Power distribution cable lean operation inspection system and method based on digital twinning

Abstract

The invention discloses a distribution cable lean operation inspection system and method based on digital twinning. The system comprises a distribution cable digital twinning panoramic modeling module, a multi-source data acquisition and fusion module and other seven modules. Constructing a panoramic model through parametric modeling, laser scanning and oblique photography to display cable channel information; multi-source data are collected, features are extracted through a long-short-term memory network, the state of the cable is evaluated through an improved hidden Markov model, an operation and maintenance strategy is formulated according to intelligent decision analysis, and real-time data interaction and visual display are achieved. The method comprises the steps of model construction, data collection and fusion, feature extraction, state evaluation, decision analysis, data interaction, visual display and the like. Real-time monitoring, accurate evaluation and intelligent operation and maintenance of the distribution cable are realized, the operation and maintenance efficiency and accuracy can be effectively improved, potential faults can be found in time, the operation and maintenance cost is reduced, safe and stable operation of a power grid is guaranteed, and technical innovation and management upgrading of the power industry are promoted.

Description

Technical Field

[0001] The present invention relates to the field of distribution cable operation and maintenance, and in particular to a distribution cable lean operation and maintenance system and method based on digital twins. Background Art

[0002] With the rapid development of the national economy and the accelerated pace of urbanization, electricity demand continues to rise. As the "aorta" of urban power transmission, distribution cables are becoming increasingly important. In urban power grids, the number of distribution cable lines of 35 kV and below continues to grow. The cableization rate of some large urban power grids has exceeded 95%, with 10 kV cable lines accounting for over 95% of medium-voltage distribution cables. However, the underground cable lines in cities under the jurisdiction of State Grid Corporation of China are complex in structure, have diverse operating environments, and feature increasing equipment density. This exposes distribution cable lines to the risk of fire and damage leading to cross-section loss, seriously threatening the operational safety of urban power grids and easily causing power outages to localized or large-scale users in the city. Against this backdrop, improving the lean management of distribution cables and enhancing intelligent operation and maintenance capabilities is imperative. Consequently, a lean operation and maintenance system and method for distribution cables based on digital twins has emerged.

[0003] Current traditional cable operation and maintenance methods have numerous shortcomings. In terms of efficiency, they primarily rely on manual inspections, which are not only inefficient but also susceptible to human interference, making them unable to meet the growing demand for cable operation and maintenance. Real-time monitoring of cable operating conditions is severely inadequate, making it difficult to detect potential safety hazards in a timely manner. Failures are often delayed until they occur, leading to power outages that significantly impact social production and life. Furthermore, the slow response to cable faults and the lack of effective early warning and emergency response mechanisms further exacerbate the losses caused by failures.

[0004] In terms of information technology, many regions lack a unified and effective information management system for distribution cable operation and maintenance. This makes it difficult to integrate and analyze maintenance data, providing ineffective support for operational and maintenance decisions, and hindering efficient operation and maintenance. Furthermore, due to the high cost of equipment upgrades, distribution cable equipment in many cities is experiencing significant aging. This aging equipment not only experiences frequent failures but also poses serious safety risks. In summary, traditional cable operation and maintenance methods are no longer adaptable to the demands of modern power grids and urgently require innovation and improvement. Summary of the Invention

[0005] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a lean operation and inspection system and method for distribution cables based on digital twins.

[0006] The technical solution adopted by the present invention is a lean operation and inspection system for distribution cables based on digital twins, which includes:

[0007] Distribution Cable Digital Twin Panoramic Modeling Module: This module uses parametric modeling and laser scanning technology to integrate system modeling information on cable channel specifications and parameters with real-world data acquired through oblique photography to construct a panoramic digital twin model of the cable channel.

[0008] Multi-source data acquisition and fusion module: used to collect multi-source data through drones and real-scene on-site acquisition, and use data transmission technology to fuse the collected thermal infrared, ultrasonic, and ground wave partial discharge data;

[0009] Long-Short-Term Memory Network Data Feature Extraction Module: This module receives data transmitted by the multi-source data acquisition and fusion module and uses the long-short-term memory network algorithm to perform time series analysis on the operating data of the distribution cables. The gating mechanism of the long-short-term memory network algorithm filters and memorizes data of different time steps, automatically extracting long-term dependency features and short-term fluctuation features from the data.

[0010] Improved Hidden Markov Model Condition Assessment Module: This module is used to assess the condition of distribution cables using an improved Hidden Markov Model based on the feature vectors output by the Long Short-Term Memory Network Data Feature Extraction Module.

[0011] Intelligent decision-making analysis module: This module is used to generate operation and maintenance strategies for different cable states based on the evaluation results of the improved hidden Markov model state assessment module, combined with the historical data and operating parameters of the distribution cables, using logical reasoning and decision tree algorithms;

[0012] Real-time data interaction module: responsible for real-time data interaction with distribution cable field equipment, other related power systems, and operation and maintenance personnel's terminal equipment, while feeding back the system's analysis results and decision-making information to external systems;

[0013] Visualization display module: used to display the model constructed by the distribution cable digital twin panoramic modeling module, the operation and maintenance strategy generated by the intelligent decision-making analysis module, and other key data through visualization technology.

[0014] Furthermore, the long short-term memory network data feature extraction module, the improved long short-term memory network model formula is as follows:

[0015] i t =σ(W ii x t +b ii +W hi h t-1 +b hi )

[0016] f t =σ(W if x t +bif +W hf h t-1 +b hf )

[0017]

[0018] o t =σ(W io x t +b io +W ho h t-1 +b ho )

[0019] h t =o t ⊙tanh(C t )

[0020] Among them, x t Represents the data input to the long short-term memory network at the current time t, including the temperature parameter T of the distribution cable t , current parameter I t , partial discharge parameter Q t , x t =[T t , I t , Q t ];h t is the hidden state output at the current moment; C t is the cell state at the current moment; i t , f t , o t They are the activation values ​​of the input gate, forget gate, and output gate respectively; is the candidate cell state; σ is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, which means element-by-element multiplication; W ii , W if , W ic , W io , W hi , W hf , W hc , W ho is the weight matrix, b ii , b if , b ic , b io , b hi , b hf , b hc , b ho is the bias vector.

[0021] Furthermore, the adaptive learning rate adjustment formula of the improved long short-term memory network model is:

[0022]

[0023] Among them, η t is the learning rate at time t, η0 is the initial learning rate, and θ represents all parameters in the long short-term memory network, including the weight matrix and bias. When the short-term memory network learns the changes in the distribution cable temperature, current, and partial discharge parameters, it automatically adjusts the learning rate according to the severity of the parameter changes.

[0024] Furthermore, the improved hidden Markov model state evaluation module, the improved hidden Markov model formula is as follows:

[0025]

[0026] Where P(O|λ) means that under the model parameter λ, the observation sequence O = {O1, O2, ..., O T}The probability of occurrence, the observation sequence O t It is composed of the feature vector output by the long short-term memory network data feature extraction module, which contains features related to the operating status of the distribution cable; N is the number of hidden states of the hidden Markov model, corresponding to different operating status levels of the distribution cable; α t-1 (i) is the forward probability of being in hidden state i at time t-1; a ij is the state transition probability from hidden state i to hidden state j, and its value is obtained based on the statistical frequency of transitions between different states in the historical operation data of the distribution cable; b j (O t ) is in hidden state j, observing O t The observation probability is determined by statistical analysis of distribution cable parameters under different operating conditions; β t (j) is the remaining observation sequence O that can be generated from the hidden state j at time t t+1 , O t+2 ,…,O T The backward probability of .

[0027] Furthermore, in the improved hidden Markov model, the state transition probability a is dynamically adjusted. ij The formula is:

[0028]

[0029] in, is the adjusted state transition probability, ω is the weight coefficient, which is determined according to the importance of the distribution cable and historical fault data; ΔT ij is the change in the operating temperature of the distribution cable when it transitions from hidden state i to hidden state j, ΔT ij =T j -T i , Ti and T j are the temperatures of the distribution cables in hidden states i and j, respectively.

[0030] Furthermore, the intelligent decision analysis module uses the following decision model to determine the operation and maintenance strategy based on the improved hidden Markov model evaluation results and the features extracted by the long short-term memory network:

[0031]

[0032] Among them, D represents the final operation and maintenance decision, D set is the set of all possible operation and maintenance decisions, p(s|O) is the probability that the distribution cable is in the hidden state s under the observation sequence O, which is obtained by the improved hidden Markov model state evaluation module; S is the total number of hidden states; U(d, s) is the utility value of taking the operation and maintenance decision d when the cable is in the hidden state s.

[0033] Furthermore, when calculating the utility value U(d, s), the aging parameter A and the load factor parameter L of the distribution cable are considered, and the formula is as follows:

[0034] U(d,s)=R(d,s)-C(d)×(1+γ×A)×(1+δ×L)

[0035] Where R(d, s) is the benefit of taking the operation and maintenance decision d when the cable is in the hidden state s, and C(d) is the cost of taking the operation and maintenance decision d. γ and δ are weight coefficients determined based on the cable type and historical data. The aging parameter A is calculated based on the cable's service life and the number of historical failures. The load factor parameter L is determined based on the ratio of the cable's real-time current to its rated current.

[0036] Furthermore, the real-time data interaction module adopts the following encryption transmission model when exchanging data with external systems:

[0037]

[0038] Among them, E(M) is the encrypted data, M is the original data to be transmitted, including the operation data of the distribution cable and the operation and maintenance decision; K is the encryption key, which is generated by the system according to a specific algorithm and is related to the identification information and timestamp of the distribution cable; H(S) is the digital signature hash value of the data sender S, which contains the sender's identity information and timestamp. The receiver can use the reverse operation to Decrypt.

[0039] Furthermore, the visualization display module adopts the following dynamic visualization model when displaying the cable channel cross-section information:

[0040]

[0041] Where V(x, y, t) is the visual display value at the coordinate (x, y) at time t, which is used to present the status information of the equipment in the cable channel, n is the number of different types of data, and w i is the weight of each data, which is determined according to the importance of the data to the evaluation of the cable operation status, f i (x, y, t) is the function value of the i-th data at time t and coordinate (x, y).

[0042] The lean operation and inspection method for distribution cables based on digital twins includes the following steps:

[0043] S1. Constructing a panoramic digital twin model: Utilizing parametric modeling and laser scanning technology, combined with on-site data acquired through oblique photography, image and coordinate matching is performed on the equipment and scene materials within the cable channel. A virtual scene model is then created using 3D modeling software. Materials and textures are added, and lighting and shadows are set. The panoramic digital twin model of the cable channel is rendered and output, presenting the distribution and operation of key equipment within the cable channel, as well as the equipment space occupancy information, in multiple expression dimensions.

[0044] S2. Collect and fuse multi-source data: Use drones and real-world on-site data collection to collect data related to distribution cables and their ancillary equipment, various sensors, and pipeline corridor scene materials. Fusion of different types of data, including thermal infrared, ultrasonic, and ground-wave partial discharge, is performed according to specific formats and transmission protocols to provide data support for subsequent analysis.

[0045] S3. Extract features using a long short-term memory network: The fused data is fed into the long short-term memory network, which uses its gating mechanism to analyze the time series of distribution cable operation data, filter and memorize data at different time steps, and automatically extract long-term dependency features and short-term fluctuation features from the data, converting them into low-dimensional, representative feature vectors.

[0046] S4. Evaluate the state based on the improved hidden Markov model: Based on the feature vectors extracted from the long short-term memory network, an improved hidden Markov model is used to define multiple hidden states corresponding to different cable operating conditions, calculate the state transition probability and observation probability, and dynamically update the cable hidden state through the forward-backward algorithm to evaluate the cable operating status;

[0047] S5. Conduct intelligent decision analysis: Based on the evaluation results of the improved hidden Markov model, combined with the historical data and operating parameters of the distribution cables, logical reasoning and decision tree algorithms are used to consider the electrical, thermal, and mechanical characteristics of the cables, as well as the costs and effects of different operation and maintenance measures, to generate operation and maintenance strategies for different cable conditions.

[0048] S6. Real-time data interaction: Through network communication technology, the system interacts with distribution cable field equipment, other related power systems, and operation and maintenance personnel's terminal equipment in real time, receives external system instructions and data, and feeds back the system's analysis results and decision-making information to the external system;

[0049] S7. Visualize key information: The constructed digital twin panoramic model, generated operation and maintenance strategy, and other key data will be presented in the form of graphics, charts, and maps through visualization technology.

[0050] Beneficial effects: The present invention proposes a lean operation and inspection system and method for distribution cables based on digital twins. In terms of improving operation and maintenance efficiency and accuracy, the system uses digital twin panoramic modeling to intuitively present the distribution and operating status of cable channel equipment, allowing operation and maintenance personnel to quickly locate problems; the long-term and short-term memory network accurately extracts data features, and the improved hidden Markov model accurately evaluates the cable status, making operation and maintenance decisions more targeted, reducing unnecessary operation and maintenance work, and improving overall efficiency. From the perspective of ensuring the safe and stable operation of the power grid, through real-time monitoring and intelligent analysis, the system can promptly discover potential safety hazards, such as predicting faults based on changes in parameters such as cable temperature, current, and partial discharge, and take measures in advance to reduce the occurrence rate of faults, ensure reliable power supply to urban power grids, and reduce the risk of large-scale power outages due to cable faults. In terms of reducing operation and maintenance costs, intelligent decision-making analysis comprehensively considers factors such as operation and maintenance costs and power outage losses to formulate strategies to avoid excessive operation and maintenance; at the same time, accurate status assessment reduces blind equipment replacement, improves equipment utilization, and reduces waste of manpower and material resources. Moreover, this system and method have promoted technological innovation and management upgrades in the power industry. The application of advanced technologies such as digital twins, long short-term memory networks, and improved hidden Markov models has provided new directions for industry development. The full-process management system and intelligent control model it has constructed provide reference for the management of power companies, helping the industry as a whole to improve its management level and achieve sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a diagram of the system module composition of the present invention;

[0052] Figure 2 The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0053] It should be noted that, unless there is a conflict, the embodiments in this application and the features described in the embodiments can be combined with each other. The application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0054] like Figure 1 As shown in the figure, the lean operation and inspection system for distribution cables based on digital twins includes the following seven modules:

[0055] Power Distribution Cable Digital Twin Panoramic Modeling Module: This module, based on parametric modeling and laser scanning technology, integrates system modeling information on cable corridor specifications and parameters with real-world on-site data acquired through oblique photography to construct a panoramic digital twin model of the cable corridor. By performing image and coordinate matching on cables, lamps, fire extinguisher boxes, cable supports, manholes and covers, cable signage, various sensors, and corridor scene materials, a virtual scene model is created using 3D modeling software. Appropriate materials and textures are added to the model, along with lighting and shadow settings, for rendering output. This module uses three dimensions of presentation: a graphical cockpit, gridded fire zones, and a modeled real-world scene to intuitively present the distribution and operation of key equipment within the cable corridor, as well as their space occupancy information.

[0056] Specifically, the distribution cable digital twin panoramic modeling module is the foundation of the entire system. It undertakes the important task of constructing a cable channel model that is highly consistent between virtual and real-world scenarios. In today's complex urban power grids, distribution cables are distributed intricately and operate in diverse environments. This module intuitively presents various information within the cable channel, providing a clear scenario foundation for subsequent operation and inspection. It integrates multiple advanced technologies to fully map the real-world cable channel into virtual space.

[0057] This module is based on parametric modeling and laser scanning technology. Parametric modeling enables precise modeling based on the specific specifications of the cable channel, such as cable length, diameter, and material. Precisely setting these parameters ensures the model is highly consistent with the actual cable. Laser scanning technology is used to obtain information about the actual topography and equipment layout of the cable channel. Its scanning accuracy can reach millimeter levels, enabling the capture of subtle structural features. Integrating real-world data obtained through oblique photography further enriches the model's details. Oblique photography captures the cable channel and its surroundings from multiple angles, adding realistic textures and colors to the model. 3D modeling software such as 3dsMax and Maya, which offer powerful modeling and rendering capabilities, is used to construct the virtual scene model. After the model is constructed, appropriate materials and textures are applied, and lighting and shadows are configured for output before rendering. The selection of materials and textures is tailored to the specific conditions, such as the cable sheath material or the metal material of the bracket, to achieve a realistic effect. Lighting and shadow settings enhance the model's three-dimensionality and realism.

[0058] This module uses three dimensions: a graphical cockpit, gridded fire zones, and modeled real-world scenarios to visually display the distribution and operation of key equipment within a cable tunnel, as well as their space usage. The graphical cockpit displays various data charts showing cable operating parameters such as current, voltage, and temperature, allowing operators to quickly understand the overall operating status of the cable. The gridded fire zones divide the cable tunnel into multiple grid zones, enabling real-time monitoring and management of fire protection in each zone. This allows operators to promptly issue alerts if a fire hazard arises in a specific area. Modeled real-world scenarios offer the most intuitive display, allowing operators to navigate through a virtual model in three-dimensional space, visualizing every detail of the cable tunnel as if they were actually there. For example, within a cable tunnel in a large city, the module's model allows operators to clearly visualize the cable routing, the location of each device, and their connections. When maintenance is required, routes and plans can be planned in advance within the model, improving efficiency and safety.

[0059] Multi-source data acquisition and fusion module: This module uses drones and real-world on-site data collection to collect data related to equipment such as cables, lamps, fire extinguisher boxes, cable brackets, manholes and covers, cable sign maintenance, universal intelligent gateways, liquid level sensors, partial discharge sensors, oxygen sensors, smoke sensors, temperature sensors, temperature and humidity sensors, and AI dome cameras, as well as tunnel scene materials. Using data transmission technology, this module fuses different types of data, including thermal infrared, ultrasonic, and geomagnetic partial discharge data, and transmits it to subsequent modules according to specific data formats and transmission protocols.

[0060] Specifically, the multi-source data acquisition and fusion module is a crucial component of the system's information acquisition. During the operation of distribution cables, a wide variety of data is generated. This data comes from various devices and sensors, and has different types and formats. This module's role is to collect and integrate this dispersed and diverse data, providing comprehensive and accurate data support for subsequent analysis and decision-making.

[0061] Data related to cables, lamps, fire extinguisher boxes, cable supports, manholes and covers, cable signage maintenance, universal smart gateways, liquid level sensors, partial discharge sensors, oxygen sensors, smoke sensors, temperature sensors, temperature and humidity sensors, AI dome cameras, and other equipment, as well as tunnel scene footage, is collected through drones and real-time on-site data collection. Drones, with their flexible maneuverability, can quickly acquire data from large cable corridors. Their onboard high-definition cameras and sensors can capture various data types, including images, temperature, and humidity. Real-time on-site data collection involves manual or automated on-site data collection to ensure accuracy and real-time data. Different sensor types have different accuracies and sampling frequencies during data collection. For example, temperature sensors can achieve an accuracy of ±0.1°C, and the sampling frequency can be set to once per minute or higher as needed. Partial discharge sensors can detect minute partial discharge signals, providing important information for cable insulation condition monitoring. Collected data, including thermal infrared, ultrasonic, and ground-wave partial discharge data, is integrated according to specific data formats and transmission protocols. The unification of data formats can facilitate subsequent data processing and analysis. The selection of transmission protocols should consider the security and real-time nature of the data, such as using the TCP / IP protocol for data transmission.

[0062] The collection and fusion of multi-source data can provide the system with comprehensive information, helping operations and maintenance personnel better understand the operating status of the cables. For example, a liquid level sensor can monitor the water level in the cable trench in real time. When the water level exceeds a certain threshold, an alarm will be issued in time to prevent the cable from being damaged by flooding. A smoke sensor can detect whether smoke is generated in the cable channel. Once smoke is detected, it indicates that there may be a fire hazard, and the system will immediately notify the operations and maintenance personnel to deal with it. Temperature and humidity sensors can monitor changes in temperature and humidity in the cable channel. Excessive temperature and humidity can affect the insulation performance of the cable. Through real-time monitoring, timely measures can be taken to adjust them. In an actual cable corridor, through the multi-source data collection and fusion module, operations and maintenance personnel can simultaneously obtain information such as the cable's electrical parameters, environmental parameters, and the operating status of the equipment, thereby comprehensively and accurately assessing the cable's operating status, identifying potential problems in advance, and addressing them.

[0063] Long-Short-Term Memory Network Data Feature Extraction Module: This module receives data from the Multi-Source Data Acquisition and Fusion Module and uses the Long-Short-Term Memory Network (LSTM) algorithm to perform time series analysis on the distribution cable operating data. Using the LSTM's gating mechanism, it filters and memorizes data at different time steps, automatically extracting long-term dependency features and short-term fluctuation characteristics from the data. This module converts high-dimensional, complex raw data into low-dimensional, representative feature vectors for processing by subsequent modules.

[0064] Specifically, the Long Short-Term Memory (LSTM) data feature extraction module is a key step in deep mining collected multi-source data. Distribution cable operational data exhibits time series characteristics, with complex dependencies between data. Traditional data analysis methods struggle to effectively process this complex time series data. However, the Long Short-Term Memory (LSTM) network offers unique advantages, capable of automatically extracting both long-term dependency features and short-term fluctuation characteristics from the data.

[0065] This module receives data from the multi-source data acquisition and fusion module and processes it using the Long Short-Term Memory (LSTM) algorithm. An LSTM is a special type of recurrent neural network that uses a gating mechanism to control the flow and memory of information. An LSTM consists of three gating units: the input gate, the forget gate, and the output gate. The input gate determines how much information from the current input data is allowed into the cell state; the forget gate determines what information in the cell state is forgotten; and the output gate determines how much information in the cell state is output to the next time step. When training an LSTM model, several parameters must be set, such as the learning rate, the number of hidden layer neurons, and the number of training iterations. The learning rate controls the step size for model parameter updates. Excessively high learning rates can prevent the model from convergence, while excessively low learning rates can slow training. The number of hidden layer neurons affects the model's expressiveness and needs to be adjusted based on the complexity of the data and the task requirements. The number of training iterations determines how fully the model is trained, and an appropriate number of iterations is typically determined through experimentation.

[0066] By extracting data features using a long short-term memory (LSTM) network, high-dimensional, complex raw data can be converted into low-dimensional, representative feature vectors. These feature vectors can more effectively reflect the operating status of cables, providing strong support for subsequent condition assessment and decision-making analysis. For example, during cable operation, parameters such as temperature and current change over time. These changes may indicate early signs of cable failure. By extracting features from these parameters using an LSTM, patterns and trends hidden in the data can be discovered. If cable temperature continues to rise over a period of time, and this rise shows a certain regularity, it may indicate an overheating fault. Based on these feature vectors, maintenance personnel can take timely measures to inspect and maintain the cable, preventing further development of the fault. In an actual power distribution system, data feature extraction using an LSTM successfully predicted cable insulation aging, enabling early cable replacement and avoiding a major power outage.

[0067] Improved Hidden Markov Model State Assessment Module: Based on the feature vectors output by the Long Short-Term Memory Network Data Feature Extraction Module, an improved Hidden Markov Model (HMM) is used to assess the state of distribution cables. This module defines multiple hidden states, each corresponding to a different cable operating condition. State transition probabilities and observation probabilities are calculated based on the feature vectors. Using a forward-backward algorithm, the hidden states of the cable are dynamically updated, providing an accurate assessment of the cable's operating status.

[0068] Specifically, the improved Hidden Markov Model (HMM) state assessment module is the core module for evaluating the operating status of distribution cables. In actual operation, the cable's state is hidden and cannot be directly observed. Its state can only be inferred through some observable parameters. The Hidden Markov Model (HMM) is a probabilistic model used to process the relationship between hidden states and observation sequences. The improved HMM can more accurately assess the operating status of cables.

[0069] This module uses an improved hidden Markov model (HMM) to assess the condition of distribution cables based on the feature vectors output by the long-short-term memory (LSTM) data feature extraction module. The improved HMM defines multiple hidden states, each corresponding to a different cable operating condition, such as normal, minor fault, or major fault. State transition probabilities and observation probabilities are calculated based on the feature vectors. The state transition probability represents the probability of transitioning from one hidden state to another, while the observation probability represents the probability of observing a particular feature vector in a given hidden state. The calculation of these probabilities requires extensive historical data and statistical analysis. In practical applications, improvements to the traditional HMM are often made, such as introducing dynamically adjusted state transition and observation probabilities. These factors, combined with real-time operating and environmental parameters of the distribution cable, allow the model to better adapt to changes in cable conditions. During model training, a forward-backward algorithm is used to estimate model parameters, namely, the state transition and observation probabilities. This iterative algorithm continuously updates parameters to minimize the error between the model output and the observed data.

[0070] The improved Hidden Markov Model (HMM) state assessment module enables accurate assessment and prediction of cable operating conditions. For example, in a cable network, this module evaluated the status of a cable and revealed that its hidden state gradually shifted from normal to a minor fault. Based on this assessment, maintenance personnel can promptly conduct further inspection and maintenance on the cable to prevent the fault from worsening. In another case, through real-time monitoring and condition assessment of the cable, an insulation breakdown failure was predicted in advance, and emergency measures were implemented before the failure occurred, ensuring the safe and stable operation of the power grid. This module can provide maintenance personnel with timely and accurate status information, helping them formulate appropriate maintenance strategies and improving the reliability and service life of the cable.

[0071] Intelligent Decision Analysis Module: Based on the evaluation results of the improved hidden Markov model state assessment module, combined with historical data and operating parameters of the distribution cables, this module uses logical reasoning and decision tree algorithms to generate operation and maintenance strategies tailored to different cable conditions. This module considers parameters such as the cable's electrical, thermal, and mechanical properties, as well as the costs and effectiveness of different operation and maintenance measures, providing operators with reasonable decision-making recommendations.

[0072] Specifically, the intelligent decision-making analysis module is the decision-making core of the entire system. It provides operators with reasonable operation and maintenance strategies based on the cable's operating status assessment results, combined with historical data and operating parameters. The operation and maintenance of distribution cables requires consideration of multiple factors, such as operation and maintenance costs, power outage losses, and repair probability. Traditional decision-making methods often fail to comprehensively and accurately account for these factors. However, the intelligent decision-making analysis module uses scientific algorithms and models to make optimal decisions.

[0073] This module uses the evaluation results of the improved hidden Markov model condition assessment module, combined with historical data and operating parameters of the distribution cables, to conduct analysis using logical reasoning and decision tree algorithms. Logical reasoning derives reasonable conclusions based on known conditions and rules; the decision tree algorithm, a tree-structured decision-making method, constructs a decision tree model by classifying and partitioning data, thereby providing appropriate decision recommendations for different situations. The decision-making process considers parameters such as the electrical, thermal, and mechanical properties of the cable, as well as the costs and effectiveness of different maintenance measures. For example, for different types of cable faults, different maintenance strategies are formulated based on their severity and scope, such as immediate maintenance, regular monitoring, and cable replacement. When evaluating the effectiveness of each maintenance strategy, factors such as maintenance costs, outage losses, and repair probability are comprehensively considered, and the feasibility and effectiveness of each strategy are evaluated using quantitative indicators.

[0074] The intelligent decision-making analysis module provides scientific and rational decision support for O&M personnel, improving the efficiency and effectiveness of O&M work. For example, in a cable failure case, the module's analysis revealed a minor insulation fault in a cable. Immediate repair could resolve the problem promptly, but it would result in a certain power outage. Regular monitoring, while potentially preventing power outages, carries the risk of the fault worsening. Taking into account factors such as O&M costs, power outage losses, and repair probability, the module recommends regular monitoring and sets an appropriate monitoring cycle. During subsequent monitoring, the module closely monitors changes in the cable's status and initiates repairs promptly if the fault shows signs of worsening. This ensures the safe operation of the power grid while reducing O&M costs. In another case, for a severely aged cable, the module, based on its historical data and current status, recommended timely replacement, thus avoiding a potential major failure and power outage.

[0075] Real-time Data Interaction Module: Responsible for real-time data interaction with distribution cable field equipment, other related power systems, and the terminal devices of operation and maintenance personnel. Through network communication technology, it transmits and shares real-time data with systems such as the power supply service command system, the power grid GIS platform, and the unified authority platform. It receives instructions and data from external systems and feeds back analysis results and decision-making information from the system to external systems.

[0076] Specifically, the real-time data exchange module bridges information communication between the system and the outside world. It is responsible for exchanging real-time data with distribution cable field equipment, other related power systems, and the terminal devices of operators and maintenance personnel. In modern power systems, close collaboration is required between all links, and the sharing and transmission of real-time data is crucial to ensuring the safe and stable operation of the power grid.

[0077] This module uses network communication technology to achieve data transmission and sharing with systems such as the power supply service command system, power grid GIS platform, and unified authority platform. Network communication technology can use wired or wireless networks, such as Ethernet, Wi-Fi, 4G / 5G, etc., and the appropriate communication method should be selected according to the actual situation. During the data transmission process, certain communication protocols, such as Modbus, IEC61850, etc., need to be followed to ensure the accuracy and reliability of the data. This module also supports receiving instructions and data from external systems, and at the same time feeds back the analysis results and decision information of this system to the external system. In order to ensure the security of the data, encryption technology will be used to encrypt the data to prevent the data from being stolen or tampered with during transmission. For example, the SSL / TLS protocol is used to encrypt the data for transmission to ensure the confidentiality and integrity of the data.

[0078] The real-time data interaction module can realize information sharing and collaborative work between the system and external systems, and improve the overall operating efficiency of the power system. For example, when a power grid failure occurs, the power supply service command system can obtain the operating status information of the distribution cable through the real-time data interaction module, promptly deploy emergency repair resources, and quickly restore power supply. The power grid GIS platform can provide the system with geographical information of the cables, helping operation and maintenance personnel to better understand the distribution of cables and plan maintenance routes. The unified authority platform can manage the access rights of the system to ensure that only authorized personnel can access and operate relevant data. In an actual power system, the real-time data interaction module realizes the seamless connection between the lean operation and inspection system of the distribution cable and other systems, improves the response speed and accuracy of fault handling, and ensures the safe and stable operation of the power grid.

[0079] Visualization Display Module: This module uses visualization technology to display the model constructed by the distribution cable digital twin panoramic modeling module, the operation and maintenance strategies generated by the intelligent decision-making analysis module, and other key data. Using graphics, charts, maps, and other formats, it intuitively presents cable channel cross-sectional information, equipment operating status, and operation and maintenance task progress, providing a clear and intuitive interface for operation and maintenance personnel.

[0080] Specifically, the visualization module serves as the user interface for system interaction. It presents complex cable operation data and analysis results to maintenance personnel in an intuitive and accessible manner. Distribution cable maintenance requires extensive data and information processing and interpretation. Without a robust visualization interface, it can be difficult for maintenance personnel to quickly and accurately access critical information. The visualization module helps maintenance personnel better understand cable operation status and make informed decisions.

[0081] This module uses visualization technology to display the model constructed by the distribution cable digital twin panoramic modeling module, the operation and maintenance strategies generated by the intelligent decision-making analysis module, and other key data. Visualization technologies include graphs, charts, maps, and other formats, such as bar charts, line charts, pie charts, and geographic information maps. When displaying cable channel cross-sectional information, 3D visualization technology can be used to enable operators to intuitively understand the internal structure and operating status of the cable. When displaying equipment operating status, different colors and icons can be used to indicate normal, abnormal, and faulty equipment states, allowing operators to quickly identify problems. To enhance visualization, animation and interactive technologies such as dynamic charts and interactive maps are also used. Operators can view detailed information and data by clicking and dragging the mouse. The display interface design adheres to ergonomic principles, with a rational layout of various elements to improve readability and ease of use.

[0082] The visualization module can provide maintenance personnel with a clear and intuitive interface, improving work efficiency and decision-making accuracy. For example, in a large cable management center, maintenance personnel can use the visualization module to view the operating status of each cable channel, equipment distribution, and the progress of maintenance tasks in real time on a large screen. When a cable fails, the system will highlight the fault location on a map and display relevant fault information and handling suggestions. Based on this information, maintenance personnel can quickly develop emergency repair plans and deploy personnel and supplies. In daily maintenance work, maintenance personnel can view historical cable data and trend analysis through the visualization interface, understand the cable's operating patterns, and make maintenance plans in advance. The visualization module can also provide decision-making support for managers. By visually displaying overall data, managers can understand the operating status of the cable system and formulate reasonable development plans and investment strategies.

[0083] The above seven modules are connected in sequence. The distribution cable digital twin panoramic modeling module outputs the constructed model to the visualization display module; the multi-source data acquisition and fusion module collects and fuses the data and transmits it to the long and short-term memory network data feature extraction module; the long and short-term memory network data feature extraction module passes the extracted feature vector to the improved hidden Markov model state evaluation module; the improved hidden Markov model state evaluation module transmits the evaluation results to the intelligent decision analysis module; the decision information generated by the intelligent decision analysis module is transmitted to the relevant systems and equipment through the real-time data interaction module, and the decision results are output to the visualization display module; the real-time data interaction module is responsible for the data interaction between each module and the external system, and the visualization display module is used to present the key data and analysis results of each module.

[0084] Preferably, in the long short-term memory network data feature extraction module, the improved long short-term memory network model formula is as follows:

[0085] i t =σ(W ii x t +b ii +W hi h t-1 +b hi )

[0086] f t =σ(W if x t +b if +W hf h t-1 +b hf )

[0087]

[0088] ot =σ(W io x t +b io +W ho h t-1 +b ho )

[0089] h t =o t ⊙tanh(C t )

[0090] Among them, x t Represents the data input to the long short-term memory network at the current time t, including the temperature parameter T of the distribution cable t , current parameter I t , partial discharge parameter Q t Wait, x t =[T t , I t , Q t ];h t is the hidden state output at the current moment; C t is the cell state at the current moment; i t , f t , o t They are the activation values ​​of the input gate, forget gate, and output gate respectively; is the candidate cell state; σ is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, which means element-by-element multiplication; W ii , W if , W ic , W io , W hi , W hf , W hc , W ho is the weight matrix, b ii , b if , b ic , b io , b hi , b hf , b hc , b ho The improved model introduces an adaptive learning rate mechanism to dynamically adjust the learning rate according to the changes in different parameters of the distribution cable, thereby enhancing the learning ability of complex changes in cable parameters.

[0091] Preferably, in the improved long short-term memory network model, the adaptive learning rate adjustment formula is:

[0092]

[0093] Among them, η tis the learning rate at time t, η0 is the initial learning rate, and θ represents all parameters in the long short-term memory network (including the weight matrix and the bias matrix). When the short-term memory network learns the changes in parameters such as the temperature, current, and partial discharge of the distribution cable, it can automatically adjust the learning rate according to the severity of the parameter changes. For parameters that change dramatically, the learning rate is reduced to avoid over-learning; for parameters that change gently, the learning rate is appropriately increased to speed up the convergence speed, thereby extracting data features more accurately.

[0094] Preferably, in the improved hidden Markov model state evaluation module, the improved hidden Markov model formula is as follows:

[0095]

[0096] Where P(O|λ) means that under the model parameter λ, the observation sequence O = {O1, O2, ..., O T}The probability of occurrence, the observation sequence O t It is composed of the feature vector output by the long short-term memory network data feature extraction module, which contains features related to the operating status of the distribution cable, such as the temperature change trend feature F T，t , current fluctuation characteristics F I，t etc.; N is the number of hidden states of the hidden Markov model, corresponding to different operating status levels of the distribution cable, such as normal state, minor fault state, serious fault state, etc.; α t-1 (i) is the forward probability of being in hidden state i at time t-1; a ij is the state transition probability from hidden state i to hidden state j, and its value is obtained based on the statistical frequency of transitions between different states in the historical operation data of the distribution cable; b j (O t ) is in hidden state j, observing O t The observation probability is determined by statistical analysis of distribution cable parameters under different operating conditions; β t (j) is the remaining observation sequence O that can be generated from the hidden state j at time t t+1 , O t+2 ,…,O T The improved model introduces dynamically adjusted state transition probability and observation probability, and combines the real-time operating parameters and environmental parameters of the distribution cable to more accurately evaluate the cable operating status.

[0097] Preferably, in the improved hidden Markov model, the state transition probability a is dynamically adjusted. ij The formula is:

[0098]

[0099] in, is the adjusted state transition probability, ω is the weight coefficient, which is determined according to the importance of the distribution cable and historical fault data; ΔT ij is the change in the operating temperature of the distribution cable when it transitions from hidden state i to hidden state j, ΔT ij =T j -T i , T i and T j are the temperatures of the distribution cable in hidden states i and j, respectively. When the cable temperature changes significantly, the state transition probability is appropriately increased to reflect the impact of temperature on the cable state change. When the temperature change is small, the state transition probability is kept relatively stable to more accurately simulate the cable state transition process.

[0100] Preferably, in the intelligent decision analysis module, the following decision model is used to determine the operation and maintenance strategy based on the evaluation results of the improved hidden Markov model and the features extracted by the long short-term memory network:

[0101]

[0102] Among them, D represents the final operation and maintenance decision, D set is the set of all possible O&M decisions, such as "immediate maintenance" and "periodic monitoring." p(s|O) is the probability that the distribution cable is in hidden state s under observation sequence O, derived by the improved hidden Markov model state assessment module. S is the total number of hidden states. U(d, s) is the utility value of taking O&M decision d when the cable is in hidden state s. This utility value comprehensively considers factors such as O&M cost, outage loss, and repair probability. For example, the "immediate maintenance" decision has a higher O&M cost, but the outage loss may be lower, and the repair probability is higher. The "periodic monitoring" decision has a lower O&M cost, but may face a certain outage risk. This decision model selects the optimal O&M strategy by considering multiple factors.

[0103] Preferably, when calculating the utility value U(d, s), the aging parameter A and the load rate parameter L of the distribution cable are considered, and the formula is as follows:

[0104] U(d,s)=R(d,s)-C(d)×(1+γ×A)×(1+δ×L)

[0105] Here, R(d, s) represents the benefit of adopting O&M decision d when the cable is in hidden state s, such as the benefit of reducing power outage duration; C(d) represents the cost of adopting O&M decision d; γ and δ are weighting coefficients determined based on cable type and historical data; the aging parameter A is calculated based on factors such as the cable's service life and number of historical failures, and the load factor parameter L is determined by the ratio of the cable's real-time current to its rated current. This formula allows for more comprehensive consideration of the cable's aging and load conditions during the decision-making process. For cables with severe aging or high load factors, the weighting of O&M costs is appropriately increased, leading to more reasonable O&M decisions.

[0106] Preferably, the real-time data interaction module adopts the following encryption transmission model when exchanging data with an external system:

[0107]

[0108] Among them, E(M) is the encrypted data, M is the original data to be transmitted, including the operation data of the distribution cable, operation and maintenance decisions, etc.; K is the encryption key, which is generated by the system according to a specific algorithm and is related to the identification information and timestamp of the distribution cable; H(S) is the digital signature hash value of the data sender S, and the digital signature contains the sender's identity information, timestamp, etc. The receiver performs the reverse operation This encrypted transmission model combines the identification information of the distribution cable and the sender's digital signature to improve the security and integrity of data transmission and prevent data from being tampered with or stolen during transmission.

[0109] Preferably, the visualization display module adopts the following dynamic visualization model when displaying the cable channel cross-section information:

[0110]

[0111] Where V(x, y, t) is the visual display value at the coordinate (x, y) at time t, which is used to present the status information of the equipment in the cable channel, such as temperature, current, etc.; n is the number of different types of data, such as temperature data, current data, etc.; w i is the weight of each data, which is determined according to the importance of the data to the evaluation of the cable operation status, such as the temperature data weight w T and current data weight w I ,and The function value of the i-th data type at time t and coordinate (x, y) is obtained by interpolating and transforming the collected raw data. This dynamic visualization model weights the display effect according to the importance of different data, highlighting key data and enabling operators to more intuitively understand key information within the cable channel.

[0112] like Figure 2 As shown in the figure, the lean operation and inspection method of distribution cables based on digital twins includes the following steps:

[0113] S1. Construct a panoramic digital twin model: Utilize parametric modeling and laser scanning technology, combined with field data obtained through oblique photography, to perform image and coordinate matching of equipment and scene materials within the cable channel. Use 3D modeling software to build a virtual scene model, add materials and textures, set lighting and shadows, and render and output a panoramic digital twin model of the cable channel. This model presents the distribution and operation of key equipment within the cable channel, as well as the space occupancy information of the equipment, in multiple expression dimensions.

[0114] S2. Collect and fuse multi-source data: Collect data related to distribution cables and their ancillary equipment, various sensors, and pipeline corridor scene materials through drones, real-life on-site collection, etc., and fuse different types of data such as thermal infrared, ultrasonic, and ground wave partial discharge according to specific formats and transmission protocols to provide data support for subsequent analysis.

[0115] S3. Extract features using long-short-term memory networks: Input the fused data into the long-short-term memory network, analyze the time series of distribution cable operation data through its gating mechanism, filter and memorize data of different time steps, automatically extract long-term dependency features and short-term fluctuation features in the data, and convert them into low-dimensional, representative feature vectors.

[0116] S4. Status evaluation based on improved hidden Markov model: Based on the feature vector extracted by the long short-term memory network, an improved hidden Markov model is used to define multiple hidden states corresponding to different cable operating conditions, calculate the state transition probability and observation probability, and dynamically update the cable hidden state through the forward-backward algorithm to evaluate the cable operating status.

[0117] S5. Conduct intelligent decision-making analysis: Based on the evaluation results of the improved hidden Markov model, combined with the historical data and operating parameters of the distribution cables, using logical reasoning and decision tree algorithms, considering the electrical, thermal, mechanical and other characteristics of the cables as well as the costs and effects of different operation and maintenance measures, generate operation and maintenance strategies for different cable states.

[0118] S6. Conduct real-time data interaction: Through network communication technology, conduct real-time data interaction with distribution cable field equipment, other related power systems, and operation and maintenance personnel terminal equipment, receive external system instructions and data, and at the same time feed back the system's analysis results and decision information to the external system.

[0119] S7. Visualize key information: The constructed digital twin panoramic model, generated operation and maintenance strategy, and other key data are displayed in graphics, charts, maps, etc. through visualization technology.

[0120] The lean operation and maintenance system and method of distribution cables based on digital twins have significant advantages and effectively overcome many shortcomings of existing technologies.

[0121] In terms of operation and inspection efficiency, the traditional method that relies on manual inspections is inefficient and easily interfered with by human factors. This system and method uses a multi-source data acquisition and fusion module, and utilizes drones, various sensors and other equipment to achieve automated, multi-dimensional data collection, greatly improving the speed of data acquisition. The long-short-term memory network data feature extraction module and the improved hidden Markov model state assessment module work together to quickly and accurately analyze the cable operation status, changing the previous slow and inaccurate manual analysis. For example, in complex cable networks, traditional manual inspections find it difficult to cover large areas in a short period of time, but this system can quickly collect and analyze data, promptly identify potential problems, and greatly improve operation and inspection efficiency.

[0122] In terms of real-time monitoring and fault warning capabilities, existing technologies are insufficient for real-time monitoring of cable operating conditions and lack an effective early warning mechanism. The new system uses a digital twin panoramic modeling module to construct a realistic cable channel model, combined with real-time collected data, to achieve comprehensive real-time monitoring of the cable operating status. Once an anomaly occurs, such as when parameters such as temperature, current, and partial discharge deviate from the normal range, the system will issue a timely warning based on the improved model. In contrast, traditional operation and maintenance methods can often only handle faults after they occur. For example, cable failures caused by overheating in the past were often discovered only when obvious abnormalities appeared. However, the new system can detect abnormal temperature changes in advance, arrange maintenance in advance, and effectively avoid faults.

[0123] From the perspective of information integration and management, the current level of informatization in distribution cable operation and inspection in many regions is low, and there is a lack of a unified information management system, resulting in fragmented data and difficulty in analysis and utilization. The digital twin-based system uses a real-time data interaction module to share and interact with multiple systems, such as the power supply service command system and the power grid GIS platform, enabling centralized data management and in-depth analysis. The visualization module presents complex data in intuitive graphics and charts, making it easier for operations and maintenance personnel to understand and make decisions. For example, when troubleshooting, operations and maintenance personnel can quickly obtain information such as cable location, surrounding environment, and historical data through a visual interface to develop a scientific repair plan. Traditional methods, however, are fraught with difficulties in obtaining this information, seriously affecting troubleshooting efficiency.

[0124] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The lean operation and inspection system for distribution cables based on digital twins is characterized by: The system includes: Distribution Cable Digital Twin Panoramic Modeling Module: This module uses parametric modeling and laser scanning technology to integrate system modeling information on cable channel specifications and parameters with real-world data acquired through oblique photography to construct a panoramic digital twin model of the cable channel. Multi-source data acquisition and fusion module: used to collect multi-source data through drones and real-scene on-site acquisition, and use data transmission technology to fuse the collected thermal infrared, ultrasonic, and ground wave partial discharge data; Long-Short-Term Memory Network Data Feature Extraction Module: This module receives data transmitted by the multi-source data acquisition and fusion module and uses the long-short-term memory network algorithm to perform time series analysis on the operating data of the distribution cables. The gating mechanism of the long-short-term memory network algorithm filters and memorizes data of different time steps, automatically extracting long-term dependency features and short-term fluctuation features from the data. Improved Hidden Markov Model Condition Assessment Module: This module is used to assess the condition of distribution cables using an improved Hidden Markov Model based on the feature vectors output by the Long Short-Term Memory Network Data Feature Extraction Module. Intelligent decision-making analysis module: This module is used to generate operation and maintenance strategies for different cable states based on the evaluation results of the improved hidden Markov model state assessment module, combined with the historical data and operating parameters of the distribution cables, using logical reasoning and decision tree algorithms; Real-time data interaction module: responsible for real-time data interaction with distribution cable field equipment, other related power systems, and operation and maintenance personnel's terminal equipment, while feeding back the system's analysis results and decision-making information to external systems; Visualization display module: used to display the model constructed by the distribution cable digital twin panoramic modeling module, the operation and maintenance strategy generated by the intelligent decision-making analysis module, and other key data through visualization technology.

2. The lean operation and inspection system for distribution cables based on digital twins according to claim 1 is characterized in that: The long short-term memory network data feature extraction module and the improved long short-term memory network model formula are as follows: i t =σ(W ii x t +b ii +W hi h t-1 +b hi ) f t =σ(W if x t +b if +W hf h t-1 +b hf ) o t =σ(W io x t +b io +W ho h t-1 +b ho ) h t =o t ⊙tanh(C t ) Among them, x t Represents the data input to the long short-term memory network at the current time t, including the temperature parameter T of the distribution cable t , current parameter I t , partial discharge parameter Q t , x t =[T t , I t , Q t ];h t is the hidden state output at the current moment; C t is the cell state at the current moment; i t , f t , o t They are the activation values ​​of the input gate, forget gate, and output gate respectively; is the candidate cell state; σ is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, which means element-by-element multiplication; W ii , W if , W ic , W io , W hi , W hf , W hc , W ho is the weight matrix, b ii , b if , b ic , b io , b hi , b hf , b hc , b ho is the bias vector.

3. The lean operation and inspection system for distribution cables based on digital twins according to claim 2 is characterized in that: The improved long short-term memory network model, the adaptive learning rate adjustment formula is: Among them, η t is the learning rate at time t, η0 is the initial learning rate, and θ represents all parameters in the long short-term memory network, including the weight matrix and bias. When the short-term memory network learns the changes in the distribution cable temperature, current, and partial discharge parameters, it automatically adjusts the learning rate according to the severity of the parameter changes.

4. The lean operation and inspection system for distribution cables based on digital twins according to claim 1 is characterized in that: The improved hidden Markov model state evaluation module, the improved hidden Markov model formula is as follows: Where P(O|λ) means that under the model parameter λ, the observation sequence O = {O1, O2, ..., O T }The probability of occurrence, the observation sequence O t It is composed of the feature vector output by the long short-term memory network data feature extraction module, which contains features related to the operating status of the distribution cable; N is the number of hidden states of the hidden Markov model, corresponding to different operating status levels of the distribution cable; α t-1 (i) is the forward probability of being in hidden state i at time t-1; a ij is the state transition probability from hidden state i to hidden state j, and its value is obtained based on the statistical frequency of transitions between different states in the historical operation data of the distribution cable; b j (O t ) is in hidden state j, observing O t The observation probability is determined by statistical analysis of distribution cable parameters under different operating conditions; β t (j) is the remaining observation sequence O that can be generated from the hidden state j at time t t+1 , O t+2 ,…,O T The backward probability of .

5. The lean operation and inspection system for distribution cables based on digital twins according to claim 4 is characterized in that: In the improved hidden Markov model, the state transition probability a is dynamically adjusted ij The formula is: in, is the adjusted state transition probability, ω is the weight coefficient, which is determined according to the importance of the distribution cable and historical fault data; ΔT ij is the change in the operating temperature of the distribution cable when it transitions from hidden state i to hidden state j, ΔT ij =T j -T i , T i and T j are the temperatures of the distribution cables in hidden states i and j, respectively.

6. The lean operation and inspection system for distribution cables based on digital twins according to claim 1 is characterized in that: The intelligent decision analysis module uses the following decision model to determine the operation and maintenance strategy based on the improved hidden Markov model evaluation results and the features extracted by the long short-term memory network: Among them, D represents the final operation and maintenance decision, D set is the set of all possible operation and maintenance decisions, p(s|O) is the probability that the distribution cable is in the hidden state s under the observation sequence O, which is obtained by the improved hidden Markov model state evaluation module; S is the total number of hidden states; U(d, s) is the utility value of taking the operation and maintenance decision d when the cable is in the hidden state s.

7. The lean operation and inspection system for distribution cables based on digital twins according to claim 6 is characterized in that: When calculating the utility value U(d, s), the aging parameter A and the load factor parameter L of the distribution cable are considered. The formula is as follows: U(d,s)=R(d,s)-C(d)×(1+γ×A)×(1+δ×L) Where R(d, s) is the benefit of taking the operation and maintenance decision d when the cable is in the hidden state s, and C(d) is the cost of taking the operation and maintenance decision d. γ and δ are weight coefficients determined based on the cable type and historical data. The aging parameter A is calculated based on the cable's service life and the number of historical failures. The load factor parameter L is determined based on the ratio of the cable's real-time current to its rated current.

8. The lean operation and inspection system for distribution cables based on digital twins according to claim 1 is characterized in that: The real-time data interaction module uses the following encryption transmission model when exchanging data with external systems: Among them, E(M) is the encrypted data, M is the original data to be transmitted, including the operation data of the distribution cable and the operation and maintenance decision; K is the encryption key, which is generated by the system according to a specific algorithm and is related to the identification information and timestamp of the distribution cable; H(S) is the digital signature hash value of the data sender S, which contains the sender's identity information and timestamp. The receiver can use the reverse operation to Decrypt.

9. The lean operation and inspection system for distribution cables based on digital twins according to claim 1 is characterized in that: The visualization module uses the following dynamic visualization model when displaying the cable channel cross-section information: Where V(x, y, t) is the visual display value at the coordinate (x, y) at time t, which is used to present the status information of the equipment in the cable channel, n is the number of different types of data, and w i is the weight of each data, which is determined according to the importance of the data to the evaluation of the cable operation status, f i (x, y, t) is the function value of the i-th data at time t and coordinate (x, y).

10. The lean operation and inspection method of distribution cables based on digital twins is characterized by: The following steps are involved: S1. Constructing a panoramic digital twin model: Utilizing parametric modeling and laser scanning technology, combined with on-site data acquired through oblique photography, image and coordinate matching is performed on the equipment and scene materials within the cable channel. A virtual scene model is then created using 3D modeling software. Materials and textures are added, and lighting and shadows are set. The panoramic digital twin model of the cable channel is rendered and output, presenting the distribution and operation of key equipment within the cable channel, as well as the equipment space occupancy information, in multiple expression dimensions. S2. Collect and fuse multi-source data: Use drones and real-world on-site data collection to collect data related to distribution cables and their ancillary equipment, various sensors, and pipeline corridor scene materials. Fusion of different types of data, including thermal infrared, ultrasonic, and ground-wave partial discharge, is performed according to specific formats and transmission protocols to provide data support for subsequent analysis. S3. Extract features using a long short-term memory network: The fused data is fed into the long short-term memory network, which uses its gating mechanism to analyze the time series of distribution cable operation data, filter and memorize data at different time steps, and automatically extract long-term dependency features and short-term fluctuation features from the data, converting them into low-dimensional, representative feature vectors. S4. Evaluate the state based on the improved hidden Markov model: Based on the feature vectors extracted from the long short-term memory network, an improved hidden Markov model is used to define multiple hidden states corresponding to different cable operating conditions, calculate the state transition probability and observation probability, and dynamically update the cable hidden state through the forward-backward algorithm to evaluate the cable operating status; S5. Conduct intelligent decision analysis: Based on the evaluation results of the improved hidden Markov model, combined with the historical data and operating parameters of the distribution cables, logical reasoning and decision tree algorithms are used to consider the electrical, thermal, and mechanical characteristics of the cables, as well as the costs and effects of different operation and maintenance measures, to generate operation and maintenance strategies for different cable conditions. S6. Real-time data interaction: Through network communication technology, the system interacts with distribution cable field equipment, other related power systems, and operation and maintenance personnel's terminal equipment in real time, receives external system instructions and data, and feeds back the system's analysis results and decision-making information to the external system; S7. Visualize key information: The constructed digital twin panoramic model, generated operation and maintenance strategy, and other key data will be presented in the form of graphics, charts, and maps through visualization technology.

Images