flow data, which may include voltage, current, power, switch status, equipment temperature, etc., is read, time-aligned, integrated, and standardized. The dispersed measurement point data is spatially correlated according to the power grid topology. Then, continuously changing analog quantities are integrated with discrete state quantities to form a power panorama state vector describing the global instantaneous state of the power system. Each element in the vector represents the instantaneous value of a key state variable. Parallel data cleaning and high-frequency feature extraction are then performed on the power panorama state vector. This involves using the parallel computing capabilities of the FPGA module to simultaneously clean multiple data points in the power panorama state vector, including residual detection and repair, range verification, and abrupt change smoothing. High-frequency feature extraction is then performed to identify derived features that can predict risks, including calculating voltage and frequency gradients / rates of change, instantaneous power spectral density, phase abrupt changes, and harmonic distortion rates. These high-frequency features are then combined into a risk feature tensor according to the power system's spatial location and feature type to structure all risk-related information. Finally, the risk feature tensor is mapped to the virtual risk space, which is a virtual working memory for real-time risk calculation tasks. When a rule meta node is triggered, the risk feature tensor data is read directly from the virtual risk space for judgment without accessing the original database, thereby ensuring high-precision and high-real-time risk prevention and control.

[0034] Step S300: Based on the risk control rule engine, perform autonomous perception and judgment of the virtual risk space based on rule meta nodes, generate risk control alarm information, and perform power feedback management based on risk control rules.

[0035] Step S300 further includes step S310, which involves activating the risk control rule engine and performing a virtual risk space inspection based on rule meta nodes according to the first rule area in the imbalance detection rule sub-library; step S320, which involves determining the risk control probability value based on the responding rule meta node if there is a rule meta node response; and step S330, which involves performing risk control management of the target power system according to the risk control probability value, wherein the risk control management dimension includes risk control alarm and feedback adjustment.

[0036] Preferably, the risk control rule engine is activated by specific events such as timers and switch changes. Then, the first rule area in the imbalance detection rule sub-library is strategically called for rapid and direct logical judgment, outputting a quantified risk probability value. Based on the risk control probability rules of the first rule area, a virtual risk space inspection is performed based on rule meta nodes. Specifically, the risk control rule engine traverses all rule meta nodes in the first rule area and applies each rule meta node in parallel to the current data in the virtual risk space, thereby performing a rapid inspection of the risk points responsible for each rule meta node. When a rule meta node responds, that is, at least one rule meta node finds that its preset trigger condition is met, such as the calculated overload risk probability is greater than 0, it is activated and an output signal is generated. Each response rule metanode outputs its calculation result, namely, determining a standardized risk control probability value. This risk control probability value is the quantified risk confidence level output by the response rule metanode. Then, based on all risk control probability values, risk control management of the target power system is performed, generating risk control alarm information and implementing power feedback management based on risk control rules. This specifically includes triggering alarm information and control commands at corresponding levels. The risk control management dimension includes risk control alarms and feedback adjustment; that is, the alarm level is determined based on the risk control probability value, and the alarm is visually displayed and notified to maintenance personnel. Finally, based on the risk type and risk control probability, corresponding control strategies are matched and executed from the balanced feedback rule sub-library, and automated control commands are executed for feedback adjustment, thus achieving a complete closed loop from perception to decision-making to execution.

[0037] Furthermore, step S330 also includes step S331, if the risk control probability value meets the preset probability value, performing rule matching in the balance feedback rule sub-library to determine the target feedback adjustment rule, wherein the preset probability value is a critical value based on risk confidence; step S332, executing the targeted risk control feedback adjustment of the target power system according to the target feedback adjustment rule.

[0038] Preferably, the preset probability value is a risk action threshold predefined according to safety criteria, used to distinguish different levels of risk. If the risk control probability value meets the preset probability value, rule matching is performed in the balanced feedback rule sub-library storing all control action rules. That is, the corresponding control instruction is searched for in the balanced feedback rule sub-library using attributes such as the current risk type, risk location, and risk level as query conditions. After a successful match, one or more rule meta-nodes are locked, thereby determining the target feedback adjustment rule, which includes various specific control logics. Based on the target feedback adjustment rule, targeted risk control feedback adjustment of the target power system is executed. That is, control instructions are sent to the power equipment directly related to the risk and the rules are predefined, and control actions are executed. Then, the input execution control action is adjusted in reverse according to the output state, thereby completing risk control and preventing local faults from evolving into large-scale power outages.

[0039] Furthermore, step S330 also includes step S333, if the risk control probability value does not meet the preset probability value, according to the second rule area of ​​the imbalance detection rule sub-library, performing association rule matching and judgment based on causal paradox rule and probability collapse rule to determine the joint judgment result; step S334, performing rule matching and feedback adjustment based on balance feedback rule sub-library on the joint judgment result.

[0040] Preferably, if the risk control probability value does not meet the preset probability value, it may be that the risk control probability value is greater than 0 but significantly lower than the preset emergency action threshold, that is, it has not yet reached the level of urgency that requires immediate automated intervention, or the rules in the first rule area have not been triggered but there is an uncertainty risk. In this case, the risk control rule engine's strategy switches from rapid screening to deep analysis mode, calls the second rule area of ​​the imbalance check rule sub-library to handle complex, implicit and dynamically evolving risks, and then performs association rule matching and judgment based on causal paradox rules and probability collapse rules. Specifically, multiple rules are activated to work together, analyze the same state from different angles and associate the analysis results. Based on the causal paradox rules, it is used to explore whether there are logical conflicts or contradictions within the system. For example, a rule determines that the opening of switch A will inevitably lead to the loss of voltage in line B, but if the timestamp of the loss of voltage in line B is detected to be earlier than the timestamp of the opening of switch A, it belongs to the causal paradox, indicating that the specific situation is different from the above rule logic. A forced erasure mechanism may be adopted to send a command to the protection device to forcibly restore the voltage of line B through the backup power supply.

[0041] Preferably, the dynamic trend of risk is analyzed based on probability collapse rules. Specifically, before a risk event occurs, precursor information is detected. At this time, the abnormal event may not have occurred yet, but the characteristics of the precursor are detected. Probability collapse is performed with the event's occurrence determined as the direction to locate the specific occurrence situation and the future occurrence node of the event for pre-risk control processing. The risk control rule engine then integrates the outputs of the causal paradox rule and the probability collapse rule to determine the joint judgment result. Finally, the joint judgment result is used as the query condition to search and match rules in the balanced feedback rule sub-library. Diagnostic feedback or preventive feedback may be determined, and then the matching rules are executed to adjust the feedback, such as switching data sources, restarting remote terminal units, and performing reversible adjustments to delay the development of risks, thereby greatly improving the intelligence level and safety margin of power system monitoring.

[0042] Furthermore, step S300 also includes step S340, storing the rule record data in a temporary database, wherein each rule record data includes an imbalance detection rule, a balance feedback rule, and a rule response result; step S350, periodically retrieving the rule record data based on the temporary database to identify rule blind spots; and step S360, updating and learning the risk control rule engine based on the rule blind spots.

[0043] Preferably, after each risk control decision is executed, the rule record data is stored in a temporary database. Each rule record data includes an imbalance detection rule, a balance feedback rule, and a rule response result. The imbalance detection rule is used to record the rule meta-nodes involved in the diagnosis, the balance feedback rule is used to record the executed feedback adjustment rule, and the rule response result is used to record the final effect of this decision, such as whether the risk was eliminated after feedback adjustment, whether false alarms or false alarms occurred. Then, the temporary database is analyzed. The rule record data is retrieved from the temporary database at a preset period for analysis to identify rule blind spots, including coverage blind spots, inaccurate judgments, and missing correlations. These represent situations where a certain risk event occurs but is missed, i.e., no imbalance detection rule is triggered; a rule is triggered and feedback is executed but proves to be a false alarm or the execution of a non-optimal feedback rule results in poor adjustment effect; multiple rules are triggered separately but there is a lack of correlation rules to identify more complex compound risks. Finally, the risk control rule engine is updated and learned based on rule blind spots. Specifically, for coverage blind spots, new rule meta-nodes are generated through rule mining based on scenario data that leads to missed reports and added to the corresponding rule library; for inaccurate judgments, the internal parameters of relevant rule meta-nodes are adjusted; for missing correlations, new correlation judgment logic is created for risk control. This process updates the risk control rule engine, ensuring that its risk identification accuracy and control precision increase, ultimately achieving intelligent and reliable power system security protection.

[0044] The risk control rule engine and the FPGA module are embedded plug-ins of the power monitoring system.

[0045] In the above text, refer to Figure 1 This paper describes in detail an alarm method for a power monitoring system based on a risk control rule engine according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes an alarm platform for a power monitoring system based on a risk control rule engine, according to an embodiment of the present invention.

[0046] The power monitoring system alarm platform based on a risk control rule engine, according to embodiments of the present invention, addresses the technical problems in existing technologies where it is difficult to perceive and respond to complex and ever-changing power risk situations in real time, resulting in insufficient alarm accuracy, poor risk prevention and control capabilities, and poor operational reliability of the power monitoring system. It achieves the technical effect of improving the alarm accuracy, risk prevention and control capabilities, and operational reliability of the power monitoring system. Figure 2 As shown, the power monitoring system alarm platform based on the risk control rule engine includes: a risk control rule engine construction unit 10, a virtual risk space update unit 20, and a risk control alarm information generation unit 30.

[0047] The risk control rule engine construction unit 10 is used to mine multi-dimensional risk control rules and encapsulate lightweight decision-making for the target power system, serving as a benchmark rule library and constructing a risk control rule engine. The virtual risk space update unit 20 is used to determine power flow data and perform data preprocessing and data environment reconstruction through interaction between the FPGA module and the front-end power monitoring components, updating the virtual risk space, wherein the virtual risk space is updated synchronously according to the collection of power flow data. The risk control alarm information generation unit 30 is used to perform autonomous perception and judgment of the virtual risk space based on rule meta nodes according to the risk control rule engine, generate risk control alarm information, and perform power feedback management based on risk control rules. The risk control rule engine and the FPGA module are embedded plug-ins of the power monitoring system.

[0048] The specific configuration of the risk control rule engine construction unit 10 will be described in detail below. The risk control rule engine construction unit 10 further includes: mining a first rule, wherein the first rule is any risk control rule of the target power system; performing training encapsulation on the first rule based on autonomous perception and lightweight rule judgment, as a first rule meta node; and completing the construction of the Nth rule meta node, as a rule meta node cluster, wherein N is the total number of rules of the target power system.

[0049] The specific configuration of the risk control rule engine construction unit 10 will be described in detail below. The risk control rule engine construction unit 10 further includes: determining rule grouping conditions, grouping the rule meta node cluster according to the rule grouping conditions, generating the benchmark rule library; formulating risk control and management logic based on the benchmark rule library and supervising training until convergence, and constructing the risk control rule engine.

[0050] The specific configuration of the risk control rule engine construction unit 10 will be described in detail below. The risk control rule engine construction unit 10 further includes: using risk control probability rules as the first group of classes, causal paradox rules as the second group of classes, probability collapse rules as the third group of classes, and feedback adjustment rules as the fourth group of classes. Here, risk control probability is defined based on risk confidence, and probability collapse is defined as directional probability based on the spatiotemporal evolution of risk. The first group of classes is written into the first rule area of ​​the imbalance detection rule sub-library, and the second and third groups of classes are written into the second rule area of ​​the imbalance detection rule sub-library. The fourth group of classes is written into the rule area of ​​the balanced feedback rule sub-library. The baseline rule library is constructed based on the imbalance detection rule sub-library and the balanced feedback rule sub-library.

[0051] The specific configuration of the virtual risk space update unit 20 will be described in detail below. The virtual risk space update unit 20 further includes: scanning the power flow data and performing multi-source data fusion and transformation to determine the power panorama state vector; using the power panorama state vector, performing parallel data cleaning and high-frequency feature extraction to determine the risk feature tensor; and mapping the risk feature tensor to the virtual risk space.

[0052] The specific configuration of the virtual risk space update unit 20 will be described in detail below. The virtual risk space update unit 20 further includes: activating the risk control rule engine and performing virtual risk space inspection based on rule meta-nodes according to the first rule area within the imbalance detection rule sub-library; if a rule meta-node response exists, determining the risk control probability value based on the responding rule meta-node; and performing risk control management of the target power system based on the risk control probability value, wherein the risk control management dimension includes risk control alarms and feedback adjustments.

[0053] The specific configuration of the virtual risk space update unit 20 will be described in detail below. The virtual risk space update unit 20 further includes: if the risk control probability value meets a preset probability value, performing rule matching in the balanced feedback rule sub-library to determine the target feedback adjustment rule, wherein the preset probability value is a critical value based on risk confidence; and performing targeted risk control feedback adjustment of the target power system according to the target feedback adjustment rule.

[0054] The specific configuration of the virtual risk space update unit 20 will be described in detail below. The virtual risk space update unit 20 further includes: if the risk control probability value does not meet the preset probability value, performing association rule matching and judgment based on the causal paradox rule and probability collapse rule according to the second rule area of ​​the imbalance detection rule sub-library to determine the joint judgment result; and performing rule matching and feedback adjustment based on the balance feedback rule sub-library on the joint judgment result.

[0055] The specific configuration of the risk control alarm information generation unit 30 will be described in detail below. The risk control alarm information generation unit 30 further includes: temporary database storage of rule record data, wherein each rule record data includes an imbalance detection rule, a balance feedback rule, and a rule response result; periodic retrieval of rule record data based on the temporary database to identify rule blind spots; and updating and learning the risk control rule engine based on the rule blind spots.

[0056] The power monitoring system alarm platform based on a risk control rule engine provided in this invention can execute the power monitoring system alarm method based on a risk control rule engine provided in any embodiment of this invention, and has the corresponding functional modules and beneficial effects of the execution method. The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for alarm of a power monitoring system based on a risk control rule engine, characterized in that, The method comprises: Multi-element risk control rule mining and lightweight decision encapsulation are performed on a target power system to serve as a benchmark rule library and to construct a risk control rule engine; Through interaction between an FPGA module and a front-end power monitoring component, power flow data are determined, data preprocessing and data environment reconstruction are performed, and a virtual risk space is updated, wherein the virtual risk space is synchronously updated according to collection of the power flow data; According to the risk control rule engine, autonomous perception and judgment based on rule meta-nodes are performed on the virtual risk space, risk control alarm information is generated, and power feedback management based on risk control rules is performed; The risk control rule engine and the FPGA module are embedded plug-ins of a power monitoring system.

2. The method of claim 1, wherein the method further comprises: The multi-element risk control rule mining and lightweight decision encapsulation performed on the target power system comprise: Mining a first rule, wherein the first rule is any risk control rule of the target power system; Performing training encapsulation based on autonomous perception and lightweight rule judgment on the first rule to serve as a first rule meta-node; Completing construction of an Nth rule meta-node to serve as a rule meta-node cluster, wherein N is a total number of rules of the target power system.

3. The method of claim 2, wherein the method further comprises: The construction of the risk control rule engine comprises: Determining rule grouping conditions, grouping the rule meta-node cluster according to the rule grouping conditions, and generating the benchmark rule library; Formulating risk control management and control logic based on the benchmark rule library and supervising training to convergence to construct the risk control rule engine.

4. The method of claim 3, wherein the risk control rule engine-based power monitoring system alarm is characterized by, The grouping of the rule meta-node cluster according to the rule grouping conditions comprises: Taking a risk control probability rule as a first group, a causal paradox rule as a second group, a probability collapse rule as a third group, and a feedback adjustment rule as a fourth group, wherein a risk control probability is defined based on a risk confidence, and probability collapse is a directional probability based on risk spatiotemporal evolution; Writing the first group into a first rule area of an imbalance detection rule sub-library, and writing the second group and the third group into a second rule area of the imbalance detection rule sub-library; Writing the fourth group into a rule area of a balance feedback rule sub-library; According to the imbalance detection rule sub-library and the balance feedback rule sub-library, the benchmark rule library is constituted.

5. The alarm method for a power monitoring system based on a risk control rule engine as described in claim 1, characterized in that, The determination of power flow data, the data preprocessing and the data environment reconstruction, and the updating of the virtual risk space comprise: Scanning the power flow data, performing multi-source data fusion and conversion, and determining a power panoramic state vector; Performing parallelized data cleaning and high-frequency feature extraction based on the power panoramic state vector, and determining a risk feature tensor; Mapping the risk feature tensor to a virtual risk space.

6. The method of claim 4, wherein the method further comprises: According to the risk control rule engine, the autonomous perception and judgment based on rule meta-nodes are performed on the virtual risk space, comprising: Through activation of the risk control rule engine, virtual risk space inspection based on rule meta-nodes is performed according to a first rule area in the imbalance detection rule sub-library; If there is a rule meta-node response, a risk control probability value based on the responding rule meta-node is determined; According to the risk control probability value, risk control management of the target power system is performed, wherein a risk control management dimension comprises risk control alarm and feedback adjustment.

7. The alarm method for a power monitoring system based on a risk control rule engine as described in claim 6, characterized in that, According to the risk control probability value, the risk control management of the target power system is performed, including: If the risk control probability value meets a preset probability value, a rule matching is performed in the balance feedback rule sub-library to determine a target feedback adjustment rule, wherein the preset probability value is a critical value based on risk confidence; According to the target feedback adjustment rule, a directional risk control feedback adjustment of the target power system is performed.

8. The alarm method for a power monitoring system based on a risk control rule engine as described in claim 7, characterized in that, According to the risk control probability value, the risk control management of the target power system is performed, including: If the risk control probability value does not meet the preset probability value, an associated rule matching and determination based on the causality paradox rule and the probability collapse rule is performed according to a second rule area of the imbalance detection rule sub-library to determine a joint determination result; The joint determination result is subjected to rule matching and feedback adjustment based on the balance feedback rule sub-library.

9. The alarm method for a power monitoring system based on a risk control rule engine as described in claim 1, characterized in that, After the power feedback management based on the risk control rule is performed, the method further includes: Temporary database storage of rule record data, wherein a rule record data contains an imbalance detection rule, a balance feedback rule and a rule response result; Based on the temporary database, periodic rule record data retrieval is performed to mine rule blind areas; According to the rule blind area, the risk control rule engine is updated and learned.

10. A power monitoring system alarm platform based on a risk control rule engine, characterized in that, The platform is used to implement the power monitoring system alarm method based on the risk control rule engine according to any one of claims 1 to 9, and the platform includes: A risk control rule engine construction unit is used to mine and encapsulate multi-element risk control rules for a target power system as a benchmark rule library and construct a risk control rule engine; A virtual risk space updating unit is used to determine power flow data and perform data preprocessing and data environment reconstruction through interaction of an FPGA module and a front-end power monitoring component to update a virtual risk space, wherein the virtual risk space is updated synchronously according to the collection of power flow data; A risk control alarm information generation unit is used to perform autonomous perception and determination of the virtual risk space based on rule meta-nodes according to the risk control rule engine to generate risk control alarm information and perform power feedback management based on the risk control rule; Wherein, the risk control rule engine and the FPGA module are embedded plug-ins of the power monitoring system.

Images