Line loss simulation application method and system based on intelligent fault diagnosis technology

Abstract

The invention discloses a line loss simulation application method and system based on an intelligent fault diagnosis technology, and relates to the technical field of intelligent operation and maintenance of a power system, and the method comprises the steps: obtaining original data through a power grid collection system, and carrying out the first feature extraction of the original data; inputting the first feature extraction result, and constructing a first diagnosis model; building a first simulation system, and carrying out line loss checking processing; and deploying the first diagnosis model into the simulation system for fault diagnosis. According to the method, the problems that the number of real transformer areas is large, and a transformer area topological line cannot be truly simulated are solved; through a virtual-real combination mode, any one-dimensional electric energy meter in a virtual environment can be mapped to a real object, so that an experimenter or a trained student can actually carry out related electrical measurement through a tool; the virtual electric energy meter and the real electric energy meter completely simulate field faults and phenomena generated by the faults.

Description

Technical Field

[0001] The present invention relates to the field of intelligent operation and maintenance technology of power systems, and specifically to a line loss simulation application method and system based on embodied intelligent fault diagnosis technology. Background Art

[0002] With the rapid development of smart grids, the widespread deployment of photovoltaic energy, electric vehicle charging piles and high-efficiency energy storage technologies, the intelligent transformation and upgrading of low-voltage distribution substations as the "nerve endings" of power transmission has become a key link in improving the quality of power supply. Under traditional management methods, problems such as information asymmetry, slow response speed and low efficiency of troubleshooting have become increasingly prominent, seriously restricting the operational efficiency and power supply safety of the power system. What is particularly important is that line loss is a key indicator that directly reflects the economic operation efficiency of the power grid. Its high-level management is of great significance for power companies to reduce costs and increase efficiency and ensure the stable operation of the system.

[0003] However, under the traditional management system, hidden problems such as electricity theft and abnormal low-voltage terminal equipment are difficult to detect in time, resulting in high line loss rates, which not only erode corporate profits but also weaken the overall efficiency and reliability of the power grid. Therefore, how to effectively monitor and optimize the line loss rate has become an urgent need to improve the stability and reliability of power supply.

[0004] Therefore, the various research units and grassroots units of the power company will continue to study and analyze various abnormalities in the substation, explore on-site governance solutions, simulate the actual substation in the laboratory through the device, simulate various abnormalities of the electric energy meter and the line, and explore rapid troubleshooting and positioning solutions. However, this pure hardware solution has relative limitations. The device contains limited metering equipment, and there are many types of line loss substations that cannot be fully simulated on the device. Currently, many physical simulation electric energy meters can simulate few types of faults and cannot fully reproduce abnormal power consumption scenarios. Not only are the simulated fault scenarios and substation types limited, but the scenes cannot be automatically switched, and there is a poor human-computer interaction experience. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing devices contain limited metering equipment, and many types of line loss areas cannot be fully simulated on the devices. Currently, many physical simulation electricity meters can simulate relatively few types of faults and cannot fully reproduce the problem of abnormal electricity consumption scenarios.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a line loss simulation application method based on embodied intelligent fault diagnosis technology, comprising:

[0008] Use the power grid acquisition system to obtain raw data and perform first feature extraction on the raw data;

[0009] Inputting the first feature extraction result to construct a first diagnostic model;

[0010] Build the first simulation system to investigate and handle line losses;

[0011] The first diagnostic model is deployed into the simulation system for fault diagnosis.

[0012] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the use of the power grid acquisition system to collect original data includes collecting substation line loss monitoring data through the power grid acquisition system and preprocessing to obtain the original data.

[0013] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the first feature extraction of the original data includes extracting features from the original data according to a time series to obtain a first feature extraction result.

[0014] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the construction of the first diagnostic model includes setting an input layer according to the result of the first feature extraction, and designing an output layer according to the line loss fault type.

[0015] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the construction of the first simulation system includes using line loss substation simulation software to build a dynamic embodied scene of line loss faults, and adjusting the line loss substation simulation device to simulate the real line loss substation scene.

[0016] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the deployment of the first diagnostic model to the simulation system includes converting the fault features extracted from the text based on the training of the big data language model into control signals, and performing control output for the virtual signal device;

[0017] Extract the data of substations with higher real line loss, extract the terminal information of low-voltage substations with higher real line loss, establish a substation line loss fault model, perform automatic simulation in the simulation software, and switch various fault modes and scenarios according to the line loss fault; select a static scenario from the intelligent line loss fault model embodied in the simulation software, and generate a dynamic scenario based on the probability of occurrence of various elements in different substations.

[0018] As a preferred solution of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in the present invention, wherein: the fault detection includes realizing automatic generation of line loss area scenes, the software extracts the features of various elements of the line loss area from the text according to the pre-trained large language model, and generates a digital model of the line loss area based on the features;

[0019] Extract the terminal information of low-voltage substations in substations with higher actual line loss, cooperate with on-site fault handling and on-site electricity theft evidence collection pictures, extract fault feature points, and input them into the intelligent perception model of line loss faults; use the large language model to generate respective fault phenomena and scenarios for each element in the substation line loss fault library and match the relevant probabilities; sample faults from the fault library according to the relevant probabilities of each element, calculate the real-time status of the dynamic embodied scenario according to the influence of each element of the line loss fault, and obtain the dynamic embodied scenario of the line loss fault.

[0020] Another object of the present invention is to provide a line loss simulation application system based on embodied intelligent fault diagnosis technology, which can solve the problem that the current substation line loss simulation technology has the inability to automatically switch scenes and has a poor human-computer interaction experience through one of the schemes of the substation line loss simulation system based on embodied intelligence.

[0021] As a preferred solution of the line loss simulation application system based on embodied intelligent fault diagnosis technology described in the present invention, it includes: an input module, a model building module, an optimization module, and a processing module;

[0022] The input module extracts feature descriptions based on real high-line loss substation data, including load rate, total number of households, distribution transformer capacity, low-voltage line length, power supply radius, substation attributes, and substation topology, builds a substation model based on refined digital modeling technology, and inputs the collected operation monitoring data, equipment life cycle data, and environmental data into the model;

[0023] The model building module extracts the characteristics of the line loss fault in the substation area according to the on-site line loss troubleshooting work order and the line loss detection data in the substation area, and establishes an intelligent model of the line loss fault based on the neural network;

[0024] The optimization module uses a large language model to analyze the elements in the line loss fault intelligent model, calculates the probability of finding line loss faults in on-site management, and improves the line loss fault model;

[0025] The processing module and the above models together constitute an embodied intelligent line loss fault model, select a static scene from the embodied intelligent line loss fault model, generate a dynamic scene according to the probability of occurrence of each element in different substations, and map the substation line loss characteristic data and the fault data of the low-voltage substation terminal to the control simulation device, and output the substation line loss nodes and fault data.

[0026] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a line loss simulation application method based on embodied intelligent fault diagnosis technology.

[0027] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a line loss simulation application method based on embodied intelligent fault diagnosis technology.

[0028] The beneficial effects of the present invention are as follows: the provided line loss simulation application method based on embodied intelligent fault diagnosis technology performs line loss simulation in the substation through a combination of virtual and real methods, which solves the problem that there are many real substations and the substation topology lines cannot be truly simulated; through a combination of virtual and real methods, any virtual DC energy meter can be mapped to the real object, allowing the experimenter or the trainee to actually perform relevant electrical measurements through tools; the virtual and real energy meters more completely simulate the on-site faults and the phenomena caused by the faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0030] Figure 1 An overall flow chart of a line loss simulation application method based on embodied intelligent fault diagnosis technology provided in the first embodiment of the present invention.

[0031] Figure 2 A composition diagram of an embodied intelligent simulation system for line loss in a substation, which provides a line loss simulation application method based on embodied intelligent fault diagnosis technology, for the third embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0033] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a line loss simulation application method based on embodied intelligent fault diagnosis technology, including:

[0034] S1: Use the power grid acquisition system to obtain raw data and perform first feature extraction on the raw data.

[0035] Furthermore, using the power grid collection system to collect raw data includes collecting substation line loss monitoring data through the power grid collection system and pre-processing to obtain the raw data.

[0036] In the embodiment of the present application, the original data may be the collected voltage, current, power factor, live line current, neutral line current, and user power consumption data.

[0037] Preprocessing can be: Time series arrangement: Arrange the line loss monitoring data in the power grid acquisition system, such as the voltage, current, power factor, live line current, neutral line current, user power consumption data, etc. collected daily, according to the time label in the collected data, that is, the time series. Missing value processing: Remove missing values ​​in the data, such as failed user data and non-check-in user data. Data division: Divide the data set into training set, validation set, and test set.

[0038] In the embodiment of the present application, the first feature extraction of the original data can be, time series features: extract features from the original data according to the time series, such as trend (compared with previous electricity consumption data, the user's electricity consumption suddenly dropped, the electricity consumption was normal during the on-site investigation, and the electricity consumption dropped after the investigation), seasonality (such as the line loss rate is large during the flood season), etc. Extract business features: Combined with the information in the on-site line loss investigation work order, extract business features related to the fault, such as the status of the metering equipment (the user privately connects the wires to bypass the electric energy meter and connects the electricity to the user's home, the current of the household line under the electric energy meter is >1.5 times the current of the electric energy meter outlet line, the system user load is low, there is a loss of pressure, loss of current, etc.), abnormal events of the metering equipment (opening the meter cover, opening the meter box, power outage events), etc.

[0039] In an optional embodiment, the first feature extraction of the original data may also be spatial feature extraction, that is, feature extraction based on the geographical location, physical distance and correlation relationship of the devices and nodes in the area. Specifically, it includes:

[0040] A1, extracts the geographic coordinates of metering equipment, transformers, and user locations to analyze the impact of the distance between devices on line loss. For example, nodes farther away may have higher line losses.

[0041] A2, builds the network graph structure between device nodes and extracts features such as node connectivity and path length. This helps identify key nodes or branches (such as branches that have a greater impact on overall line loss) and analyze the propagation mode of faults.

[0042] A3, extract the characteristics of load distribution by spatial region, such as branches with concentrated loads in certain areas, areas with high power density in local areas, etc. This can help identify high line loss areas or abnormal areas.

[0043] A4, analyzes the cluster distribution characteristics of equipment in geographic space, combined with the load characteristics within the cluster, to discover the differences in line losses of equipment in densely populated areas.

[0044] It should be noted that the feature extraction method can provide a perspective based on geographic location and physical layout to help identify line losses or faults caused by uneven spatial distribution, thereby improving the line loss analysis and fault diagnosis model.

[0045] S2: Input the first feature extraction result to construct a first diagnosis model.

[0046] Furthermore, a first feature extraction is performed on the original data, and features are extracted from the original data according to a time series to obtain a first feature extraction result.

[0047] In the embodiment of the present application, constructing the first diagnostic model may be to design an RNN model.

[0048] B1, input layer: set the input layer according to the result of feature extraction, and input a feature vector at each time step (for example, if a user privately connects wires around the electric energy meter and connects electricity to the user's home, the feature model is: the current of the household line under the electric energy meter > the current of the electric energy meter outlet line).

[0049] B2, hidden layer: Design a hidden layer composed of multiple RNN units to capture the time dependency and long-term correlation in the data (such as the household line current under the electricity meter > 1.5 times the meter outgoing line current is bypassing the electricity theft).

[0050] B3, output layer: Design the output layer according to the line loss fault type. Deploy the trained line loss model to the simulation system to detect line loss faults in the substation area.

[0051] In an optional embodiment, constructing the first diagnostic model may also be based on a convolutional neural network (CNN) model design, using the feature extraction capability of CNN to identify spatial patterns and local correlations to improve the detection accuracy of line loss faults.

[0052] Step 1: The input layer is set as a feature matrix based on the result of feature extraction, which can contain a variety of time series features and business features (such as current, voltage, load information of different nodes, etc.), and the data is formatted into a multi-dimensional matrix input suitable for convolution operation.

[0053] In step 2, multiple convolutional layers are set to capture spatial features and local abnormal patterns in the feature matrix (for example, the current and voltage differences of some specific devices are significantly higher than other nodes). The convolution kernel can extract specific local patterns for identifying potential fault signs.

[0054] Step 3: Reduce the feature dimension and the amount of computation while retaining the key features extracted in the convolution layer, thereby improving the computational efficiency of the model. The feature outputs of the convolution layer and the pooling layer are integrated through the fully connected layer to generate a comprehensive feature vector for identifying complex fault modes. Multiple neurons can be set to improve the accuracy of fault identification. The output layer is designed according to the line loss fault type to generate classification or regression output.

[0055] It should be noted that when the simulation software is running automatically, the data of each branch node in the substation is automatically calculated according to the line loss fault switching various fault modes and scenarios.

[0056] Embodiment 2: The following is an embodiment of the present invention, which provides a line loss simulation application method based on embodied intelligent fault diagnosis technology, specifically including:

[0057] S3: Build the first simulation system to investigate and handle line losses.

[0058] In an embodiment of the present application, building a first simulation system may be to use line loss substation simulation software to build a dynamic embodied scenario of line loss faults, and adjust the line loss substation simulation device to simulate a real line loss substation scenario, including C1 to C4.

[0059] C1, Initialization: Set the initial parameters of the simulation software, including line loss area, node type (PQ, PV, balancing node, etc.), line impedance, transformer parameters, etc. At the same time, define various possible line loss fault modes and scenarios. Fault mode switching: Automatically switch to the specified line loss fault mode according to simulation requirements. According to the specified line loss fault mode, the corresponding node load and line impedance parameters will change.

[0060] C2, the simulation software uses the fast decoupling method to calculate the data of each branch node in the substation: according to the above selected line loss substation, node type, line impedance and other given operating conditions and system wiring conditions, the operating status of the entire substation, node voltage, power flowing through the equipment, power loss of the system, etc. Its main purpose is to calculate the parameters of the operating status of equipment and users in the power network.

[0061] C3, calculation by fast decoupling method: Initialization voltage: Usually the initial value of the voltage of the same type of real line loss substation is used, or the voltage value set by the student can be used as the initialization voltage.

[0062] C4, decoupling iteration: decompose the Jacobian matrix into real and imaginary parts, ignoring the coupling terms between them. Solve the unbalanced amount of active power and reactive power respectively, and update the real and imaginary parts of the voltage. Iterate until the convergence condition is met (such as the power unbalance amount is less than the set threshold). Output node data: After the calculation is completed, output the key parameters such as voltage, current, power, etc. of each branch node in the substation.

[0063] In an optional embodiment, the first simulation system may also be an intelligent fault diagnosis simulation system based on a neural network and an expert system. Specifically, it includes:

[0064] Step 1, feature extraction and data preprocessing: extract features (such as current, voltage, load rate, etc.) from the real-time monitoring data and historical operation data of the substation, and perform data cleaning and normalization to ensure the accuracy of data input.

[0065] Step 2, preliminary diagnosis: Input the features into the rule base and use the inference engine to quickly judge typical line loss faults (such as unauthorized wiring, open meter cover, etc.). If a rule is met, the system directly marks the fault type; otherwise, it switches to the neural network for in-depth diagnosis.

[0066] Step 3, complex pattern recognition: input the preprocessed data into the neural network model to identify complex or atypical fault patterns. The neural network automatically adapts to multiple factors such as load fluctuations and abnormal power consumption behavior based on the learned historical features, and outputs the fault classification probability.

[0067] Step 4, result fusion and decision support: The diagnostic results of the expert system and the neural network are integrated. If the two diagnoses are consistent, the fault type is confirmed; if the results conflict, the weight algorithm of the decision support module is used for fusion to determine the final diagnostic result and generate fault location or optimization suggestions. Feedback and continuous learning: The system feeds back the diagnostic results to the rule base and neural network model, and uses new fault data to regularly update and optimize the model to improve the accuracy and generalization ability of fault diagnosis.

[0068] It should be noted that the text knowledge base construction: the text knowledge base is constructed based on the big language model, which contains the standard process of line loss troubleshooting, safety operation specifications, equipment use, troubleshooting methods, historical case analysis, etc. Through deep learning and natural language processing technology, the big language model can understand and parse these text information to form a comprehensive understanding of the line loss troubleshooting task. Combined with the real-time video data collected by the camera on the simulation device, the big language model integrates text knowledge with visual information. It can match abstract text descriptions with specific visual elements by analyzing the failure phenomenon of the metering equipment in the simulation device in the video and combining the line loss scene of the substation displayed by the simulation software, thereby enhancing the ability to understand complex operating environments. During the process of line loss troubleshooting by trainees, the embodied intelligent big language model can analyze the operation sequence in the video stream in real time and compare it with the standard process template. By identifying key operation steps and evaluating the logical relationship and time sequence between steps, the model can determine whether the trainee's operation process is correct and provide feedback or guidance when necessary.

[0069] S4: deploying the first diagnostic model to the simulation system for fault diagnosis.

[0070] In an embodiment of the present application, deploying the first diagnostic model to the simulation system includes converting fault features extracted from text based on big data language model training into control signals and performing control output on virtual signal devices.

[0071] Furthermore, the data of substations with higher actual line losses are extracted, the terminal information of low-voltage substations in substations with higher actual line losses is extracted, a substation line loss fault model is established, and automatic simulation is performed in the simulation software, and various fault modes and scenarios are switched according to the line loss fault; a static scenario is selected from the intelligent line loss fault model embodied in the simulation software, and a dynamic scenario is generated according to the probability of occurrence of various elements in different substations.

[0072] Furthermore, fault detection includes automatic generation of line loss area scenarios. The software extracts features of various elements of the line loss area from text based on pre-trained large language models, and generates a digital model of the line loss area based on the features.

[0073] Furthermore, the terminal information of low-voltage substations in substations with higher actual line loss is extracted, and the fault feature points are extracted in conjunction with on-site fault handling and on-site electricity theft evidence collection pictures, and input into the intelligent perception model of line loss faults; the large language model is used to generate respective fault phenomena and scenarios for each element in the substation line loss fault library and match the relevant probabilities; faults are sampled from the fault library according to the relevant probabilities of each element, and the real-time status of the dynamic embodied scenario is calculated according to the influence of each element of the line loss fault, so as to obtain the dynamic embodied scenario of the line loss fault.

[0074] Furthermore, through embodied intelligence transformation based on the line loss troubleshooting scenario pre-trained with the large language model, the industrial camera captures the line loss troubleshooting images of the trainees in the simulation device and processes them to determine whether the line loss troubleshooting process is correct and whether the operation meets the requirements of the safety operation specifications.

[0075] In addition to the operation process, the large language model can also evaluate the trainees' operation behavior based on safe operation specifications. It predicts potential safety risks by analyzing the trainees' clothing (such as whether they wear protective equipment), operation posture (such as whether they maintain a safe distance), tool usage, etc., and can issue warnings or suggestions as needed.

[0076] Furthermore, the embodied intelligent large language model has a strong situational awareness capability. It can dynamically adjust the analysis strategy and evaluation criteria according to the environmental information in the video (such as the device status (the simulation device adopts a modular setting, and different modules will be replaced according to different competition requirements); personnel distribution (in the competition, teachers, students or safety officers, operators, etc. will be assigned according to the competition content)) to adapt to different investigation scenarios and conditions.

[0077] Furthermore, as new data is continuously added and feedback from students is collected, the large language model can continue to learn and optimize its analysis model and prediction capabilities. Through the iterative process of machine learning, the model can continuously improve recognition accuracy, evaluation precision, and response speed.

[0078] It should be noted that the large language model can provide real-time feedback and guidance when students are performing troubleshooting operations. Through speech recognition and synthesis technology, the model can interact with students in natural language to answer questions, point out errors, and provide improvement suggestions. Based on the students' operating performance and learning progress, the large language model can also tailor a personalized learning path for them. By recommending targeted training tasks and simulation scenarios, it helps students quickly master line loss troubleshooting skills and improve safety awareness.

[0079] Embodiment 3, as Figure 2 The figure shows an embodiment of the present invention, which provides a line loss simulation application system based on embodied intelligent fault diagnosis technology. Specifically, it includes:

[0080] Step 1: Extract the data of the real high-line loss area, including load rate, total number of households, distribution transformer capacity, low-voltage line length, power supply radius, area attributes, and area topology, extract feature descriptions, build an area model based on digital modeling technology, and input the operation monitoring data of the acquisition system, equipment life cycle data, and environmental data into the model, so that the real-time status information of the physical entity equipment can be accurately mapped to the digital model of the area, thereby simulating the line loss area. (This step is the digital modeling of the area)

[0081] Step 2: Extract the low-voltage terminal information of the substation with higher real line loss, including equipment parameters, feature descriptions, and operation data of concentrators, special transformers, electric energy meters, and transformers. (This step is to model the substation metering equipment)

[0082] Step 3: Extract the characteristics of the line loss fault in the substation based on the on-site line loss inspection work order and the substation line loss detection data, and establish an intelligent model of line loss fault based on the neural network. (This step is to establish the substation line loss fault model)

[0083] Step 4: Extract fault feature points in cooperation with on-site fault handling and on-site electricity theft evidence collection pictures, and input them into the intelligent perception model of line loss faults; (the above 4 steps are to establish an intelligent perception model of line loss faults)

[0084] The simulation software automatically runs the simulation, switches various fault modes and scenarios according to the line loss fault, and automatically calculates the data of each branch node in the substation (the load at the fault point will change according to different line loss faults, and the data of each node will be re-output);

[0085] So far, a normal software simulation substation can run normally; (the above steps establish a static line loss scenario in the substation, including normally operating concentrators, special transformers, energy meters, transformers, etc., and also include metering equipment that causes high line loss faults. The faulty equipment outputs fault phenomena according to the line loss intelligent perception model, which can be used by students to troubleshoot on the system)

[0086] Select a static scenario from the embodied intelligent line loss fault model in the simulation software, and generate a dynamic scenario based on the probability of occurrence of each element in different substations. (This step establishes a dynamic substation line loss scenario based on the above static scenario).

[0087] The input module extracts feature descriptions based on the actual data of substations with higher line losses, including load rate, total number of households, distribution transformer capacity, low-voltage line length, power supply radius, substation attributes, and substation topology, builds a substation model based on refined digital modeling technology, and inputs the collected operation monitoring data, equipment life cycle data, and environmental data into the model.

[0088] The model building module extracts the line loss fault characteristics of the substation according to the on-site line loss inspection work order and the substation line loss detection data, and establishes an intelligent model of line loss fault based on the neural network.

[0089] The optimization module uses a large language model to analyze the various elements in the line loss fault intelligent model, calculates the probability of discovering line loss faults during on-site management, and improves the line loss fault model.

[0090] Processing module, the above models together constitute an embodied intelligent line loss fault model, select a static scene from the embodied intelligent line loss fault model, generate a dynamic scene according to the probability of occurrence of each element in different substations, and map the substation line loss characteristic data and the fault data of the low-voltage substation terminal to the control simulation device, and output the substation line loss nodes and fault data.

[0091] Among them, various line loss factors such as energy meters, mutual inductors, topological relationships, etc. can be selected and set on the software as needed to generate dynamic embodied intelligent scenarios, which are mapped to physical devices such as energy meters, mutual inductors, concentrators, etc., and various faults are generated on the physical devices at the same time;

[0092] Through embodied intelligence, the line loss troubleshooting requirements pre-trained based on the large language model are converted, and the trainee training images collected by industrial cameras are processed to determine whether the line loss troubleshooting process and operations meet the line loss management requirements.

[0093] Example 4: The following is an embodiment of the present invention, which provides a line loss simulation application method based on embodied intelligent fault diagnosis technology. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0094] Select a static scenario from the embodied intelligent line loss fault model in the simulation software, and generate a dynamic scenario based on the probability of occurrence of each element in different substations.

[0095] Static scenario selection: Select a static scenario from the static embodied scenario data set of the simulation software as the starting point of the dynamic scenario. This static scenario includes power supply radius, substation attributes, substation topology, etc.

[0096] Data preparation: Build or adjust the area setting library, including the user information that may be involved in the area (such as residential areas, villas, factories, etc.) and their basic attributes (such as capacitive power consumption, inductive power consumption, etc.).

[0097] Use the large language model to generate or improve activity-related probabilities, including the probability of weather, different users, holidays and other factors (such as metering failures, electricity theft failures, technical factors, etc.) occurring on different devices (such as transformers, electricity meters, etc.).

[0098] Dynamic scene generation:

[0099] 1: Select the static scene to start and initialize the status of each element in the stage area.

[0100] For example, if you select a line loss substation in a mountainous area, the substation has a large power supply radius and old lines, initialize the status of various elements such as the substation power supply radius, substation attributes, substation topology, and metering equipment.

[0101] 2: According to the characteristics of the substation and the line loss abnormal probability database, simulate the abnormalities of the substation within the specified time period. For each abnormality, automatically match the device and location where the abnormality occurs according to the abnormal content a and the probability P(l|a) of the device l.

[0102] According to the characteristics of the mountainous area: large power supply radius, old lines, high forest coverage, prominent line tree problems, low-voltage overhead line leakage, user theft, three-phase imbalance and other abnormalities that affect the line loss in the area are more likely. For example, in the simulation of abnormal line loss in the mountainous area from July to August, according to the abnormal content a (including the line leakage caused by thunderstorms in July and August affecting the line trees in the area, causing incomplete carrier signals and failure to collect user data, transformer line abnormalities causing three-phase imbalance; factors such as sudden increase in high-temperature electricity consumption in July and August, user theft, and user over-capacity electricity consumption causing metering equipment failure, etc.); and the device l probability P (l|a) automatically match the device and location where the abnormality occurs. For example (the line has a high probability of leakage caused by thunderstorms in July and August affecting the line trees in the area, and the corresponding abnormal device is the low-voltage overhead line, and the location is the forest area).

[0103] 3: For each line loss anomaly, further simulate its impact on the device. If the abnormal activity affects device o, determine whether the impact occurs based on the impact probability P(o|a,l).

[0104] For example, during July and August, the temperature is high and the air conditioning fee increases dramatically. We can make judgments based on the impact probability. For example, there is a high probability that some users will steal electricity from the metering equipment. There is also a high probability that the high temperature will cause users to exceed their electricity capacity, thus causing the metering equipment to malfunction.

[0105] 4: According to the occurrence of activities and the change of equipment status, calculate and update the status S_t of the station area at each time t. This includes simulating and displaying the current status, location and fault phenomenon of the equipment on the simulation software, and displaying the corresponding fault status on the corresponding simulation equipment on the simulation device.

[0106] 5: Repeat steps 2 to 4 until the simulation ends.

[0107] Table 1 Part of the simulation experiment data

[0108]

[0109] It can be seen from the above experimental data that the embodied intelligent line loss fault model of the present invention significantly improves the comprehensiveness and accuracy of line loss anomaly detection by automatically generating dynamic scenarios and introducing the comprehensive influence of multiple factors such as user load and weather changes. A comprehensive analysis of the impact of line loss, experimental data shows that from early July to the end of August, the increase in line loss caused by various abnormal events was higher than the average line loss value simulated by the prior art (about 3.5%). Especially during the thunderstorm in early July, the low-voltage overhead lines in the mountainous area experienced an 8.5% increase in leakage, and the risk of metering equipment failure caused by users' over-capacity use during high temperatures reached 50%, which is much higher than the risk prediction of traditional line loss models. The dynamically generated scenarios fully demonstrate the high responsiveness of the present invention to abnormal events under the influence of the external environment, and can significantly improve the line loss management effect.

[0110] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0111] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0112] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0113] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A line loss simulation application method based on embodied intelligent fault diagnosis technology, characterized in that: include: Use the power grid acquisition system to obtain raw data and perform first feature extraction on the raw data; Inputting the first feature extraction result to construct a first diagnostic model; Build the first simulation system to investigate and handle line losses; The first diagnostic model is deployed into the simulation system for fault diagnosis.

2. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 1, characterized in that: The method of collecting raw data using the power grid collection system includes collecting substation line loss monitoring data through the power grid collection system and pre-processing to obtain raw data.

3. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 2, characterized in that: The extracting the first feature from the original data includes extracting features from the original data according to a time series to obtain a first feature extraction result.

4. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 3, characterized in that: The constructing of the first diagnostic model includes setting an input layer according to the result of the first feature extraction and designing an output layer according to the line loss fault type.

5. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 4, characterized in that: The construction of the first simulation system includes using line loss substation simulation software to build a dynamic embodied scenario of line loss faults, and adjusting the line loss substation simulation device to simulate a real line loss substation scenario.

6. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 5, characterized in that: The deploying the first diagnostic model to the simulation system includes converting the fault features extracted from the text based on the big data language model training into control signals, and performing control output for the virtual signal device; Extract the data of substations with higher real line loss, extract the terminal information of low-voltage substations with higher real line loss, establish a substation line loss fault model, perform automatic simulation in the simulation software, and switch various fault modes and scenarios according to the line loss fault; select a static scenario from the intelligent line loss fault model embodied in the simulation software, and generate a dynamic scenario based on the probability of occurrence of various elements in different substations.

7. The line loss simulation application method based on embodied intelligent fault diagnosis technology according to claim 6, characterized in that: The fault detection includes realizing automatic generation of line loss area scenes, the software extracts features of various elements of the line loss area from text according to the pre-trained large language model, and generates a digital model of the line loss area based on the features; Extract the terminal information of low-voltage substations in substations with higher actual line loss, cooperate with on-site fault handling and on-site electricity theft evidence collection pictures, extract fault feature points, and input them into the intelligent perception model of line loss faults; use the large language model to generate respective fault phenomena and scenarios for each element in the substation line loss fault library and match the relevant probabilities; sample faults from the fault library according to the relevant probabilities of each element, calculate the real-time status of the dynamic embodied scenario according to the influence of each element of the line loss fault, and obtain the dynamic embodied scenario of the line loss fault.

8. A system using the line loss simulation application method based on embodied intelligent fault diagnosis technology as claimed in any one of claims 1 to 7, characterized in that: It includes input module, model building module, optimization module and processing module; The input module extracts feature descriptions based on real high-line loss substation data, including load rate, total number of households, distribution transformer capacity, low-voltage line length, power supply radius, substation attributes, and substation topology, builds a substation model based on refined digital modeling technology, and inputs the collected operation monitoring data, equipment life cycle data, and environmental data into the model; The model building module extracts the characteristics of the line loss fault in the substation area according to the on-site line loss troubleshooting work order and the line loss detection data in the substation area, and establishes an intelligent model of the line loss fault based on the neural network; The optimization module uses a large language model to analyze the elements in the line loss fault intelligent model, calculates the probability of finding line loss faults in on-site management, and improves the line loss fault model; The processing module and the above models together constitute an embodied intelligent line loss fault model, select a static scene from the embodied intelligent line loss fault model, generate a dynamic scene according to the probability of occurrence of each element in different substations, and map the substation line loss characteristic data and the fault data of the low-voltage substation terminal to the control simulation device, and output the substation line loss nodes and fault data.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the line loss simulation application method based on embodied intelligent fault diagnosis technology described in any one of claims 1 to 7 are implemented.

Images