Power distribution network non-power-cut operation simulation drilling and normative evaluation method and system

Abstract

The invention discloses a power distribution network non-power-cut operation simulation drill and normalization evaluation method and system, and the method comprises the following steps: constructing a virtual environment of power distribution network non-power-cut operation, and carrying out the virtualization modeling of a typical operation task in the virtual environment; performing a virtual drill in the virtual environment based on a result of the virtualized modeling; in the virtual drilling process or after the virtual drilling process is finished, the execution process of the standardized operation steps is evaluated based on the constructed normative evaluation model; wherein the normative evaluation model integrates hard rule verification based on a rule engine and operation time sequence rationality analysis based on a machine learning model; and generating an evaluation report containing quantitative scores and risk points, and performing superposition display on the risk points in the form of warning marks when the execution process of the simulation drill is reproduced in the virtual environment. According to the invention, real-time and intelligent operation normative evaluation and feedback can be realized.

Description

Technical Field

[0001] This invention relates to the field of power system training technology, specifically to a method and system for simulation and standardized evaluation of live-line work in power distribution networks. Background Technology

[0002] Live-line work on power distribution networks is a crucial aspect of power system maintenance and repair, requiring personnel to possess advanced skills and adhere to strict operational procedures. Traditional training methods primarily rely on hands-on drills, which suffer from high safety risks, high costs, and subjective evaluation.

[0003] Existing technologies include some virtual reality-based training systems, but most focus on scenario simulation and lack intelligent, standardized evaluation mechanisms. They cannot analyze the rationality of operational sequences in real time, and their evaluation models often rely on fixed rules, resulting in poor adaptability. Therefore, a system and method integrating virtual simulation and intelligent evaluation are needed to improve training efficiency and safety. Summary of the Invention

[0004] To address the technical problems of existing technologies, such as the lack of intelligent and standardized evaluation mechanisms, the inability to analyze the rationality of operation sequences in real time, and the poor adaptability of evaluation models, this invention provides a method and system for simulation and standardized evaluation of live-line work in power distribution networks. This solution utilizes virtualization modeling and a hybrid evaluation model to achieve standardized training and automated scoring of operational procedures, thereby reducing the risks associated with hands-on training.

[0005] In a first aspect, the present invention provides a method for simulating and evaluating the standardization of live-line work in power distribution networks, comprising the following steps: S101: Construct a virtual environment for live-line work on the power distribution network, and perform virtualization modeling of typical work tasks in the virtual environment; wherein, the virtualization modeling includes: decomposing the work process of the typical work task into multiple standardized operation steps, and using a state machine model to define the triggering conditions and state transition logic of each standardized operation step. S102: Based on the results of the virtualization modeling, a virtual exercise is performed in the virtual environment; wherein, the virtual character model is driven by animation skeletal binding technology, and the mechanical interaction between the operation tool and the virtual environment is simulated based on the physics engine, so as to visually reproduce the execution process of the standardized operation steps; S103: During or after the virtual exercise, the execution process of the standardized operation steps is evaluated based on the constructed normative evaluation model; wherein the normative evaluation model integrates hard rule verification based on a rule engine and operation sequence rationality analysis based on a machine learning model; S104: Generate an evaluation report containing quantitative scores and risk points, and display the risk points as warning markers when reproducing the execution process of the simulation exercise in the virtual environment.

[0006] Furthermore, the typical work tasks include at least one of insulation shielding, wire splicing, and switchgear maintenance, and each state in the state machine model corresponds to a standardized operation step. The transition between states is based on the triggering conditions, which include a signal triggered by the completion of the previous standardized operation step or a predefined event occurring in the virtual environment.

[0007] Furthermore, the method of driving the virtual character model through animation skeletal rigging technology and simulating the mechanical interaction between the operating tool and the virtual environment based on a physics engine includes: The animation skeletal binding technology is used to drive the hand movements of the virtual character model, thereby causing the virtual operating tools to produce corresponding movements. Based on the movement of the operating tool, the physics engine simulates the force feedback effect generated when it interacts with devices in the virtual environment.

[0008] Furthermore, the hard rule verification based on the rule engine verifies hard indicators including at least one of the following: safe distance threshold, tool usage sequence, and protective equipment wearing status, wherein the safe distance threshold is dynamically set based on electrical safety specifications.

[0009] Furthermore, the operation timing rationality analysis based on the machine learning model includes: using a timing sequence model to analyze the execution sequence of the standardized operation steps to detect abnormal situations such as step jumps or reversed order, wherein the timing sequence model is constructed based on a long short-term memory network or a hidden Markov model.

[0010] Furthermore, the normative evaluation model is a hybrid evaluation model combining the rule engine and the machine learning model, wherein: The rule engine outputs binary verification results; The machine learning model outputs a probability score for the timing of operations; The hybrid evaluation model weights and fuses the binary verification result with the probability score to generate an overall evaluation score.

[0011] Furthermore, the evaluation report also includes: a score for each standardized operating step, a detailed description of the risk points, and corresponding improvement suggestions; Once generated, the evaluation report is stored in a database and can be exported as a structured document.

[0012] Furthermore, the display of risk points as overlaid warning markers includes: In the reproduced image, different color codes are used to highlight the locations of erroneous operations with different risk levels; When a high-risk operation is triggered, voice annotations will be provided to indicate the error.

[0013] On the other hand, the present invention also provides a power distribution network live-line operation simulation exercise and standardization evaluation system for implementing the method described above, including: The 3D real-scene simulation module is used to build a virtual environment for uninterrupted power distribution network operations; The virtualization modeling module is used to perform virtualization modeling of typical job tasks in the virtual environment; the virtualization modeling includes: decomposing the job process of the typical job task into multiple standardized operation steps, and using a state machine model to define the triggering conditions and state transition logic of each standardized operation step; The physics simulation module is used to perform virtual drills in the virtual environment; wherein, the virtual character model is driven by animation skeletal binding technology, and the mechanical interaction between the operation tool and the virtual environment is simulated based on the physics engine, so as to visually reproduce the execution process of the standardized operation steps; The evaluation model module, including a rule engine submodule and a machine learning submodule, is used to construct a normative evaluation model; during or after the virtual exercise, the execution process of the standardized operation steps is evaluated based on the constructed normative evaluation model. The scoring report module is used to generate an evaluation report that includes quantitative scores and risk points; The replay module is used to overlay the risk points as warning markers when reproducing the execution process of the simulation exercise in the virtual environment.

[0014] Furthermore, the system also includes a data storage module for storing operation records, evaluation model parameters, and historical quantitative scoring data, and supports data mining for model optimization; the 3D real-scene simulation module is implemented based on the Unity 3D or Unreal Engine platform.

[0015] Compared with existing technologies, the core innovation of this invention lies in the deep integration of a hybrid evaluation model that combines a state machine model with a rule engine and machine learning, thereby achieving real-time and intelligent evaluation and feedback of operational norms.

[0016] The present invention has at least the following beneficial effects: 1) Standardized and intelligent evaluation of the training process was achieved. A state machine model was used to break down the complex live-line work process into standardized operating steps, with each state corresponding to a standardized operating step, and state transitions based on explicit triggering conditions. This structured modeling approach makes the training process repeatable and comparable, providing a basic framework for standardized evaluation. Combined with a hybrid evaluation model, it ensures both strict adherence to basic safety regulations and the identification of complex temporal anomaly patterns, significantly improving the accuracy and comprehensiveness of the evaluation.

[0017] 2) It provides a more realistic and immersive operating experience. A lifelike working environment is constructed based on a 3D real-scene simulation platform. Animation skeletal binding technology enables more precise synchronization between operator gestures and tool movements. Combined with a physics engine, it simulates realistic mechanical interaction effects (such as the torque feedback of a wrench tightening a bolt). This multi-sensory immersive experience allows trainees to obtain an operating feel close to actual work in a virtual environment, effectively improving the authenticity and effectiveness of the training.

[0018] 3) A multi-dimensional intelligent evaluation system was established. A hybrid evaluation model combining a rule engine and machine learning was adopted. The rule engine is responsible for verifying hard indicators such as safety distance thresholds, tool usage order, and protective equipment wearing status, while the machine learning model uses an LSTM network to analyze the rationality of the operation sequence and detect anomalies such as step jumps or reversed order. The two are weighted and fused to generate an overall evaluation score, forming a comprehensive evaluation mechanism that combines rigid constraints with flexible evaluation.

[0019] 4) Real-time feedback and continuous improvement are achieved. The system can generate scoring reports and mark risk points in real time during virtual drills. During playback, it highlights the locations of erroneous operations at different risk levels using color coding, supplemented by voice annotations to provide error prompts. This immediate and intuitive feedback mechanism helps trainees correct errors promptly and provides data support for the continuous improvement of training effectiveness.

[0020] 5) Enhanced training safety and cost-effectiveness. Virtual training completely eliminates safety risks associated with hands-on training, significantly reducing equipment wear and tear and training costs. The system supports repeated practice and playback analysis, allowing trainees to master operational skills in a risk-free environment, greatly shortening the training cycle and improving training efficiency. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating the steps of the power distribution network live-line operation simulation and standardization evaluation method provided in this embodiment of the invention.

[0023] Figure 2 This is a schematic diagram illustrating the working principle of the power distribution network live-line operation simulation and standardization evaluation method provided in this embodiment of the invention.

[0024] Figure 3 The flowchart illustrates the process of the power distribution network live-line operation simulation and standardization evaluation method provided in this embodiment of the invention.

[0025] Figure 4 This is a schematic diagram of the state machine model of the present invention during insulation shielding operations.

[0026] Figure 5 This is a schematic diagram of the workflow of the hybrid evaluation model of the present invention.

[0027] Figure 6 This is a schematic diagram of the hardware configuration of the power distribution network live-line operation simulation and standardization evaluation system of the present invention.

[0028] Figure 7 This is a schematic diagram of the architecture of the power distribution network live-line operation simulation exercise and standardization evaluation system of the present invention. Detailed Implementation

[0029] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0031] The implementation of the method of this invention is based on a 3D reality simulation platform (such as Unity3D or Unreal Engine), integrating a physics engine (such as NVIDIA PhysX) and a machine learning framework (such as TensorFlow). The system hardware includes a high-performance computer, a VR headset, motion capture equipment, etc.

[0032] In an embodiment of the present invention, reference is made to... Figure 1 As shown, the core of this method specifically includes the following steps: S101: Construct a virtual environment for live-line work on the power distribution network, and perform virtualization modeling of typical work tasks within this virtual environment. Virtualization modeling includes: breaking down the workflow of typical work tasks into multiple standardized operation steps, and defining the triggering conditions and state transition logic for each standardized operation step using a state machine model; S102: Based on the results of virtualization modeling, perform virtual drills in a virtual environment. This includes: driving virtual character models using animation skeletal rigging technology and simulating the mechanical interactions between the operating tools and the virtual environment based on a physics engine, to visually reproduce the execution process of standardized operating procedures; S103: During or after the virtual exercise, the execution process of standardized operating procedures is evaluated based on the constructed normative evaluation model. This constructed normative evaluation model integrates hard rule verification based on a rule engine with the rationality analysis of the operation sequence based on a machine learning model. S104: Generate an evaluation report containing quantitative scores and risk points, and display the risk points as warning markers when reproducing the execution process of the simulation exercise in the virtual environment.

[0033] Figure 2 A schematic diagram illustrating an embodiment of the method of the present invention is shown. Figure 2 As shown, the core execution logic and principles of the live-line work simulation and standardization evaluation method for power distribution networks are broken down into four key steps using a hierarchical modular flowchart. The process begins with the "building a virtual environment and virtualization modeling" step, which is the foundation of the entire method. First, a virtual environment for live-line work in power distribution networks is built using a 3D reality simulation platform based on Unity 3D or Unreal Engine, creating digital models of power distribution network equipment, tools, and work scenarios. Then, virtualization modeling is performed for typical work tasks such as insulation shielding, conductor splicing, and switchgear maintenance, breaking down the work process into multiple reusable standardized operation steps. Finally, a state machine model is used to define the triggering conditions and state transition logic for each standardized operation step, clarifying the sequence and triggering rules of the work process.

[0034] After completing the environment and process modeling, the "virtual exercise" phase begins. Animation skeletal rigging technology is used to establish the skeletal structure of the virtual character model. Inverse kinematics algorithms drive the virtual character's hand movements, achieving synchronization between the virtual character and the trainee's actions. Simultaneously, based on a physics engine, rigid body dynamics principles and collision detection algorithms are used to simulate the mechanical interaction between the operating tools and equipment in the virtual environment, such as the torque transmission when a wrench tightens a bolt. Furthermore, a spring-damped system model calculates force feedback signals, providing trainees with a more realistic tactile experience and visually reproducing the execution process of standardized operating procedures.

[0035] During and after the exercise, the "standardization evaluation" phase commences. This phase employs a hybrid evaluation model. On one hand, a rule engine verifies hard indicators such as safety distance thresholds, tool usage sequence, and protective equipment wearing status, dynamically setting safety distance thresholds based on electrical safety regulations and outputting a binary verification result of "compliant / non-compliant." On the other hand, a machine learning model based on Long Short-Term Memory (LSTM) networks or Hidden Markov Models analyzes the execution sequence of standardized operating procedures, detecting anomalies such as step jumps and reversed sequences, and outputting a probability score for the operation timing. Finally, the binary verification result and the probability score are weighted and fused to generate an overall evaluation score, achieving a comprehensive evaluation of the standardization of operations.

[0036] After the evaluation is completed, the process proceeds to the "Generate Evaluation Report and Display Risks" stage. Based on the overall evaluation score and the evaluation results of each step, an evaluation report is generated, which includes quantitative scores, detailed descriptions of risk points, and improvement suggestions. The generated report is stored in the database and can be exported as a structured document. Simultaneously, when reproducing the simulated exercise in the virtual environment, different color codes are used to highlight the locations of erroneous operations at different risk levels. When a high-risk operation is triggered, voice annotations are provided to indicate the error, intuitively presenting the risk points.

[0037] The entire process breakdown diagram, through clear module division and logical connections, organically combines virtual modeling, drill execution, standardized evaluation, and result output, ensuring the orderly implementation and efficient execution of power distribution network live-line operation simulation drills and standardized evaluation methods.

[0038] In conjunction with the steps of the method of this invention, step S101 mainly includes the construction of a 3D real-scene simulation platform, virtualization modeling of typical work tasks, and the setting or definition of a state machine model. The construction of the 3D real-scene simulation platform includes: building a virtual environment based on a game engine (such as Unity3D or Unreal Engine), and creating digital models of power distribution network equipment, tools, and work scenarios through 3D modeling technology. Scene graph management technology is used to organize various objects in the virtual environment to ensure rendering efficiency and scene realism.

[0039] Typical job task virtualization modeling includes: analyzing or decomposing the power distribution network live-line work process using domain expert knowledge to extract key operation nodes or key operation steps. A component-based architecture is adopted, where each job task is decomposed into multiple reusable standardized operation units or standardized operation steps.

[0040] The state machine model setting or definition includes: establishing a power distribution network uninterrupted operation process model using finite state machine theory, and defining the state set S={s1,s2,...,s...} nThe transition function is T: S×E→S, where E is the set of events. The job flow is visually represented by a state transition diagram, where each state corresponds to a specific operation step, and state transitions are triggered by predefined events.

[0041] The set of states S = {s1, s2, ..., s} n This is a series of discrete, clear, and ordered "snapshots" or "stages" obtained by "slicing" the entire complex live-line maintenance process of the power distribution network. In the system, each state corresponds to a standardized operating procedure. The following example illustrates this (using insulation shielding work as an example): S1: Initial state — "Workers are in position, ready to test for electricity"; S2: "Electrical testing in progress"; S3: "Install near-edge phase shield"; S4: "Install the far-side phase shield"; S5: "Install the mid-phase shield"; S6: End Status - "Operation completed, all shielding covers installed".

[0042] In the system of this invention, this is manifested in the fact that each state corresponds to a standardized operation step.

[0043] The transition function T: S×E→S is the decision-making and rule function of the entire state machine, defining "what the next state should be if event E occurs in the current state S". S×E represents the combination of the "current state" and the "occurring event". →S indicates that the result of the function is the next "state". This is the program logic written in the simulation platform (usually a switch-case or if-else statement), which continuously monitors the current state and the events that have occurred, and updates the current state according to predefined rules.

[0044] In step S101 above, the logical sequence of the work process is clearly defined through a state machine model, avoiding arbitrariness in training. At the same time, standardized operating procedures facilitate unified training standards and assessment requirements, the construction of a virtual environment reduces the cost and risk of physical training, and the state machine model provides a structured foundation for subsequent standardized evaluation.

[0045] The implementation of step S102 mainly includes: animation skeleton binding, physics engine simulation and mechanical feedback implementation.

[0046] Animation skeletal rigging specifically includes: establishing the skeletal hierarchy of the virtual character using skeletal animation technology; calculating joint angles using inverse kinematics algorithms to ensure coordination between gestures and tool operations; and mapping skeletal motion onto the character model surface using skinning weights. Skinning weights are an influence factor allocation system used to control how the 3D model mesh deforms with the movement of its internal bones. The skinning weight system mainly includes correlations (which bones influence a vertex), weight values ​​(the magnitude of influence of each bone), and a weight distribution map (a color map for visualization and editing). These three elements are input into a linear blending skinning algorithm, and the output is the final result.

[0047] The physics engine simulation specifically includes: establishing a physical interaction model between tools and equipment based on the principles of rigid body dynamics; using collision detection algorithms (such as the GJK algorithm, i.e., Gilbert–Johnson–Keerthi, also known as the GJK distance algorithm; or bounding box hierarchy) to detect the contact state; and calculating contact forces and moments through a constraint solver.

[0048] The mechanical feedback implementation specifically includes: establishing a force feedback model based on a spring-damped system, F = -kΔx - cv, where k is the stiffness coefficient, c is the damping coefficient, Δx is the displacement, and v is the relative velocity. The corresponding force feedback signal is then output through a tactile device.

[0049] The physics engine simulation in step S102 above provides a realistic operating experience, enhancing the immersion of the training. The mechanical feedback mechanism allows trainees to perceive the force applied during operation, cultivating fine motor skills. Animated skeletal rigging ensures natural and smooth movements, improving the realism of the training. Through this simulated real-world operation training, the risk of equipment damage due to improper operation is reduced.

[0050] The implementation of step S103 mainly includes rule engine construction, machine learning model design, and hybrid evaluation model fusion.

[0051] The following detailed explanation of the construction process of the hybrid evaluation model is provided through specific embodiments. The rule engine construction specifically includes: employing a rule engine based on the Rete algorithm to define premise-conclusion style evaluation rules; and establishing a rule base that incorporates professional knowledge in areas such as safety specifications, operational procedures, and quality standards.

[0052] The machine learning model design specifically includes: using time-series data analysis methods to collect operation sequence data as training samples; employing an LSTM network to capture long-term dependencies; and the network structure including an input layer, multiple LSTM layers, a fully connected layer, and an output layer.

[0053] The hybrid evaluation model fusion specifically includes: establishing a multi-source information fusion framework, with the rule engine outputting discrete compliance judgments and the machine learning model outputting continuous probability scores. A weighted average method is used for result fusion: Score = α·R_rule + β·P_ml, where α + β = 1.

[0054] The rule engine in step S103 ensures the strict implementation of basic safety specifications, the machine learning model can identify complex time-series patterns and discover potential risks, the hybrid evaluation model combines the advantages of deterministic and probabilistic evaluation, and improves the comprehensiveness of the evaluation, and the adaptive learning capability enables the system to continuously optimize the evaluation accuracy as experience accumulates.

[0055] The implementation of step S104 mainly includes the generation of scoring reports, the marking of risk points, and the implementation of playback functions.

[0056] The rating report generation process specifically includes: establishing a template-based report generation system, defining the report structure and content elements, and using natural language generation technology to convert structured evaluation data into readable text descriptions.

[0057] Risk point labeling specifically includes: establishing a risk labeling system based on spatial location information, locating risk positions using a three-dimensional coordinate system, and visually displaying risk areas in a virtual scene using graphic overlay technology.

[0058] The replay function specifically includes: establishing a data structure for operation records, storing time-series status information and event logs, and reproducing the training process through scenario reenactment technology, supporting multi-perspective observation and analysis.

[0059] The real-time scoring in step S104 above provides immediate feedback, making it easier for trainees to correct mistakes promptly. Visualized risk point annotations enhance the effectiveness of safety awareness training. The playback function supports in-depth analysis and teaching discussions. Structured analysis reports facilitate training effectiveness evaluation and record management.

[0060] Preferably, typical work tasks include at least one of insulation shielding, wire splicing, and switchgear maintenance, and each state in the state machine model corresponds to a standardized operation step. Transitions between states are based on predefined trigger conditions, including operation completion signals or environmental events. This preferred approach is implemented by: establishing a work task library and designing state machine models for typical tasks such as insulation shielding, wire splicing, and switchgear maintenance. A unified modeling language is used to define the state transition diagrams to ensure model consistency and maintainability. The main benefits are: coverage of major work types, further improving the system's practicality and applicability.

[0061] Preferably, the physics engine simulates tool operation, including torque feedback for tightening bolts with a wrench, and synchronizes operator gestures with tool movement through animated skeletal rigging. The physics engine calculates the force feedback effect based on Newtonian mechanics principles. This preferred step involves: establishing a physics-based torque calculation model, considering factors such as thread friction and material elasticity; and using haptic rendering technology to convert the calculation results into haptic signals. The main benefits are: providing a realistic operating feel and better cultivating trainees' force control abilities.

[0062] Preferably, the hard indicators verified by the rule engine include at least one of the following: safe distance threshold, tool usage sequence, and protective equipment wearing status, wherein the safe distance threshold is dynamically set based on electrical safety specifications. This preferred step involves establishing a multi-dimensional safety evaluation indicator system, including spatial distance, time series, and equipment status. The analytic hierarchy process (AHP) is preferably used to determine the weights of each indicator. The main benefits are: comprehensive coverage of safety elements and a better establishment of a systematic safety evaluation system.

[0063] Preferably, the machine learning model analysis of the temporal rationality of operations includes using a time-series model to detect skipped or reversed steps. This time-series model is constructed based on a Long Short-Term Memory (LSTM) network or a Hidden Markov Model and trained using historical operational data. The implementation of this preferred step includes: designing an LSTM network structure where the input layer receives temporal features, the hidden layer captures long-term dependencies, and the output layer generates anomaly probabilities. A sliding window method is used to process variable-length sequences. The main benefits are: effectively identifying complex temporal anomaly patterns and further improving risk warning capabilities.

[0064] Preferably, the normative evaluation model is a hybrid evaluation model combining a rule engine and machine learning. The rule engine outputs binary verification results, while the machine learning model outputs probability scores, and these scores are weighted and fused to generate an overall evaluation score. This preferred step involves: establishing a fusion algorithm based on DS evidence theory, handling the uncertainty information of the rule engine and machine learning model, and designing an adaptive weight adjustment mechanism to further improve the reliability and robustness of the evaluation results.

[0065] Preferably, the scoring report includes scores for operational steps, descriptions of risk points, and improvement suggestions. The report is automatically stored after generation and can be exported as a structured document. Implementation of this preferred step includes: designing an XML-formatted report template that supports structured data population; and establishing a report version management mechanism to ensure the integrity of historical records. The main benefits of this step include: providing standardized evaluation output, making comparison and analysis easier.

[0066] Preferably, during playback, overlaying warning markers includes highlighting the location of the erroneous operation in the 3D reality scene, differentiating risk levels through color coding, and providing voice annotations to indicate the error. This preferred step involves: establishing a visual coding scheme based on color psychology, where red represents high risk, yellow represents medium risk, and green represents safe. A hierarchical detail model of spatial annotation is designed. The main benefits include: a more intuitive display of risk distribution and enhanced training effectiveness.

[0067] In embodiments of the present invention, such as Figure 3 The diagram illustrates the workflow of a simulation exercise and standardized evaluation method for live-line work in power distribution networks. The process begins by constructing a 3D virtual environment to provide a scenario foundation for the work simulation. Next, the work process is broken down into standardized steps to facilitate subsequent standardized operation and evaluation. Then, a state machine model is used to define states and transitions, clarifying the state change logic at each stage of the work. A physics engine simulation tool is then used to make the simulation more closely resemble actual work. Finally, real-time operational data is collected to provide data support for subsequent evaluation.

[0068] The process then moves to the standardized evaluation model construction phase, which consists of two parts: rule engine verification and machine learning analysis. Rule engine verification covers aspects such as safety distance thresholds, tool usage order, protective equipment wearing, and the rationality of operation sequence; machine learning analysis includes step skip detection and action smoothness analysis. Based on these verification and analysis results, a comprehensive evaluation result is generated, leading to a scoring report and risk warnings. Afterwards, the exercise is replayed and analyzed, with warning markers overlaid, and finally, the process "ends." The entire process, through virtual environment simulation, standardized operation breakdown, and multi-dimensional evaluation, achieves a standardized evaluation of live-line work simulation exercises in power distribution networks.

[0069] In an embodiment of the present invention, Figure 4 This diagram illustrates the state transitions of a state machine model during insulation shielding operations. Initially, after personnel are in position, the process enters the "voltage testing" state. This testing requires meeting strict criteria: the voltage detector must be used correctly and testing time must be sufficient. When testing is complete and the duration is ≤30 seconds, the state transitions to "Installing near-side phase shielding." After the near-side phase is installed and the torque meets the requirements, a state verification process is performed to ensure the previous steps were completed, the tool usage sequence was correct, and the safety distance meets requirements. Then, the process transitions to "Installing far-side phase shielding." When the far-side phase is installed and the safety distance is ≥0.7m, the process transitions to "Installing middle-phase shielding." Once the middle phase is installed and fully shielded, the operation is complete, and the process ends. This state transition diagram clearly shows the states and transition conditions at each stage of the insulation shielding operation, ensuring the operation is performed according to standardized procedures.

[0070] In an embodiment of the present invention, Figure 5This is a schematic diagram of the workflow of the hybrid evaluation model. First, operational data is input, followed by data preprocessing, including noise removal and missing data completion, to prepare for subsequent analysis. Then, feature extraction is performed, followed by a process flow divided into a rule engine processing flow and a machine learning processing flow. The rule engine processing flow sequentially performs safety distance verification, tool sequence verification, protective equipment inspection, and operation time limit verification. The machine learning processing flow first extracts temporal features, then analyzes them using an LSTM network to identify abnormal patterns, and finally outputs a probability score.

[0071] The evaluation results from the rule engine and machine learning are then fused and weighted to obtain a comprehensive evaluation score. For example, the risk level is determined based on the comprehensive evaluation score: 91-100 is considered safe, 81-90 is low risk, 61-80 is medium risk, and ≤60 is high risk. Finally, a report is generated and output based on the risk level. This process combines the determinism of the rule engine with the intelligence of machine learning to achieve accurate evaluation of the task.

[0072] In embodiments of the present invention, a power distribution network live-line working simulation exercise and standardization evaluation system is also provided to implement the method described above. (Reference) Figure 6 , 7 The hardware configuration and simulation implementation of the system in this embodiment are shown in the figure. The system mainly includes: a 3D real-scene simulation module, a virtualization modeling module, a physical simulation module, an evaluation model module, a scoring report module, and a playback module.

[0073] Among them, the 3D real-scene simulation module is used to build a virtual environment for uninterrupted power distribution network operations.

[0074] The virtualization modeling module is used to virtualize typical job tasks, break down the job process into standardized operation steps, and integrate state machine model to define trigger conditions and state transition logic.

[0075] The physics simulation module provides mechanical feedback through animation skeleton rigging and physics engine simulation tools.

[0076] The evaluation model module, which includes a rules engine submodule and a machine learning submodule, is used to build prescriptive evaluation models.

[0077] The scoring report module is used to generate scoring reports and mark risk points in real time.

[0078] The playback module is used to support the overlay of warning markers on the 3D reality scene during exercise playback.

[0079] The implementation of the above system includes: adopting a layered architecture design, with each module communicating through well-defined interfaces; and establishing a message bus-based data exchange mechanism to ensure loose coupling between modules. The main benefits are: further improving the system's maintainability and scalability, and better supporting independent upgrades of functional modules.

[0080] Preferably, the system also includes a data storage module for storing operation records, evaluation model parameters, and historical scoring data, and supports data mining for model optimization. An operational record, evaluation parameter, and scoring result are stored using a relational database schema. A data mining pipeline is established to support feature engineering, model training, and performance evaluation. This better enables persistent data management and intelligent analysis, supporting continuous system optimization.

[0081] In an embodiment of the present invention, Figure 6 and Figure 7 The hardware configuration and system architecture of the power distribution network live-line working simulation exercise and standardization evaluation system were showcased. The top layer is the user operation layer, where users operate the system using VR headsets, controllers, and motion capture equipment. These devices are connected to a real-time ray tracing module, which provides realistic lighting and shadow effects for the virtual scene. The results of the real-time ray tracing are input into the physics engine calculation module, which calculates and simulates physical interactions such as tool operations. Figure 6 As shown, the hardware foundation and software module composition of the power distribution network live-line operation simulation and standardization evaluation system are presented in the form of a "hardware support + software module" combination diagram, which clarifies the collaborative logic between hardware and software and provides a guarantee for the stable operation and functional realization of the system.

[0082] The hardware support platform is the fundamental carrier for system operation, mainly including VR headsets and controllers, motion capture equipment, high-performance computers, and auxiliary hardware. VR headsets (such as the Meta Quest 3) support 4K resolution to ensure visual fidelity in the virtual environment; the controllers have vibration modules that can output mechanical feedback. Motion capture equipment (such as OptiTrack) has a sampling rate of ≥120Hz, capturing the user's hand and body movements in real time, providing data for driving the virtual character model's movements. High-performance computers are equipped with high-performance CPUs (such as Intel i9) and GPUs (such as NVIDIA RTX 4090), providing sufficient computing power for 3D rendering, physics engine calculations, and machine learning model operation. Auxiliary hardware includes audio equipment (outputting voice annotation prompts) and data storage servers (storing system operation data), collectively forming the system's hardware support system.

[0083] The software modules are the core implementation carriers of the system's functions, encompassing a 3D real-scene simulation module, a virtualization modeling module, a physical simulation module, an evaluation model module, a scoring report module, a playback module, and a data storage module. The 3D real-scene simulation module, based on the Unity 3D or Unreal Engine platform, constructs a virtual environment for uninterrupted power distribution network operations. The virtualization modeling module breaks down typical work tasks into standardized operating steps, using a state machine model to define trigger conditions and state transition logic. The physical simulation module drives virtual character models through animation skeletal binding technology, simulating the mechanical interaction between tools and the environment based on a physics engine, achieving visualized re-enactment of drills. The evaluation model module includes a rule engine sub-module (verifying hard indicators) and a machine learning sub-module (analyzing the rationality of operation sequence), constructing a hybrid evaluation model. The scoring report module generates evaluation reports containing quantitative scores, risk points, and improvement suggestions. The playback module overlays risk warning markers during drill re-enactment. The data storage module stores operation records, model parameters, and historical scoring data, supporting data mining and model optimization.

[0084] During system operation, hardware and software modules work together: motion capture equipment collects trainee motion data and transmits it to the physical simulation module to drive the virtual character; the 3D real-scene simulation module presents the virtual environment; the evaluation model module analyzes the operation data in real time and outputs evaluation results; the scoring report module generates reports; and the playback module calls the stored data for review, forming a functional closed loop to ensure that the system achieves the core objective of power distribution network live-line operation simulation and standardized evaluation.

[0085] like Figure 7 As shown, the architecture design of the power distribution network live-line operation simulation exercise and standardization evaluation system is presented in the form of a four-layer hierarchical architecture diagram. It clarifies the functional positioning, module composition and data interaction relationship of each layer, and provides clear guidance for system design, development and maintenance.

[0086] The top layer is the "user interface layer," which serves as the interaction point between the system and users, directly facing trainees and training administrators. It includes a 3D virtual reality display interface, an interactive operation interface, an evaluation report display interface, and a replay analysis interface. The 3D virtual reality display interface uses VR or desktop display methods to present a virtual work environment and operation process, supporting viewpoint switching. The interactive operation interface provides virtual tool selection, operation command input, and force feedback adjustment functions, supporting VR controller or mouse and keyboard interaction. The evaluation report display interface presents the overall score, risk level, step score, risk points, and improvement suggestions, supporting report querying and export. The replay analysis interface includes a training timeline, risk marker toggle, voice prompt toggle, and screenshot / screen recording functions, facilitating trainee review and administrator feedback.

[0087] The second layer is the "business logic layer," the core execution layer of the system's functions. It implements the core functions of live-line operation simulation and standardized evaluation in power distribution networks, encompassing a 3D reality simulation module, a virtualization modeling module, a physical simulation module, an evaluation model module, a scoring report module, and a playback module. These modules work collaboratively: the 3D reality simulation module constructs the virtual environment; the virtualization modeling module defines standardized operating procedures; the physical simulation module enables visualized execution of the exercise; the evaluation model module completes the evaluation of operational standardization; the scoring report module generates evaluation reports; and the playback module supports exercise debriefing. Data interaction between modules is achieved through a message bus, ensuring smooth functional flow.

[0088] The third layer is the "data access layer," which serves as a bridge between the business logic layer and the data storage layer. It is responsible for reading, writing, and processing data, preventing the business logic layer from directly manipulating the database, thus improving system security and maintainability. It includes operation record access interfaces, evaluation model parameter access interfaces, and historical score data access interfaces. The operation record access interface supports reading and writing operation data and multi-condition queries; the evaluation model parameter access interface enables loading and updating model parameters; and the historical score data access interface supports reading, writing, and statistical analysis of score data, providing data support for the business logic layer.

[0089] The bottom layer is the "data storage layer," which is the persistent storage layer for system data. It uses a relational database (such as MySQL) or a non-relational database (such as MongoDB) to store databases for operation records, evaluation model parameters, and historical scores. The operation record database stores training data such as student ID, task type, operation steps, operation time, and operation parameters. The evaluation model parameter database stores configuration data such as rule engine thresholds, machine learning model structure parameters, and hybrid evaluation weight coefficients. The historical score database stores student training evaluation results, providing data support for system data traceability, model optimization, and training effect analysis.

[0090] This four-layer architecture follows the principle of "high cohesion and low coupling". Each layer has independent functions and works in synergy. It supports independent upgrades and maintenance of each layer, ensuring that the system has good scalability and maintainability, and meets the long-term use needs of power distribution network live-line operation simulation and standardization evaluation.

[0091] The results of the physics engine's calculations are stored as operation logs and simultaneously input into the hybrid evaluation model for real-time analysis. Operation logs are transmitted via a network communication unit (including a gigabit Ethernet interface supporting multi-terminal synchronization). After real-time analysis, the hybrid evaluation model outputs scoring reports, risk warnings, and replay data to support job evaluation and debriefing. The entire architecture encompasses user operation, scenario simulation, data processing, and evaluation analysis, forming a complete job simulation and evaluation system.

[0092] It should be noted that in this paper, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply these relationships. There is no such actual relationship or order between entities or operations. Furthermore, the terms "including" and "package" do not apply. The word "comprise" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0093] In this embodiment of the invention, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0094] The various embodiments in this specification are described in a related manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0095] In particular, the device embodiment is basically similar to the method embodiment, so the description is relatively simple. For relevant details, please refer to the description of the method embodiment.

[0096] For ease of description, the above apparatus is described by dividing it into various functional units / modules. Of course, in implementing this invention, the functions of each unit / module can be implemented in one or more software and / or hardware.

[0097] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0098] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for simulating and evaluating the standardization of live-line work in power distribution networks, characterized in that, Includes the following steps: S101: Construct a virtual environment for live-line work on the power distribution network, and perform virtualization modeling of typical work tasks in the virtual environment; wherein, the virtualization modeling includes: decomposing the work process of the typical work task into multiple standardized operation steps, and using a state machine model to define the triggering conditions and state transition logic of each standardized operation step. S102: Based on the results of the virtualization modeling, a virtual exercise is performed in the virtual environment; wherein, the virtual character model is driven by animation skeletal binding technology, and the mechanical interaction between the operation tool and the virtual environment is simulated based on the physics engine, so as to visually reproduce the execution process of the standardized operation steps; S103: During or after the virtual exercise, the execution process of the standardized operation steps is evaluated based on the constructed normative evaluation model; wherein the normative evaluation model integrates hard rule verification based on a rule engine and operation sequence rationality analysis based on a machine learning model; S104: Based on the evaluation, generate an evaluation report containing quantitative scores and risk points, and when reproducing the execution process of the simulation exercise in the virtual environment, display the risk points as overlays in the form of warning markers.

2. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The typical work tasks include at least one of insulation shielding, wire splicing, and switchgear maintenance. Each state in the state machine model corresponds to a standardized operation step. The transition between states is based on the triggering conditions, which include a signal triggered by the completion of the previous standardized operation step or a predefined event occurring in the virtual environment.

3. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The method of driving virtual character models through animation skeletal rigging technology and simulating the mechanical interaction between operating tools and the virtual environment based on a physics engine includes: The animation skeletal binding technology is used to drive the hand movements of the virtual character model, thereby causing the virtual operating tools to produce corresponding movements. Based on the movement of the operating tool, the physics engine simulates the force feedback effect generated when it interacts with devices in the virtual environment.

4. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The hard rule verification based on the rule engine verifies hard indicators including at least one of the following: safe distance threshold, tool usage sequence, and protective equipment wearing status, wherein the safe distance threshold is dynamically set based on electrical safety specifications.

5. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The operation timing rationality analysis based on the machine learning model includes: using a timing sequence model to analyze the execution sequence of the standardized operation steps to detect abnormal situations such as step jumps or reversed order, wherein the timing sequence model is constructed based on a long short-term memory network or a hidden Markov model.

6. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The normative evaluation model is a hybrid evaluation model that combines the rule engine and the machine learning model, wherein: The rule engine outputs binary verification results; The machine learning model outputs a probability score for the timing of operations; The hybrid evaluation model weights and fuses the binary verification result with the probability score to generate an overall evaluation score.

7. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The evaluation report also includes: a score for each standardized operating step, a detailed description of the risk points, and corresponding improvement suggestions; Once generated, the evaluation report is stored in a database and can be exported as a structured document.

8. The method for simulation and standardization evaluation of live-line work in power distribution networks according to claim 1, characterized in that, The method of displaying risk points as overlay warning markers includes: In the reproduced image, different color codes are used to highlight the locations of erroneous operations with different risk levels; When a high-risk operation is triggered, voice annotations will be provided to indicate the error.

9. A system for simulating and evaluating the standardization of live-line work in power distribution networks, used to implement the method for simulating and evaluating the standardization of live-line work in power distribution networks as described in any one of claims 1-8, characterized in that, include: The 3D real-scene simulation module is used to build a virtual environment for uninterrupted power distribution network operations; The virtualization modeling module is used to perform virtualization modeling of typical job tasks in the virtual environment; the virtualization modeling includes: decomposing the job process of the typical job task into multiple standardized operation steps, and using a state machine model to define the triggering conditions and state transition logic of each standardized operation step; The physics simulation module is used to perform virtual drills in the virtual environment; wherein, the virtual character model is driven by animation skeletal binding technology, and the mechanical interaction between the operation tool and the virtual environment is simulated based on the physics engine, so as to visually reproduce the execution process of the standardized operation steps; The evaluation model module, including a rule engine submodule and a machine learning submodule, is used to construct a normative evaluation model; during or after the virtual exercise, the execution process of the standardized operation steps is evaluated based on the constructed normative evaluation model. The scoring report module is used to generate an evaluation report that includes quantitative scores and risk points; The replay module is used to overlay the risk points as warning markers when reproducing the execution process of the simulation exercise in the virtual environment.

10. The power distribution network live-line operation simulation and standardization evaluation system according to claim 9, characterized in that, The system also includes a data storage module for storing operation records, evaluation model parameters, and historical quantitative scoring data, and supports data mining for model optimization; the 3D real-scene simulation module is implemented based on the Unity 3D or Unreal Engine platform.

Images