Early warning emergency processing system and method based on electric power operation

Abstract

The invention discloses an early warning emergency processing system and method based on electric power operation, and relates to the technical field of monitoring analysis, and the method comprises the steps: obtaining an electric power equipment operation state data set, and generating an electric power operation three-dimensional dynamic model; performing abnormal operation feature extraction on the power equipment operation state data set to obtain an equipment abnormal feature map, performing dynamic risk prediction on the equipment abnormal feature map in combination with environment monitoring data, and generating a power equipment risk level distribution map; performing potential safety hazard correlation analysis according to the power equipment risk level distribution diagram and the power operation three-dimensional dynamic model, and generating a power safety early warning optimization model; and performing emergency plan matching degree calculation based on the electric power safety early warning optimization model, generating a multi-stage emergency response scheme set, performing dynamic priority adjustment on the multi-stage emergency response scheme set, executing electric power operation emergency processing operation, and outputting an alarm signal to an alarm device. The method has the effect of improving the emergency processing efficiency.

Description

Technical Field

[0001] This application relates to the technical field of monitoring and analysis, and particularly relates to an early warning and emergency handling system and method based on power operation. Background Art

[0002] With the expansion of the scale of the power system and the improvement of equipment complexity, the safety of power operation faces severe challenges.

[0003] In related technologies, traditional power safety monitoring mainly relies on a single sensor (such as an infrared thermometer, a partial discharge detector) to obtain the status data of power equipment, and combines a threshold alarm mechanism for risk early warning. As a result, there are prominent problems in power operation safety protection, such as a high false alarm rate of early warning and a lag in emergency response, which seriously restricts the safe operation level of power operation and there is room for improvement. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, this application provides an early warning and emergency handling system and method based on power operation.

[0005] In a first aspect, this application provides an early warning and emergency handling method based on power operation, including the following steps: Step S1: Multidimensional data collection is carried out on the power operation site through a multi-source sensor array to obtain a power equipment operation status data set, and multi-source heterogeneous data fusion is performed on the power equipment operation status data set to generate a three-dimensional dynamic model of power operation; Step S2: Abnormal operation feature extraction is carried out on the power equipment operation status data set according to the deep residual network model to obtain an equipment abnormal feature map, and dynamic risk prediction is carried out on the equipment abnormal feature map in combination with environmental monitoring data to generate a power equipment risk level distribution map; Step S3: Safety hazard correlation analysis is carried out according to the power equipment risk level distribution map and the three-dimensional dynamic model of power operation, a power accident evolution prediction model is constructed, and dynamic analysis is carried out on the power accident evolution prediction model to generate a power safety early warning optimization model; Step S4: Based on the power safety early warning optimization model, the matching degree of the emergency plan is calculated to generate a multi-level emergency response plan set, and dynamic priority adjustment is carried out on the multi-level emergency response plan set according to real-time risk evolution data, and power operation emergency handling operations are performed based on the adjusted multi-level emergency response plan, and an alarm signal is output to the alarm device.

[0006] Preferably, step S1 includes the following steps: Step S11: Deploy a multi-source sensor array at key nodes of power equipment, carry out multidimensional data collection on the power operation site based on the multi-source sensor array, and confirm a power equipment operation status data set; Step S12: Align the time stamps and register the spatial coordinates of the power equipment operation status data set, and then construct a multi-source heterogeneous data fusion matrix; Step S13: Use a three-dimensional convolutional neural network to perform feature space mapping on the multi-source heterogeneous data fusion matrix to generate a three-dimensional dynamic model of power operation.

[0007] Preferably, step S2 includes the following steps: Step S21: Perform spatio-temporal benchmark unification processing on the power equipment operation status data set to generate a multi-modal feature tensor; Step S22: Construct a deep residual network model, and extract an abnormal feature map according to the deep residual network model and the multi-modal feature tensor; Step S23: Allocate risk weights to the abnormal feature map through an attention mechanism to generate a risk feature heat map; Step S24: Combine environmental monitoring data, where the environmental monitoring data includes environmental temperature and humidity data and wind speed and direction data, construct an environment-equipment association model, and generate a risk spatio-temporal distribution matrix according to the environment-equipment association model; Step S25: Perform dynamic risk assessment based on the risk feature heat map and the risk spatio-temporal distribution matrix to generate a risk level distribution map of power equipment.

[0008] Preferably, step S3 includes the following steps: Step S31: Identify the characteristics of high-risk areas according to the risk level distribution map of power equipment, and then generate an association knowledge graph of equipment status - environmental factors - historical accidents; Step S32: Perform power accident evolution prediction on the association knowledge graph of equipment status - environmental factors - historical accidents, and then construct an optimized power accident evolution prediction model; Step S33: Perform safety dynamic analysis on the three-dimensional dynamic model of power operation according to the power accident evolution prediction model to generate a power safety early warning model.

[0009] Preferably, step S33 includes the following steps: Step S331: Analyze the power accident evolution path of the three-dimensional dynamic model of power operation, and extract the evolution path data of key accident trigger nodes; Step S332: Based on the evolution path data of key accident trigger nodes, perform dynamic simulation of accident chain reactions on the three-dimensional dynamic model of power operation to generate a multi-scenario accident evolution dynamic data set; Step S333: Combine the multi-scenario accident evolution dynamic data set with environmental monitoring data to estimate the safety risk probability of the dynamic evolution process, and output risk probability distribution data; Step S334: Calculate the warning threshold of the entrained energy of the accident shock wave through a hydrodynamic simulation algorithm, and then construct an energy propagation attenuation model; Step S335: Perform coupling optimization of warning parameters according to the risk probability distribution data and the energy propagation attenuation model to generate an optimized power safety warning model.

[0010] Preferably, step S4 includes the following sub-steps: Step S41: Based on the risk probability distribution data and the energy propagation attenuation model in the optimized power safety warning model, construct an emergency plan matching degree feature matrix, and generate a quantified data set of plan matching degrees according to the emergency plan matching degree feature matrix; Step S42: According to the quantified data set of plan matching degrees and the preset response level division rules, perform multi-level response arrangement on the emergency plans in the emergency resource library through a fuzzy decision tree algorithm to generate a multi-level emergency response plan set; Step S43: Dynamically correct the priority of each emergency response plan in the multi-level emergency response plan set in combination with the dynamic risk gradient value in the real-time risk evolution data, and output a dynamic response priority sequence; Step S44: Based on the highest-level emergency response plan in the dynamic priority sequence, parse the emergency operation instruction set corresponding to the emergency response plan through a three-dimensional space coordinate mapping engine, drive the power operation terminal to execute the emergency operation instruction set, and output an alarm signal to the alarm device.

[0011] Preferably, after driving the power operation terminal to execute the emergency operation instruction set, it further includes: Establish an emergency treatment effect feedback mechanism, collect power equipment status feedback data after emergency treatment through an Internet of Things terminal; and adjust and update the optimized power safety warning model based on the power equipment status feedback data.

[0012] In a second aspect, the present application provides a warning and emergency treatment system based on power operation, including: A data collection module, configured to perform multi-dimensional data collection on the power operation site through a multi-source sensor array, obtain a power equipment operation status data set, and perform multi-source heterogeneous data fusion on the power equipment operation status data set to generate a three-dimensional dynamic model of power operation; A power equipment risk level distribution map acquisition module, configured to extract abnormal operation characteristics from the power equipment operation status data set according to a deep residual network model to obtain an equipment abnormal characteristic map, and perform dynamic risk prediction on the equipment abnormal characteristic map in combination with environmental monitoring data to generate a power equipment risk level distribution map; A power safety warning optimization model generation module, which is used to perform safety hazard correlation analysis based on the power equipment risk level distribution map and the three-dimensional dynamic model of power operation, construct a power accident evolution prediction model, and perform dynamic analysis on the power accident evolution prediction model to generate a power safety warning optimization model; An emergency handling module, which is used to calculate the matching degree of the emergency plan based on the power safety warning optimization model, generate a multi-level emergency response plan set, dynamically adjust the priority of the multi-level emergency response plan set according to real-time risk evolution data, perform power operation emergency handling operations based on the adjusted multi-level emergency response plan, and output an alarm signal to the alarm device.

[0013] In summary, the present application includes the following beneficial technical effects: The embodiment of the present application provides a warning emergency handling method based on power operation. By performing multi-source heterogeneous data fusion on the power equipment operation status data set, a three-dimensional dynamic model of power operation is generated, and abnormal operation characteristics of the power equipment operation status data set are extracted according to the deep residual network model to obtain an equipment abnormal feature map, and then a power equipment risk level distribution map is generated. Furthermore, safety hazard correlation analysis is performed based on the power equipment risk level distribution map and the three-dimensional dynamic model of power operation to generate a power safety warning optimization model, and the matching degree of the emergency plan is calculated according to the power safety warning optimization model to generate a multi-level emergency response plan set. The dynamic priority of the multi-level emergency response plan set is adjusted according to real-time risk evolution data, and power operation emergency handling operations are performed based on the adjusted multi-level emergency response plan, and an alarm signal is output to the alarm device. Therefore, the situation where the safety protection of power operation has problems such as high warning false alarm rate and prominent emergency response lag caused by relying on a single sensor to obtain power equipment status data and combining a threshold alarm mechanism for risk warning is effectively reduced, thereby effectively improving the warning emergency response efficiency of power operation. Description of the Drawings

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0015] Figure 1 It is a flowchart of the warning emergency handling method based on power operation in the embodiment of the present application.

[0016] Figure 2 It is a schematic diagram of the warning emergency handling system based on power operation in the embodiment of the present application. Detailed Embodiments

[0017] The following will further elaborate on this application in conjunction with the attached Figure 1-2 drawings. Embodiment 1

[0018] This application embodiment discloses a warning and emergency handling method based on power operations.

[0019] Referring to Figure 1 , a warning and emergency handling method based on power operations includes the following steps: Step S1: Use a multi-source sensor array to collect multi-dimensional data from the power operation site, obtain a power equipment operation status data set, and perform multi-source heterogeneous data fusion on the power equipment operation status data set to generate a three-dimensional dynamic model of power operations; Step S2: Extract abnormal operation characteristics from the power equipment operation status data set according to the deep residual network to obtain an equipment abnormal characteristic map, and combine environmental monitoring data to perform dynamic risk prediction on the equipment abnormal characteristic map to generate a power equipment risk level distribution map; Step S3: Perform safety hazard correlation analysis based on the power equipment risk level distribution map and the three-dimensional dynamic model of power operations, construct a power accident evolution prediction model, and perform dynamic analysis on the power accident evolution prediction model to generate a power safety warning optimization model; Step S4: Calculate the matching degree of the emergency plan based on the power safety warning optimization model to generate a multi-level emergency response plan set, dynamically adjust the priorities of the multi-level emergency response plan set according to real-time risk evolution data, and perform power operation emergency handling operations based on the adjusted multi-level emergency response plan, and output an alarm signal to the alarm device.

[0020] It should be noted that Step S1 includes the following steps: Step S11: Deploy a multi-source sensor array at key nodes of power equipment, collect multi-dimensional data from the power operation site based on the multi-source sensor array, and identify a power equipment operation status data set; Specifically, deploying a multi-source sensor array at key nodes of power equipment specifically includes: Install an infrared thermal imager at the transformer bushing to collect temperature field distribution data at a frequency of 0.5 Hz; Set a UHF partial discharge sensor at the breaker contact to collect discharge pulse waveforms in the 40 - 300 MHz frequency band; Install a three-axis vibration sensor on the outer shell of the power equipment, and set the sampling rate to 10 kHz to collect mechanical vibration spectra; Install a laser gas analyzer in the gas chamber of the gas-insulated switchgear to monitor the purity and decomposition product concentration of SF6 gas in real time; Step S12: Perform timestamp alignment and spatial coordinate registration on the power equipment operation status dataset, and then construct a multi-source heterogeneous data fusion matrix; Specifically, perform timestamp alignment and spatial coordinate registration on the power equipment operation status dataset, and then construct a multi-source heterogeneous data fusion matrix. Since the time starting points and frequencies of data collected by different sensors may be different, timestamp alignment is required first. For example, an infrared thermal imager collects data every 2 seconds, while a three-axis vibration sensor collects 10,000 times per second. By adjusting the timestamps, the data from different sensors can have a corresponding relationship in time. At the same time, considering that different sensors are installed at different positions on the power equipment, spatial coordinate registration is performed to associate the data collected by each sensor with the spatial position information of the equipment. After timestamp alignment and spatial coordinate registration, different types of data are organized according to certain rules to construct a multi-source heterogeneous data fusion matrix. The multi-source heterogeneous data fusion matrix contains various data from different sensors, providing a unified data structure for subsequent analysis; Step S13: Use a three-dimensional convolutional neural network to perform feature space mapping on the multi-source heterogeneous data fusion matrix to generate a three-dimensional dynamic model of power operation.

[0021] Specifically, use a three-dimensional convolutional neural network to perform feature space mapping on the multi-source heterogeneous data fusion matrix to generate a three-dimensional dynamic model of power operation. The three-dimensional convolutional neural network can automatically learn the features in the multi-source heterogeneous data fusion matrix and map them to a new feature space. During the training process, a large number of multi-source heterogeneous data fusion matrices are used as inputs, and key features in the data are extracted through operations such as convolution and pooling in the network. For example, by learning the hot spot area features in the temperature field distribution data, the fault features in the discharge pulse waveform data, and the abnormal frequency features in the mechanical vibration spectrum data, etc. After training, the three-dimensional convolutional neural network can generate a three-dimensional dynamic model of power operation that reflects the operation status of the power equipment according to the input multi-source heterogeneous data fusion matrix. The three-dimensional dynamic model of power operation can intuitively display the operation of the power equipment at different times and spaces, helping the staff to discover potential fault hazards in a timely manner.

[0022] It should be noted that Step S2 includes the following steps: Step S21: Perform spatio-temporal reference unification processing on the power equipment operation status dataset to generate a multi-modal feature tensor; Step S22: Construct a deep residual network model, and extract an abnormal feature map according to the deep residual network model and the multi-modal feature tensor; Step S23: Perform risk weight assignment on the abnormal feature map through an attention mechanism to generate a risk feature heat map; Step S24: Combine the environmental monitoring data, which includes environmental temperature and humidity data and wind speed and direction data, to construct an environment-equipment association model, and generate a risk spatio-temporal distribution matrix according to the environment-equipment association model; Step S25: Conduct dynamic risk assessment based on the risk characteristic heat map and the risk spatio-temporal distribution matrix to generate a power equipment risk level distribution map.

[0023] Specifically, in step S21: Perform spatio-temporal benchmark unification processing on the power equipment operation status data set to generate a multi-modal feature tensor. The power equipment operation status data set contains various data from different sensors, such as temperature field distribution, discharge pulse waveform, mechanical vibration spectrum, etc. The above data may have differences in time and space. For the convenience of subsequent analysis, spatio-temporal benchmark unification processing is required. In terms of time, calibrate the timestamps of the data collected by different sensors so that they have a unified starting time and time interval. For example, unify the data time of different devices such as infrared thermal imagers and ultra-high frequency partial discharge sensors to the second-level accuracy. In terms of space, associate the data with the spatial coordinates of the equipment according to the installation positions of the sensors on the power equipment. After the spatio-temporal benchmark unification processing, organize the different-modal data into a multi-modal feature tensor according to certain rules. For example, take the temperature data, vibration data, etc. as different dimensions of the tensor to form a multi-dimensional array. This multi-modal feature tensor contains multi-faceted information about the operation status of the power equipment; Step S22: Construct a deep residual network model, and extract an abnormal feature map according to the deep residual network model and the multi-modal feature tensor. Among them, the deep residual network model is a powerful deep learning model that can process complex high-dimensional data. When constructing the deep residual network model, determine parameters such as the number of network layers, the number of nodes, and the activation function. Input the multi-modal feature tensor into the deep residual network model. The network automatically learns the features in the data through operations such as multi-layer convolution, pooling, and residual connection. During the learning process, the deep residual network model will identify the feature patterns in the normal operation state. When there are parts in the input data that are different from the normal pattern, it will extract abnormal features. For example, when there are abnormal frequency components in the mechanical vibration spectrum of the power equipment, the deep residual network model can capture these abnormalities and present them in the form of a map to form an abnormal feature map, intuitively showing the possible abnormalities of the equipment; Step S23: Perform risk weight assignment on the abnormal feature map through the attention mechanism to generate a risk feature heat map. The attention mechanism is a method that enables the model to focus on important information. For the abnormal feature map, analyze the importance of each feature through the attention mechanism and assign corresponding risk weights to different abnormal features. For example, for the abnormal feature of the discharge pulse waveform related to the insulation performance of electrical equipment, a higher risk weight is assigned because insulation problems may lead to serious equipment failures; while for some minor abnormal features of temperature fluctuations, a lower risk weight is assigned. Generate a risk feature heat map according to the assigned risk weights. In the heat map, the darker the color, the higher the risk weight, that is, the greater the impact of the abnormal features in this area on the operation risk of electrical equipment; Step S24: Combine the environmental monitoring data, which includes environmental temperature and humidity data and wind speed and direction data, to construct an environment-equipment association model. Generate a risk spatio-temporal distribution matrix based on the environment-equipment association model. Environmental factors have an important impact on the operation state of electrical equipment. Combine the environmental temperature and humidity data and wind speed and direction data with the operation state data of electrical equipment to construct an environment-equipment association model. In the embodiments of the present application, methods such as regression analysis and machine learning can be used to establish the environment-equipment association model. For example, through the analysis of historical data, it is found that when the environmental temperature is too high, the temperature of the transformer winding of electrical equipment will increase accordingly, thereby increasing the risk of equipment overheating. Calculate the impact of environmental factors on the operation risk of equipment at different times and spaces according to the environment-equipment association model to generate a risk spatio-temporal distribution matrix. The risk spatio-temporal distribution matrix records the risk situations faced by electrical equipment at different time points and different geographical locations; Step S25: Perform dynamic risk assessment based on the risk feature heat map and the risk spatio-temporal distribution matrix to generate a risk level distribution map of electrical equipment. Comprehensively consider the information in the risk feature heat map and the risk spatio-temporal distribution matrix, and perform dynamic risk assessment on electrical equipment. Combine the risk weights in the risk feature heat map with the spatio-temporal information in the risk spatio-temporal distribution matrix to calculate the comprehensive risk value of electrical equipment at different times and spaces. According to the size of the comprehensive risk value, divide the operation area of electrical equipment into different risk levels, such as low risk, medium risk, high risk, etc. Finally, generate a risk level distribution map of electrical equipment, and use different colors or symbols in the map to represent different risk levels, intuitively showing the risk distribution of electrical equipment in the entire monitoring area, providing a decision-making basis for the maintenance and management of electrical equipment.

[0024] It should be noted that step S3 includes the following steps: Step S31: Identify the characteristics of high-risk areas based on the risk level distribution map of electrical equipment, and then generate an association knowledge graph of equipment status - environmental factors - historical accidents; Step S32: Perform power accident evolution prediction on the associated knowledge graph of equipment status - environmental factors - historical accidents, and then construct an optimized model for power accident evolution prediction; Step S33: Conduct safety dynamic analysis on the three-dimensional dynamic model of power operation according to the power accident evolution prediction model, and generate a power safety warning model.

[0025] Specifically, in step S31: Identify the characteristics of high-risk areas based on the power equipment risk level distribution map, and then generate an associated knowledge graph of equipment status - environmental factors - historical accidents. The power equipment risk level distribution map visually presents the risk levels of power equipment in different regions. Through in-depth analysis of the power equipment risk level distribution map, the characteristics of high-risk areas can be identified. For example, in the power equipment risk level distribution map, it is found that power equipment in certain regions frequently shows high risk levels. By further exploring the status of power equipment in the above regions, it may be found that the equipment has serious aging, frequent partial discharge, etc.; at the same time, combining with environmental monitoring data, it may be found that the environmental temperature and humidity in the above high-risk areas have been in a range that is not conducive to equipment operation for a long time, or the wind speed is relatively high. Referring to historical accident records, check whether similar power accidents have occurred in the above regions and the specific circumstances of the accidents. Integrate the relevant information of power equipment status, environmental factors, and historical accidents to construct an associated knowledge graph. In the knowledge graph, equipment status information (such as transformer temperature, discharge condition of the circuit breaker, etc.), environmental factors (temperature, humidity, wind speed and direction, etc.), and historical accidents (accident time, cause, scope of influence, etc.) are interconnected through various relationships to form a comprehensive knowledge network. For example, through the knowledge graph, it can be clearly seen that under a certain specific environmental condition, a certain state of power equipment is likely to trigger a specific type of historical accident, thus providing a rich knowledge basis for subsequent analysis.

[0026] Step S32: Perform power accident evolution prediction on the associated knowledge graph of equipment status - environmental factors - historical accidents, and then construct an optimized power accident evolution prediction model. Based on the constructed associated knowledge graph of equipment status - environmental factors - historical accidents, methods such as data analysis and machine learning are used for power accident evolution prediction. In the embodiments of the present application, a graph neural network algorithm can be utilized to analyze the interactions and influences among various factors in the knowledge graph and predict the possible development trends of power accidents under different conditions. Suppose the associated knowledge graph of equipment status - environmental factors - historical accidents shows that when the degree of equipment aging reaches a certain threshold and the environmental humidity is high, accidents of equipment insulation breakdown have occurred. Through algorithm learning of the association relationship, the possibility of future occurrence of similar accidents and the possible evolution paths of the accidents can be predicted. For example, the accident may first start to deteriorate from a certain key component of the power equipment and then gradually affect the entire equipment system. According to the prediction results, the model is optimized by adjusting the parameters and structure of the model to enable it to more accurately predict the evolution of power accidents. For instance, continuously adding new historical accident data and real-time monitoring data to update the knowledge base of the model and improve the prediction accuracy of the model, and finally constructing an optimized power accident evolution prediction model.

[0027] Step S33: Perform safety dynamic analysis on the three-dimensional dynamic model of power operation according to the power accident evolution prediction model to generate a power safety warning model. The three-dimensional dynamic model of power operation visually displays the operating status of power equipment and the operating environment. By combining the power accident evolution prediction model with the three-dimensional dynamic model of power operation, the equipment status and environmental factors in the three-dimensional dynamic model of power operation are monitored and analyzed in real time. For example, in the three-dimensional dynamic model of power operation, parameters such as the temperature and vibration of the equipment and information such as the temperature, humidity, and wind speed of the environment are displayed in real time. The power accident evolution prediction model analyzes whether there are potential safety risks in the current operating status of the equipment and how these risks may evolve based on the above real-time data and its own prediction algorithm. If the prediction model finds that the temperature of a certain component of the equipment rises abnormally and, according to historical data and the knowledge graph, this situation may lead to equipment failure, then a corresponding safety warning will be issued. By continuously performing safety dynamic analysis on the three-dimensional dynamic model of power operation and adjusting the warning strategy according to the analysis results, a power safety warning model is finally generated. The power safety warning model can issue safety warnings at different levels in a timely and accurate manner according to the real-time status of the equipment and environmental changes, reminding the staff to take corresponding measures to prevent the occurrence of power accidents.

[0028] Further, step S33 includes the following steps: Step S331: Perform power accident evolution path analysis on the three-dimensional dynamic model of power operation to extract the evolution path data of key accident trigger nodes; Step S332: Based on the evolution path data of critical accident trigger nodes, perform dynamic simulation of accident chain reactions on the three-dimensional dynamic model of power operation to generate a multi-scenario accident evolution dynamic data set; Step S333: Combine the multi-scenario accident evolution dynamic data set with environmental monitoring data to estimate the safety risk probability of the dynamic evolution process and output risk probability distribution data; Step S334: Calculate the warning threshold of the entrained energy of the accident shock wave through the fluid mechanics simulation algorithm, and then construct an energy propagation attenuation model; Step S335: Perform coupling optimization of warning parameters according to the risk probability distribution data and the energy propagation attenuation model to generate an optimized power safety warning model.

[0029] Specifically, in step S331: Analyze the power accident evolution path of the three-dimensional dynamic model of power operation, and extract the evolution path data of critical accident trigger nodes. The three-dimensional dynamic model of power operation contains rich information such as the real-time operating status, spatial layout, and operation environment of power equipment. By analyzing the three-dimensional dynamic model of power operation, simulate the possible occurrence and development process of power accidents. For example, in the transformer area, if it is monitored that the oil temperature continues to rise and exceeds the normal range, and at the same time the local discharge phenomenon intensifies, then this may be a critical accident trigger node. Through the study of historical data and relevant knowledge graphs, combined with real-time monitoring data, analyze the evolution path of the accident starting from this node, including the equipment components that may be affected subsequently, the direction of accident diffusion, etc. Extract the relevant data of the above critical accident trigger nodes, such as trigger time, initial state parameters, subsequent state changes, etc., to form evolution path data, providing a basis for subsequent analysis; Step S332: Based on the evolution path data of critical accident trigger nodes, perform dynamic simulation of accident chain reactions on the three-dimensional dynamic model of power operation to generate a multi-scenario accident evolution dynamic data set. Use the evolution path data obtained in step S331 to simulate the accident chain reaction in the three-dimensional dynamic model of power operation. Considering the complexity of the power system, an accident may trigger a series of chain reactions, affecting multiple devices and regions. For example, a transformer failure may cause abnormal voltage, which in turn affects the connected circuit breakers and other electrical equipment. By setting different initial conditions and parameters, simulate multiple possible accident scenarios. For example, assume different severity levels of transformer failures or changes in environmental conditions (such as temperature, humidity), and simulate the evolution process of the accident in the above cases respectively. Record the development process of the accident, the equipment involved, state changes, etc. in each scenario to generate a multi-scenario accident evolution dynamic data set; Step S333: Combine the multi-scenario accident evolution dynamic dataset with environmental monitoring data to estimate the safety risk probability of the dynamic evolution process, and output risk probability distribution data. Combine various accident scenarios in the multi-scenario accident evolution dynamic dataset with real-time environmental monitoring data (such as environmental temperature and humidity, wind speed and direction, etc.). Use probability statistics methods and machine learning algorithms to analyze the possibility of various accident scenarios occurring under different environmental conditions. For example, through the analysis of historical data, it is found that in a high-temperature and high-humid environment, the insulation performance of power equipment decreases, and the probability of insulation breakdown accidents increases. According to the characteristics of different scenarios in the multi-scenario accident evolution dynamic dataset and the real-time values of environmental monitoring data, calculate the probability of accidents occurring in each scenario, and finally output risk probability distribution data. The risk probability distribution data reflects the safety risk probability situation faced by the power system under different accident scenarios and environmental conditions; Step S334: Calculate the warning threshold of the entrained energy of the accident shock wave through a hydrodynamic simulation algorithm, and then construct an energy propagation attenuation model. When a power accident occurs, an accident shock wave may be generated, and the energy it entrains will pose a threat to surrounding equipment and personnel. Use the hydrodynamic simulation algorithm to simulate the propagation process of the accident shock wave in the power operation environment, considering the influence of environmental factors (such as air density, obstacles, etc.) on the shock wave propagation, calculate the entrained energy of the shock wave at different positions, and determine a reasonable warning threshold based on the calculation results, that is, when the entrained energy of the shock wave reaches or exceeds this threshold, a warning needs to be issued. At the same time, through the analysis of the shock wave energy at different distances and positions, construct an energy propagation attenuation model to describe the law of energy attenuation as the propagation distance increases; Step S335: Coupling and optimizing the warning parameters according to the risk probability distribution data and the energy propagation attenuation model to generate a power safety warning optimization model. Comprehensively consider the risk probability distribution data obtained in Step S333 and the energy propagation attenuation model constructed in Step S334 to optimize the warning parameters. For example, according to the risk probability of different accident scenarios, adjust the relevant parameters in the energy propagation attenuation model to make the warning more accurate and timely. If the probability of a certain accident scenario occurring is relatively high, and the energy of the shock wave it generates may cause serious impacts on key equipment, then correspondingly lower the warning threshold in this scenario. By continuously adjusting and optimizing the warning parameters, generate a power safety warning optimization model. The power safety warning optimization model can accurately issue safety warnings based on the real-time operating status of power equipment, environmental conditions, and the possibility of accident evolution, providing a strong guarantee for the safe operation of the power system.

[0030] It should be noted that Step S4 includes the following sub-steps: Step S41: Based on the risk probability distribution data and the energy propagation attenuation model in the power safety warning optimization model, construct an emergency plan matching degree feature matrix, and generate a plan matching degree quantization data set according to the emergency plan matching degree feature matrix; Step S42: According to the plan matching degree quantization data set and the preset response level division rules, perform multi-level response arrangement on the emergency plans in the emergency resource library through the fuzzy decision tree algorithm, and generate a multi-level emergency response plan set; Step S43: Combine the dynamic risk gradient value in the real-time risk evolution data to dynamically correct the priorities of the emergency response plans in the multi-level emergency response plan set, and output a dynamic response priority sequence; Step S44: Based on the emergency response plan with the highest level in the dynamic priority sequence, parse the emergency operation instruction set corresponding to the emergency response plan through the three-dimensional space coordinate mapping engine, drive the power operation terminal to execute the emergency operation instruction set, and output an alarm signal to the alarm device.

[0031] Specifically, in step S41: Based on the risk probability distribution data and the energy propagation attenuation model in the power safety warning optimization model, construct an emergency plan matching degree feature matrix, and generate a plan matching degree quantization data set according to the emergency plan matching degree feature matrix. The risk probability distribution data in the power safety warning optimization model reflects the possibility of different accident scenarios occurring, and the energy propagation attenuation model describes the propagation and attenuation of energy after an accident occurs. Combining the above two key pieces of information, analyze the matching degree of each emergency plan in the emergency plan library with different risk scenarios. For example, for a high-risk scenario that may be caused by a transformer failure, its risk probability is relatively high and it may generate a strong accident shock wave. Some emergency plans may focus more on dealing with transformer failures and can effectively cope with the impact brought by the shock wave, while other plans may be weaker in this regard. By evaluating the performance of each plan in dealing with different risk characteristics (such as fault type, energy influence range, etc.), construct an emergency plan matching degree feature matrix. Each row in the matrix represents an emergency plan, each column represents a risk characteristic, and the matrix element represents the matching degree score of the plan corresponding to the risk characteristic. According to the above emergency plan matching degree feature matrix, further organize and quantify the relevant data to generate a plan matching degree quantization data set, providing a basis for subsequent plan arrangement; Step S42: According to the pre-plan matching degree quantization data set and the preset response level division rules, use the fuzzy decision tree algorithm to perform multi-level response arrangement on the pre-plans in the emergency resource library, and generate a multi-level emergency response plan set. The preset response level division rules are usually formulated according to the severity and impact scope of risks. For example, they are divided into three levels: low, medium, and high. Using the fuzzy decision tree algorithm, with the pre-plan matching degree quantization data set as the input, combined with the response level division rules, classify and arrange the emergency pre-plans. The fuzzy decision tree algorithm can handle the uncertainty and ambiguity in the data and more accurately evaluate the suitability of the pre-plans for different risk levels. For example, for a pre-plan with a high matching degree and can effectively respond to high-risk scenarios, arrange it into the high response level; for a pre-plan with a moderate matching degree and can respond to medium risks, arrange it into the medium response level; for a pre-plan mainly for low-risk scenarios, arrange it into the low response level. Through the above multi-level response arrangement, generate an emergency response plan set containing different response levels, so that there are corresponding pre-plans to choose from when facing risks of different levels; Step S43: Combine the dynamic risk gradient value in the real-time risk evolution data to dynamically correct the priorities of each emergency response plan in the multi-level emergency response plan set, and output a dynamic response priority sequence. The real-time risk evolution data reflects the real-time changes of risks in the power system, and the dynamic risk gradient value in it represents the speed and trend of risk changes. According to the dynamic risk gradient value, adjust the priorities of each plan in the multi-level emergency response plan set. For example, if the current dynamic risk gradient value shows that the risk is increasing rapidly, the priority of a certain plan originally belonging to the medium response level may be increased because of the timeliness and effectiveness of its response measures; while if the effect of a certain plan may be inferior to other plans under the current risk change situation, its priority will be reduced. By continuously adjusting the priorities of each plan according to the real-time risk evolution data, output a dynamic response priority sequence to ensure that the most suitable emergency response plan for the current risk situation can be preferentially executed at any time; Step S44: Based on the emergency response plan with the highest level in the dynamic priority sequence, parse the emergency operation instruction set corresponding to the emergency response plan through the three-dimensional space coordinate mapping engine, drive the power operation terminal to execute the emergency operation instruction set, and output an alarm signal to the alarm device. In the dynamic response priority sequence, select the emergency response plan with the highest level. The three-dimensional space coordinate mapping engine can map various operation instructions in the emergency response plan to the actual spatial positions and states of power equipment. For example, the emergency response plan may include operation instructions such as cutting off the power supply at a specific location and starting standby equipment. The three-dimensional space coordinate mapping engine will parse these instructions, determine the three-dimensional space coordinates of the corresponding power equipment, and drive the power operation terminal (such as intelligent switches, automated control systems, etc.) to perform corresponding operations according to the instructions to address safety risks in the power system. At the same time, output an alarm signal to the alarm device to notify relevant personnel of the current risk situation and the emergency operations being carried out, so that they can take further measures or conduct supervision in a timely manner.

[0032] Further, after driving the power operation terminal to execute the emergency operation instruction set, it also includes: Establish an emergency treatment effect feedback mechanism, collect power equipment status feedback data after emergency treatment through the Internet of Things terminal; and adjust and update the power safety early warning optimization model based on the power equipment status feedback data.

[0033] Specifically, establish an emergency treatment effect feedback mechanism. The core of the emergency treatment effect feedback mechanism lies in using the Internet of Things terminal to achieve real-time and accurate collection of the status of power equipment. The Internet of Things terminal, as a bridge connecting power equipment and the data processing system, is widely deployed at various key positions of power equipment. For example, various sensors such as temperature sensors, vibration sensors, and current sensors are installed on equipment such as transformers, circuit breakers, and switch cabinets. The sensors can monitor the operating parameters of the equipment in real time, such as temperature, vibration frequency, and current magnitude. When the emergency treatment measures are executed, the Internet of Things terminal starts to work and continuously collects power equipment status feedback data. For example, after a power accident occurs and emergency operations such as cutting off the power supply of some circuits and starting the standby power supply are carried out, the Internet of Things terminals installed on the relevant equipment will collect various parameters of the equipment under the new power supply state. The temperature sensor of the transformer will feedback the temperature change of the transformer when it is powered by the standby power supply, and the current sensor will monitor whether the current of the equipment is stable within the normal range, etc.; Process and analyze the power equipment status feedback data collected. Since the amount of data collected is huge and may contain noise, preprocessing operations such as data cleaning and filtering are required to remove outliers and interference factors to ensure the accuracy and reliability of the data. For example, for the data collected by temperature sensors, if there are instantaneous extreme temperature values that do not match the data of other sensors, it may be caused by sensor failures or external interference, and these abnormal data need to be excluded. After preprocessing the data, use data analysis algorithms and models for in-depth analysis. Statistical analysis methods can be used to calculate statistics such as the mean and standard deviation of various device parameters to understand the overall change trend of the power equipment status; machine learning algorithms can also be used to classify and predict the device status to determine whether the device has returned to normal operation or there are potential failure risks; Based on the analysis results of the power equipment status feedback data, adjust and update the power safety warning optimization model. The power safety warning optimization model is a complex system that includes multiple sub-models such as a risk probability distribution model and an energy propagation attenuation model. According to the feedback data, if it is found that there is a large deviation between the actual occurrence probability of some risk scenarios and the probability predicted by the model, the risk probability distribution model needs to be adjusted. For example, after emergency treatment, it is found that although the predicted probability of a specific equipment failure scenario in the model is low, it actually occurred and had a greater impact. At this time, it is necessary to re-analyze the relevant factors of the above failure scenario, such as the aging degree of the equipment and environmental conditions, and adjust the probability parameters of the above scenario in the model. For the energy propagation attenuation model, if the feedback data shows that the propagation range and energy attenuation of the accident shock wave do not match the model prediction, the energy propagation attenuation model also needs to be corrected. For example, it is found that the actual propagation distance of the shock wave is farther than the model prediction, which may be due to insufficient consideration of environmental factors (such as building layout, air density, etc.) in the energy propagation attenuation model, and the relevant parameters in the energy propagation attenuation model need to be updated. Embodiment 2

[0034] The embodiment of the present application also discloses a warning and emergency processing system based on power operations.

[0035] Refer to Figure 2 , a warning and emergency processing system based on power operations, including: A data acquisition module, used to perform multi-dimensional data acquisition on the power operation site through a multi-source sensor array, obtain a power equipment operation status data set, and perform multi-source heterogeneous data fusion on the power equipment operation status data set to generate a three-dimensional dynamic model of power operations; The power equipment risk level distribution map acquisition module is used to extract abnormal operation characteristics from the power equipment operation status data set according to the deep residual network model, obtain the equipment abnormal feature map, and perform dynamic risk prediction on the equipment abnormal feature map in combination with the environmental monitoring data to generate the power equipment risk level distribution map; The power safety early warning optimization model generation module is used to perform safety hazard correlation analysis according to the power equipment risk level distribution map and the three-dimensional dynamic model of power operation, construct a power accident evolution prediction model, and perform dynamic analysis on the power accident evolution prediction model to generate a power safety early warning optimization model; The emergency handling module is used to calculate the matching degree of the emergency plan based on the power safety early warning optimization model, generate a multi-level emergency response plan set, perform dynamic priority adjustment on the multi-level emergency response plan set according to the real-time risk evolution data, perform power operation emergency handling operations based on the adjusted multi-level emergency response plan, and output an alarm signal to the alarm device.

[0036] The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of this technology can make various modifications or supplements to the described specific embodiments or use similar methods to replace them. As long as they do not deviate from the concept of the invention, they should fall within the protection scope of the present invention.

[0037] In the description of this specification, the description with reference to terms such as "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0038] The above-disclosed preferred embodiments of the present invention are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the present invention to only the specific implementation manners. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and utilize the present invention well.

Claims

1. A warning and emergency handling method based on electric power operations, characterized in that, Including the following steps: Step S1: Multidimensional data acquisition of the power operation site is carried out through a multi-source sensor array to obtain a dataset of the operating state of power equipment, and multi-source heterogeneous data fusion is performed on the dataset of the operating state of power equipment to generate a three-dimensional dynamic model of power operation; Step S2: Extract abnormal operation features from the dataset of the operating state of power equipment according to the deep residual network model to obtain an equipment abnormal feature map, and perform dynamic risk prediction on the equipment abnormal feature map in combination with environmental monitoring data to generate a risk level distribution map of power equipment; Step S3: Conduct safety hazard correlation analysis according to the risk level distribution map of power equipment and the three-dimensional dynamic model of power operation, construct an electric power accident evolution prediction model, and perform dynamic analysis on the electric power accident evolution prediction model to generate an optimized model for electric power safety warning; Step S4: Calculate the matching degree of the emergency plan based on the optimized model for electric power safety warning to generate a multi-level emergency response plan set, dynamically adjust the priority of the multi-level emergency response plan set according to real-time risk evolution data, and perform emergency handling operations for power operation based on the adjusted multi-level emergency response plan, and output an alarm signal to the alarm device.

2. The early warning and emergency handling method based on power operation according to claim 1, wherein, Step S1 includes the following steps: Step S11: Deploy a multi-source sensor array at key nodes of power equipment, carry out multidimensional data acquisition of the power operation site based on the multi-source sensor array, and confirm a dataset of the operating state of power equipment; Step S12: Perform timestamp alignment and spatial coordinate registration processing on the dataset of the operating state of power equipment, and then construct a multi-source heterogeneous data fusion matrix; Step S13: Use a three-dimensional convolutional neural network to perform feature space mapping on the multi-source heterogeneous data fusion matrix to generate a three-dimensional dynamic model of power operation.

3. The early warning and emergency handling method based on power operation according to claim 1, wherein, Step S2 includes the following steps: Step S21: Perform spatio-temporal benchmark unification processing on the dataset of the operating state of power equipment to generate a multi-modal feature tensor; Step S22: Construct a deep residual network model, and extract an abnormal feature map according to the deep residual network model and the multi-modal feature tensor; Step S23: Perform risk weight assignment on the abnormal feature map through an attention mechanism to generate a risk feature heat map; Step S24: Combine environmental monitoring data, where the environmental monitoring data includes environmental temperature and humidity data and wind speed and direction data, construct an environment-equipment association model, and generate a risk spatio-temporal distribution matrix according to the environment-equipment association model; Step S25: Perform dynamic risk assessment based on the risk feature heat map and the risk spatio-temporal distribution matrix to generate a risk level distribution map of power equipment.

4. The early warning and emergency handling method based on electric power operation according to claim 1, characterized in that Step S3 includes the following steps: Step S31: Identify the characteristics of high-risk areas according to the risk level distribution map of power equipment, and then generate an association knowledge graph of equipment status - environmental factors - historical accidents; Step S32: Perform electric power accident evolution prediction on the association knowledge graph of equipment status - environmental factors - historical accidents, and then construct an optimized model for electric power accident evolution prediction; Step S33: Perform safety dynamic analysis on the three-dimensional dynamic model of power operation according to the electric power accident evolution prediction model to generate an electric power safety warning model.

5. A warning and emergency handling method based on electric power operation according to claim 4, characterized in that, Step S33 includes the following steps: Step S331: Conduct an analysis of the evolution path of power accidents on the three-dimensional dynamic model of power operations, and extract the evolution path data of key accident trigger nodes; Step S332: Based on the evolution path data of key accident trigger nodes, conduct a dynamic simulation of the accident chain reaction on the three-dimensional dynamic model of power operations to generate a multi-scenario accident evolution dynamic data set; Step S333: Combine the multi-scenario accident evolution dynamic data set with environmental monitoring data to estimate the probability of safety risks in the dynamic evolution process, and output risk probability distribution data; Step S334: Calculate the warning threshold of the entrained energy of the accident shock wave through a fluid mechanics simulation algorithm, and then construct an energy propagation attenuation model; Step S335: Perform coupling optimization of warning parameters according to the risk probability distribution data and the energy propagation attenuation model to generate an optimized power safety warning model.

6. The early warning and emergency handling method based on power operation according to claim 5, characterized in that, Step S4 includes the following sub-steps: Step S41: Based on the risk probability distribution data and the energy propagation attenuation model in the optimized power safety warning model, construct an emergency plan matching degree feature matrix, and generate a plan matching degree quantization data set according to the emergency plan matching degree feature matrix; Step S42: According to the plan matching degree quantization data set and the preset response level division rules, perform multi-level response arrangement on the emergency plans in the emergency resource library through a fuzzy decision tree algorithm to generate a multi-level emergency response plan set; Step S43: Dynamically correct the priority of each emergency response plan in the multi-level emergency response plan set in combination with the dynamic risk gradient value in the real-time risk evolution data, and output a dynamic response priority sequence; Step S44: Based on the highest-level emergency response plan in the dynamic priority sequence, parse the emergency operation instruction set corresponding to the emergency response plan through a three-dimensional space coordinate mapping engine, drive the power operation terminal to execute the emergency operation instruction set, and output an alarm signal to the alarm device.

7. The early warning and emergency handling method based on electric power operation according to claim 6, characterized in that, After driving the power operation terminal to execute the emergency operation instruction set, it also includes: Establish an emergency treatment effect feedback mechanism, collect power equipment status feedback data after emergency treatment through an Internet of Things terminal; and adjust and update the optimized power safety warning model based on the power equipment status feedback data.

8. An early warning and emergency handling system based on electric power operations, which is applied to an early warning and emergency handling method based on electric power operations according to any one of the above claims 1-7, characterized in that, It includes: A data acquisition module, which is used to collect multi-dimensional data on the power operation site through a multi-source sensor array, obtain a power equipment operation status data set, and perform multi-source heterogeneous data fusion on the power equipment operation status data set to generate a three-dimensional dynamic model of power operations; A power equipment risk level distribution map acquisition module, which is used to extract abnormal operation characteristics from the power equipment operation status data set according to a deep residual network model to obtain an equipment abnormal characteristic map, and perform dynamic risk prediction on the equipment abnormal characteristic map in combination with environmental monitoring data to generate a power equipment risk level distribution map; An optimized power safety warning model generation module, which is used to conduct a safety hazard correlation analysis based on the power equipment risk level distribution map and the three-dimensional dynamic model of power operations, construct a power accident evolution prediction model, and perform dynamic analysis on the power accident evolution prediction model to generate an optimized power safety warning model; An emergency handling module, which is used to calculate the matching degree of emergency plans based on the power safety early warning optimization model, generate a multi-level emergency response plan set, dynamically adjust the priorities of the multi-level emergency response plan set according to real-time risk evolution data, perform emergency handling operations for power operations based on the adjusted multi-level emergency response plan, and output an alarm signal to the alarm device.

Images