Intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling

Abstract

The invention relates to an intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling, and belongs to the technical field of intelligent detection of power systems. The method comprises the following steps: dynamically splitting a power grid panoramic frame into an equipment physical entity layer and a static background environment layer, and generating a panoramic display picture through cloud and edge parallel rendering and a terminal HMD picture synthesis technology; establishing real-time bidirectional mapping between the physical entity and the virtual twinborn body of the power grid equipment, and synchronizing sensor data to the digital twinborn model; adopting an improved permutation entropy algorithm to extract equipment operation characteristic quantity, and dynamically distributing rendering pipeline resources in combination with a Bayesian network diagnosis model and an analytic hierarchy process evaluation system; and constructing an equipment state quality index, and quantitatively evaluating the health degree of the current equipment and predicting the future degradation trend by dynamically comparing the characteristic parameters with an expected value and combining health factor correction and a time sequence nonlinear prediction model. According to the invention, full-life-cycle efficient management and control of the operation state of the power grid equipment can be realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent detection of power systems, and relates to an intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling. Background Technique

[0002] Intelligent operation and maintenance of power grid equipment is a key technology to ensure the reliability of power systems. Traditional methods mainly rely on manual regular inspections, offline tests, or threshold alarm mechanisms based on single-source sensor data. For example, early warning is carried out through vibration spectrum thresholds, oil chromatographic analysis, or fixed-period mechanical deformation detection, and empirical threshold rules or offline simulation models are used to evaluate the health status of equipment. However, there are three core bottlenecks in the existing technical system: First, the real-time performance of model iteration is poor. Conventional digital twin systems adopt a full-data synchronization modeling strategy, and non-critical parameters occupy more than 70% of the edge computing resources, resulting in the update cycle of the twin model of key equipment being extended to 3-5 seconds, and the two-way synchronization delay between transient physical quantities and virtual models reaching 300-800 ms, making it difficult to accurately map gradual defects such as insulation aging and contact wear. Second, the reliability of cross-domain diagnosis is low. Existing methods mostly rely on the generation of isolated criteria for single-modal data and lack the joint analysis of the coupling characteristics of electromagnetic-thermal-mechanical multi-physical fields. Especially in the spatio-temporal matching of vibration signals and transient current waveforms, there is a time sequence misalignment of ≥20%, resulting in the multi-dimensional feature fusion error rate rising to 15%-22%, and the misjudgment rate of defect types exceeding the upper limit allowed by the operation and maintenance regulations. Third, the cloud-edge collaboration efficiency is weak. The cloud adopts a rigid resource allocation strategy, and the real-time rendering accuracy loss of high-value parameters exceeds 30%. The edge side is limited by the fixed-priority data scheduling mechanism, and the transmission packet loss rate of key status quantities soars to 25% during peak load periods, causing the dynamic evaluation deviation of the equipment health index to expand to 10%-15%.

[0003] With the development of digital twin technology and cloud-edge collaborative computing, existing technologies attempt to achieve state visualization by constructing virtual models of equipment and combining real-time sensor data. For example, a monitoring system based on cloud rendering sends three-dimensional models to the terminal for display, but it still has technical bottlenecks: the model update lags behind, only synchronizing data periodically and unable to achieve two-way dynamic mapping of microsecond-level transient features; the rendering efficiency is low, without differentiating between high-dynamic equipment entities and static backgrounds, resulting in idle edge computing resources; the problem of feature redundancy is prominent, using a single algorithm to extract feature parameters and ignoring the spatio-temporal correlation modeling of multi-modal data.

[0004] In recent years, deep learning technology has been introduced into the field of intelligent operation and maintenance of power grid equipment, such as using convolutional neural networks to fuse multi-sensor data to improve fault diagnosis accuracy. However, such methods have inherent limitations: insufficient content adaptability, reliance on fixed training data sets, resulting in poor generalization ability for sudden working conditions; rigid probability distribution models, assuming that the potential features obey a single Gaussian distribution, unable to describe complex non-Gaussian probability structures such as insulator discharge; potential representation redundancy, the geometric topology and behavior rule encoding of the digital twin model do not differentiate the allocation of bit rates, resulting in a waste of storage and computing resources.

[0005] Current improvement schemes attempt to improve system adaptability by updating model parameters online or introducing attention mechanisms. However, frequent retraining significantly increases edge computing overhead, and the probability modeling of potential representations is not optimized in combination with the physical characteristics of the equipment, resulting in a disconnect between model updates and the actual physical state. The above problems restrict the coordinated optimization of real-time, accuracy, and resource efficiency in the intelligent operation and maintenance of power grid equipment. An innovative method that integrates cloud-edge collaborative rendering architecture, dynamic probability distribution modeling, and fine-grained potential representation optimization is urgently needed to achieve efficient and accurate management and control of the entire life cycle of equipment. Summary of the invention

[0006] In view of this, the purpose of the present invention is to provide an intelligent operation and maintenance method for power transmission network equipment based on dynamic coupling of digital twins, aiming to achieve the collaborative optimization requirements of intelligent operation and maintenance of power grid equipment in terms of real-time, accuracy and resource efficiency.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] A method for intelligent operation and maintenance of power transmission network equipment based on dynamic coupling of digital twins, the method comprising the following steps:

[0009] Dynamically split the grid panoramic frame into the equipment physical entity layer and the static background environment layer, and generate a panoramic display screen through cloud-side and edge-side parallel rendering and terminal HMD screen synthesis technology;

[0010] Based on the digital twin model MDS, a real-time bidirectional mapping between the physical entity and the virtual twin of the power grid equipment is established, and sensor data is synchronized to the digital twin model;

[0011] The improved permutation entropy algorithm is used to extract the equipment operation characteristics, and the Bayesian network diagnosis model and hierarchical analysis evaluation system are combined to dynamically allocate rendering pipeline resources.

[0012] Construct an equipment status quality index, dynamically compare characteristic parameters with expected values, combine health factor correction and time series nonlinear prediction model, quantitatively evaluate the current equipment health and predict future degradation trends.

[0013] Further, the process of dynamically splitting the power grid panoramic frame and then generating a panoramic display image through cloud and edge-side parallel rendering and terminal HMD image synthesis technologies includes:

[0014] The power grid panoramic view is dynamically split for rendering by dividing it into a physical entity layer and a static background layer; among them, the device physical entity layer includes transformers, circuit breakers, and user interaction components; the static background layer includes static terrain and pole terminals.

[0015] The physical entity layer is rendered and modeled at the edge side using a hierarchical rendering method, and at the same time, the static background layer is rendered and modeled at the cloud side using a stable rendering method; among them, the hierarchical rendering method means that for the key areas of the physical entity layer, a rendering model with a triangle mesh number not lower than the high-precision threshold is used for modeling, and for the non-key areas of the physical entity layer, a rendering model with a triangle mesh number not exceeding the low-poly threshold is used for modeling; the stable rendering method means that a rendering model with a fixed rendering frame rate is used to model the static background. In this embodiment, offline baked lighting and pre-computed global illumination technologies are used.

[0016] The rendering results of the cloud and the edge side are fused through a screen synthesis formula with dynamically adjusted transparency weights. The screen synthesis formula is:

[0017] I final = α·I phy +(1 - α)·I env

[0018] In the formula, α is the transparency weight of the entity layer, which is dynamically adjusted according to the user interaction focus; I phy represents the data entity layer image rendered by the edge side, which contains dynamic content related to the user interaction focus, and I env represents the static background layer image rendered by the cloud, which contains environmental or non-interactive static content.

[0019] Further, the process of real-time bidirectional mapping between the power grid device physical entity and the virtual twin is as follows:

[0020] The real-time current data I(t) drives the update of the electromagnetic field of the virtual model to achieve the mapping from physical to virtual. The formula is:

[0021]

[0022] In the formula: μ0 is the magnetic permeability of vacuum, and r is the wire radius;

[0023] Based on the contact wear rate data predicted by the digital twin model, the system dynamically adjusts the adjustment step parameter of the tap changer actuator, that is, the gear change amount for a single action, to achieve the feedback from virtual to physical;

[0024] Based on the priority formula, the LOD level is dynamically adjusted by combining the current and the temperature change rate. The formula is as follows:

[0025]

[0026] In the formula: ω1 + ω2 = 1, which respectively represent the priority ratios of current change and temperature change; ΔI and ΔT are the change rates of current and temperature respectively; I max represents the maximum allowable current change rate threshold of the system (unit: A / s), and T max represents the maximum allowable temperature change rate threshold of the system (unit: °C / s); when the current change rate ΔI or the temperature change rate ΔT exceeds the threshold, the LOD level of the corresponding area is automatically increased to the highest precision.

[0027] Furthermore, in the process of improving the permutation entropy algorithm to extract the device operation characteristic quantities, a sliding window adaptive threshold τ is introduced to optimize the extraction accuracy of the non-linear dynamic characteristics of the device operation characteristic quantities. The expression is as follows:

[0028]

[0029] In the formula: m is the embedding dimension, and π i is the symbol sequence pattern, and p(π i ) represents the occurrence probability of the symbol pattern π i . The adaptive threshold τ is determined according to the maximum value of the precision max(H p ).

[0030] Furthermore, in the Bayesian network diagnosis model, based on the device historical fault data and real-time characteristic parameters, the fault type confidence level and risk level are output. Among them, the conditional probability table is constructed based on the historical fault library to output the fault type confidence level. The expression is as follows:

[0031]

[0032] In the formula: F k is the k-th type of fault, E is the real-time characteristic evidence, P(F k ) represents the prior probability, and P(E∣F k ) represents the conditional probability. The subscript "i" represents all fault types; if the confidence level of a certain type of fault exceeds the confidence level threshold, a warning signal is triggered.

[0033] Furthermore, in the analytic hierarchy process evaluation system, the multi-dimensional characteristic parameters are fused through the weight allocation algorithm to generate the global state quality index. Among them, the weight ζ i is calculated by combining multi-dimensional parameters, and the global quality index Z Lk (y) is generated:

[0034]

[0035] Where: ζ i is the combined weight, and S i (·) is the feature score of the i-th item; the multi-dimensional parameter is the vibration feature including the environmental parameter and the device status parameter, where the environmental parameter includes temperature and humidity, and the device status parameter includes energy consumption and running duration.

[0036] Furthermore, the dynamic resource allocation process of the rendering pipeline includes:

[0037] Scheduling the rendering tasks of the cloud and edge ends based on the user's bid priority; optimizing the wireless channel transmission delay by using the TDMA (Time Division Multiple Access) technology. Among them, the formula for optimizing the transmission delay at the edge end by the TDMA time slot allocation strategy is:

[0038]

[0039] Where: L e is the edge node load, N node is the total number of nodes, and T frame is the total duration of the TDMA frame period;

[0040] Construct a mechanism for scheduling cloud rendering tasks, and give a pricing model to schedule tasks based on the economic weight ρ and the QoS score preset by the user. The formula is:

[0041] Bid = ρ·P + (1 - ρ)·QoS, ρ ∈ [0, 1]

[0042] Where: ρ ∈ [0, 1] is the economic weight, and Q os is the service quality score.

[0043] Furthermore, obtain the power grid equipment status quality index by combining the weight values of the power grid equipment characteristic parameters The formula is:

[0044]

[0045] Where: Z lk (·) represents the status quality index when the power grid equipment is in a normal state, y is the number of iterations, and ζ k is the combined weight of the k-th characteristic parameter.

[0046] Furthermore, the formula for dynamically adjusting the status quality index Z by combining the health factor correction is: a L(y) The formula is:

[0047]

[0048] where: v(·) is the health factor. When determining the health factor based on the device operation duration t and the material degradation rate γ, the corrected quality index is expressed as:

[0049]

[0050] Among them, the more severely aged the device is, the greater the amplitude of the exponential decay.

[0051] Furthermore, in the time series non - linear prediction model, a polynomial fitting and LSTM network fusion algorithm is adopted to output the future state quality index The formula is:

[0052]

[0053] In the formula: The periodic trend is captured by polynomial terms, ζ0 is the fitting coefficient, n is the polynomial degree, and LSTm(·) is to learn non - linear mutation characteristics.

[0054] The beneficial effects of the present invention are as follows:

[0055] First of all, through the dynamic hierarchical rendering strategy, the present invention efficiently splits the power grid panoramic frame into the device physical entity layer and the static background environment layer. By using the parallel rendering technology of the cloud and the edge side, combined with the terminal HMD picture synthesis, the real - time and high - fidelity interaction between the device physical parameters and the virtual model is realized. This strategy not only significantly improves the timeliness of power grid device status monitoring, but also ensures the accuracy of monitoring data, laying a solid foundation for the stable and efficient operation of the power system.

[0056] Secondly, in terms of data processing and fault prediction, the present invention introduces an improved permutation entropy algorithm, a Bayesian network diagnosis model and an analytic hierarchy process evaluation system. Under the joint action of these advanced algorithms and models, the operation characteristic quantities of the device can be accurately extracted, the confidence of the fault type can be accurately inferred, and the status of power grid devices can be comprehensively and accurately monitored. At the same time, by dynamically allocating rendering pipeline resources, such as LOD dynamic optimization and TDMA transmission strategy, the real - time performance and rendering efficiency of the system are further improved, ensuring the timeliness and effectiveness of monitoring data.

[0057] In addition, by constructing the device state quality index, dynamically comparing the characteristic parameters with the expected values, combining the health factor correction and the time series non - linear prediction model, the present invention can accurately evaluate the current health state of the device and predict its future degradation trend. This function provides strong support for the preventive maintenance and full - life - cycle management of power grid devices, helps to reduce the device failure rate, extend the service life of the device, and improve the overall efficiency and safety of power grid operation.

[0058] Other advantages, objects, and features of the present invention will be set forth to some extent in the following description, and to some extent, will be apparent to those skilled in the art based on an examination of the following, or can be learned from the practice of the present invention. The objects and other advantages of the present invention can be achieved and obtained through the following description. Brief Description of the Drawings

[0059] In order to make the objects, technical solutions, and advantages of the present invention clearer, the present invention will be described in detail preferably with reference to the accompanying drawings, where:

[0060] Figure 1 is a schematic flowchart of an intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to an embodiment of the present invention;

[0061] Figure 2 is a schematic diagram of the dynamic hierarchical rendering strategy process according to an embodiment of the present invention;

[0062] Figure 3 is a schematic diagram of the two-way mapping and LOD dynamic adjustment process of the digital twin model according to an embodiment of the present invention;

[0063] Figure 4 is a schematic diagram of the feature extraction and resource dynamic allocation process according to an embodiment of the present invention;

[0064] Figure 5 is a schematic diagram of the health assessment and degradation prediction process according to an embodiment of the present invention. Detailed Embodiments

[0065] The following illustrates the embodiments of the present invention through specific specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the drawings provided in the following embodiments only illustrate the basic concept of the present invention schematically. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0066] Among them, the drawings are only for illustrative purposes, showing only schematic diagrams, not physical diagrams, and should not be construed as a limitation to the present invention; in order to better illustrate the embodiments of the present invention, some components in the drawings will be omitted, enlarged, or reduced, and do not represent the dimensions of actual products; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0067] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "front", "rear", etc. indicating the orientation or positional relationship, they are based on the orientation or positional relationship shown in the accompanying drawings. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the accompanying drawings are only for illustrative purposes and should not be construed as a limitation of the present invention. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances.

[0068] Please refer to Figures 1 to 5 , which is an intelligent operation and maintenance method for power grid equipment based on digital twin dynamic coupling.

[0069] Embodiment

[0070] Figure 1 shows the overall flow schematic diagram of the intelligent operation and maintenance method for power grid equipment based on digital twin dynamic coupling in this embodiment. Among them, the method includes the following steps:

[0071] S1. Dynamically split the power grid panoramic frame into the device physical entity layer and the static background environment layer, and generate a panoramic display screen through cloud, edge-side parallel rendering and terminal HMD screen synthesis technology;

[0072] S2. Establish a real-time two-way mapping between the power grid device physical entity and the virtual twin based on the digital twin model MDS, and synchronize the sensor data to the twin model;

[0073] S3. Adopt an improved permutation entropy algorithm to extract the device operation characteristic quantities, and combine the Bayesian network diagnosis model and the analytic hierarchy process evaluation system to dynamically allocate rendering pipeline resources (LOD dynamic optimization, TDMA transmission);

[0074] S4. Construct a device state quality index, and by dynamically comparing the characteristic parameters with the expected values, combined with health factor correction and time series non-linear prediction model, quantitatively evaluate the current device health and predict the future degradation trend.

[0075] In step S1 of this embodiment, when dynamically splitting the power grid panoramic frame, the device physical entity layer includes transformers, circuit breakers, and user interaction components, and the static background environment layer includes static terrain and pole terminals. The physical entity layer rendering adopts a multi-level LOD (Level of Detail) optimization strategy, allocating high-precision models (triangle mesh count ≥ 500,000) to key devices (such as circuit breaker contacts), and reducing non-critical areas to low-poly models (triangle mesh count ≤ 50,000). The static background layer uses offline baked lighting and precomputed global illumination (Precomputed Radiance Transfer, PRT) technology, and the rendering frame rate is stable at 60 FPS.

[0076] For specific implementation details, please refer to the appendix Figure 2 , including:

[0077] Adopt split rendering to dynamically split the power grid panorama into a physical entity layer and a static background layer; among them, the device physical entity layer includes transformers, circuit breakers, and user interaction components; the static background layer includes static terrain and pole terminals;

[0078] Adopt a hierarchical rendering method to render and model the physical entity layer, and at the same time adopt a stable rendering method to render and model the static background layer; among them, the hierarchical rendering method means using a rendering model with a triangle mesh count not lower than the high-precision threshold for key areas of the physical entity layer for modeling, and using a rendering model with a triangle mesh count not exceeding the low-poly threshold for non-critical areas of the physical entity layer for modeling; the stable rendering method means using a rendering model with a fixed rendering frame rate to model the static background. In this embodiment, offline baked lighting and precomputed global illumination technology are adopted;

[0079] Through a screen synthesis formula with dynamic adjustment of transparency weights, fuse the rendering results of the cloud and the edge (where cloud rendering corresponds to the static background layer and edge segment rendering corresponds to the data entity layer). The screen synthesis formula is:

[0080] I final = α·I phy +(1 - α)·I env

[0081] In the formula, α ∈ [0.7, 0.9] is the transparency weight of the entity layer, dynamically adjusted by the user interaction focus; I phy represents the data entity layer screen generated by edge-side rendering, including dynamic content related to the user interaction focus, and I env represents the static background layer screen generated by cloud rendering, including environmental or non-interactive static content.

[0082] In step S2 of this embodiment, the virtual-real bidirectional mapping mechanism of the digital twin model MDS includes: real-time synchronization and feedback between physical entity parameters (dimensions, current / voltage data) and virtual models (geometric structures, behavior rules); dynamically adjusting the detail level of the twin model (LOD) based on the rendering task priority.

[0083] Based on the digital twin model MDS, a real-time bidirectional mapping of physical-virtual parameters is established to achieve data synchronization and precision optimization between the physical entity and the virtual model. For specific implementation details, please refer to the appendix Figure 3 , which is a schematic diagram of the bidirectional mapping and LOD dynamic adjustment of the digital twin model in the embodiment of the present invention. Establishing a bidirectional mapping mechanism from physical to virtual specifically includes the following steps:

[0084] The real-time current data I(t) drives the update of the electromagnetic field of the virtual model to achieve the mapping from physical to virtual. The formula is:

[0085]

[0086] In the formula: μ0 is the magnetic permeability of vacuum, and r is the radius of the wire.

[0087] Based on the contact wear rate data predicted by the digital twin model, the system dynamically adjusts the adjustment step parameter of the tap changer actuator (i.e., the gear change amount for a single action) to achieve the feedback from virtual to physical.

[0088] Based on the priority formula, the LOD level is dynamically adjusted in combination with the current and temperature change rates. The formula is:

[0089]

[0090] In the formula: ω1 + ω2 = 1, which respectively represent the priority proportions of current change and temperature change; ΔI and ΔT are the change rates of current and temperature respectively; I max represents the maximum allowable current change rate threshold of the system (unit: A / s), and T max represents the maximum allowable temperature change rate threshold of the system (unit: °C / s). When the current change rate ΔI or the temperature change rate ΔT exceeds the threshold, the LOD level of the corresponding area is automatically increased to the highest precision.

[0091] In step S3 of this embodiment, for specific implementation details, please refer to the appendix Figure 4 , which is a schematic diagram of feature extraction and resource dynamic allocation in the embodiment of the present invention. An improved permutation entropy algorithm and multi-model fusion are used to optimize the fault prediction. Specifically, it includes the following steps:

[0092] Improve the permutation entropy algorithm, introduce a sliding window adaptive threshold τ, and optimize the extraction accuracy of the non-linear dynamic characteristics of the device operation feature quantity. Its expression is:

[0093]

[0094] where: m is the embedding dimension (the value range is 3 - 7), π i is the i-th symbol sequence pattern, p(π i ) represents the occurrence probability of the symbol pattern π i . The symbol pattern set {π i} is dynamically updated through a sliding window. When the standard deviation σ of H p (m) > 0.2, the window length L = 2m is automatically adjusted. The influence of the time delay τ is implicitly included in the formula (indirectly reflected by the sliding window length L = mτ).

[0095] The Bayesian network diagnosis model outputs the confidence level of the fault type and the risk level based on the historical fault data and real-time characteristic parameters of the device. In this embodiment, a Bayesian network is used to construct a conditional probability table based on the historical fault library and output the confidence level of the fault type. Its expression is:

[0096]

[0097] where: F k is the k-th type of fault, E is the real-time characteristic evidence, P(F k ) represents the prior probability, P(E∣F k ) represents the conditional probability, and the subscript "i" represents all fault types. If the confidence level of a certain type of fault exceeds the confidence level threshold, a warning signal is triggered. In this embodiment, the confidence level threshold is set to 0.8.

[0098] The analytic hierarchy process evaluation system fuses multi-dimensional characteristic parameters through the weight allocation algorithm (AHP) to generate a global state quality index. The weight ζ i is calculated by the analytic hierarchy process. By integrating multi-dimensional parameters such as the comprehensive environment (temperature, humidity, etc.), device status (energy consumption, operation duration, etc.), and vibration characteristics, a global quality index is generated:

[0099]

[0100] where: ζ i is the combined weight, and S i (·) is the i-th feature score.

[0101] The dynamic allocation of rendering pipeline resources includes: scheduling cloud and edge rendering tasks based on the user bid priority; optimizing the wireless channel transmission delay using TDMA (Time Division Multiple Access) technology. The TDMA time slot allocation strategy is adopted to optimize the edge transmission delay. Its formula is:

[0102]

[0103] where: L e is the load of the edge node, N node is the total number of nodes, T frame is the total duration of the TDMA frame period.

[0104] Build a mechanism for cloud rendering task scheduling. Based on the economic weight ρ (preset by the user) and the scoring QoS, a pricing model is given to schedule tasks. The formula is:

[0105] Bid = ρ·P + (1 - ρ)·QoS, ρ ∈ [0, 1]

[0106] where: ρ ∈ [0, 1] is the economic weight, Q os is the service quality score.

[0107] In step S4 of this embodiment, the power grid equipment status quality index is obtained by combining the weight values of the power grid equipment characteristic parameters The formula is:

[0108]

[0109] where: represents the status quality index when the power grid equipment is in a normal state, y is the number of iterations, ζ k is the combined weight of the kth characteristic parameter.

[0110] In step S4 of this embodiment, a time series prediction model that constructs the equipment status quality index and integrates polynomial fitting and LSTM is implemented to achieve quantitative health assessment and accurate prediction of degradation trends. For the specific implementation details, please refer to the appendix Figure 5 , which is a schematic diagram of the health assessment and degradation prediction model of this embodiment of the present invention. Constructing the equipment status quality index specifically includes:

[0111] Combining the health factor correction to dynamically adjust the status quality index Z a L(y) The formula is:

[0112]

[0113] where: v() is the health factor.

[0114] In this embodiment, according to the equipment operation duration t and the material degradation rate γ, the quality index is corrected The formula is:

[0115]

[0116] Among them, the exponential decay amplitude of the equipment with serious aging (such as t > 10 years) increases.

[0117] The time series non - linear prediction model adopts the polynomial fitting and LSTM network fusion algorithm to output the future state quality index The formula is:

[0118]

[0119] Where: ζ0 is the fitting coefficient, and n is the polynomial degree.

[0120] In this embodiment, the polynomial fitting and LSTM fusion algorithm is adopted to predict the state index in the next 30 days Its formula is:

[0121]

[0122] Where: The polynomial term captures the periodic trend, n is the polynomial degree, and LSTM(·) is to learn the non - linear mutation characteristics.

[0123] First of all, the present invention realizes the real - time interaction and high - fidelity monitoring of the physical parameters of the equipment and the virtual model by dynamically splitting the grid panoramic frame into the equipment physical entity and the static background environment, and adopting the cloud - edge parallel rendering and terminal HMD picture synthesis technology. This innovative strategy not only improves the timeliness of the grid equipment operation state monitoring, but also ensures the accuracy of the monitoring results, providing a strong guarantee for the stable operation of the power system.

[0124] Secondly, the present invention introduces the improved permutation entropy algorithm, the Bayesian network diagnosis model and the analytic hierarchy process evaluation system. By accurately extracting the operation characteristic quantities of the equipment and inferring the confidence level of the fault types, it realizes the comprehensive and accurate monitoring of the grid equipment state. At the same time, by dynamically allocating the rendering pipeline resources, the rendering efficiency is optimized, and the real - time performance of the system is further improved.

[0125] In addition, the present invention also constructs the equipment state quality index. By dynamically comparing the characteristic parameters with the expected values, combined with the health factor correction and the time series non - linear prediction model, it realizes the quantitative evaluation of the current health degree of the equipment and the accurate prediction of the future degradation trend. This function provides a scientific basis for the preventive maintenance and full - life - cycle management of grid equipment, helps to reduce the equipment failure rate, and improves the safety and reliability of the grid operation.

[0126] In summary, the intelligent operation and maintenance method of the transmission grid equipment based on digital twin coupling drive proposed by the present invention has achieved remarkable technical effects in improving the timeliness and accuracy of grid equipment monitoring, optimizing the rendering efficiency, realizing the quantitative evaluation of equipment health degree and predicting the degradation trend, providing a new solution for the intelligent operation and maintenance management of the power system.

[0127] It should be noted that the dynamic parameters in the above embodiments can be adjusted according to the specific device type. Those skilled in the art can make equivalent replacements or extensions to the algorithm parameters, model structures, and interaction logics without departing from the core idea of the present invention. For example, replacing LSTM with a GRU network, adjusting the weight allocation strategy of the Gaussian mixture model, etc. Such modifications should all be regarded as within the protection scope of the present invention.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the purpose and scope of the present technical solution, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling, characterized in that: The method comprises the following steps: Dynamically split the grid panoramic frame into the equipment physical entity layer and the static background environment layer, and generate a panoramic display screen through cloud-side and edge-side parallel rendering and terminal HMD screen synthesis technology; Based on the digital twin model MDS, a real-time bidirectional mapping between the physical entity and the virtual twin of the power grid equipment is established, and sensor data is synchronized to the digital twin model; The improved permutation entropy algorithm is used to extract the equipment operation characteristics, and the Bayesian network diagnosis model and hierarchical analysis evaluation system are combined to dynamically allocate rendering pipeline resources. Construct an equipment status quality index, dynamically compare characteristic parameters with expected values, combine health factor correction and time series nonlinear prediction model, quantitatively evaluate the current equipment health and predict future degradation trends.

2. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 1, wherein: The process of dynamically splitting the grid panoramic frame and then generating a panoramic display screen through cloud-side and edge-side parallel rendering and terminal HMD screen synthesis technology includes: The grid panorama is dynamically split into a physical entity layer and a static background layer by using split rendering. The physical entity layer of the equipment includes transformers, circuit breakers and user interaction components; the static background layer includes static terrain and pole terminals. A hierarchical rendering method is used to render and model the physical entity layer at the edge, and a stable rendering method is used to render and model the static background layer at the cloud; wherein the hierarchical rendering method refers to modeling the key areas of the physical entity layer using a rendering model with a number of triangles not less than a high-precision threshold, and modeling the non-critical areas of the physical entity layer using a rendering model with a number of triangles not exceeding a low-poly threshold; the stable rendering method refers to modeling the static background using a rendering model with a fixed rendering frame rate. In this embodiment, offline baking lighting and pre-calculated global illumination technology are used; The picture synthesis formula is adjusted dynamically through the transparency weight, integrating the cloud and edge rendering results. The picture synthesis formula is: I final = α·I phy +(1 - α)·I env where α is the transparency weight of the entity layer, which is dynamically adjusted according to the user interaction focus; I phy represents the data entity layer image generated by edge - side rendering, including dynamic content related to the user interaction focus, I env represents the static background layer image generated by cloud - side rendering, including environmental or non - interactive static content.

3. An intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 1, characterized in that: The process of real-time bidirectional mapping between the physical entity and the virtual twin of the power grid equipment is as follows: Real-time current data I(t) drives the virtual model electromagnetic field update to achieve mapping from physical to virtual. The formula is: Where: μ0 is the vacuum magnetic permeability, R is the wire radius; Based on the contact wear rate data predicted by the digital twin model, the system dynamically adjusts the adjustment step parameters of the tap changer actuator, that is, the gear position change amount of a single action, to achieve virtual to physical feedback; Based on the priority formula, the LOD level is dynamically adjusted in combination with the current and temperature change rate. The formula is: Where: ω1 + ω2 = 1, which respectively represent the priority ratios of current change and temperature change; ΔI and ΔT are the change rates of current and temperature respectively; I max represents the maximum allowable current change rate threshold of the system (unit: A / s), and T max represents the maximum allowable temperature change rate threshold of the system (unit: °C / s); when the current change rate ΔI or the temperature change rate ΔT exceeds the threshold, the LOD level of the corresponding area is automatically increased to the highest precision.

4. An intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 1, characterized in that: In the process of extracting equipment operation characteristics using the improved permutation entropy algorithm, a sliding window adaptive threshold τ is introduced to optimize the accuracy of extracting the nonlinear dynamic characteristics of equipment operation characteristics. The expression is: where: m is the embedding dimension, π i is the symbol sequence pattern, p(π i ) represents the occurrence probability of the symbol pattern π i , and the adaptive threshold τ is determined according to the maximum precision max(H p ).

5. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 4, characterized in that: In the Bayesian network diagnosis model, the fault type confidence and risk level are output based on the historical fault data and real-time characteristic parameters of the equipment. The conditional probability table is constructed based on the historical fault library, and the fault type confidence is output. The expression is: Where: F k is the k-th type of fault, E is the real-time feature evidence, P(F k ) represents the prior probability, P(E∣F k ) represents the conditional probability, and the subscript "i" represents all fault types; if the confidence level of a certain type of fault exceeds the confidence level threshold, a warning signal is triggered.

6. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 5, wherein: In the hierarchical analysis and evaluation system, multi-dimensional feature parameters are fused through a weight allocation algorithm to generate a global state quality index, where the weight ζ is calculated by combining multi-dimensional parameters i , and a global quality index is generated Where: ζ i is the combined weight, S i (·) is the feature score of the i-th item; the multi-dimensional parameter is the vibration feature including environmental parameters and equipment status parameters, where the environmental parameters include temperature and humidity, and the equipment status parameters include energy consumption and operation duration.

7. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 6, wherein: The rendering pipeline resource dynamic allocation process includes: Schedule cloud and edge rendering tasks based on user bid priorities; optimize the wireless channel transmission delay using TDMA (Time Division Multiple Access) technology. Among them, the formula for optimizing the edge transmission delay with the TDMA time slot allocation strategy is: Where: L e is the load of the edge node, N node is the total number of nodes, T frame is the total duration of the TDMA frame period; Build a mechanism for scheduling cloud rendering tasks. Based on the economic weight ρ and the scoring QoS preset by the user, give a bid model to schedule tasks, and its formula is: Bid=ρ·P+(1-ρ)·QoS, ρ∈[0,1] where: ρ ∈ [0, 1] is the economic weight, and Q os is the service quality score.

8. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 1, wherein: Obtaining the Grid Equipment Status Quality Index by Combining the Weight Values of Grid Equipment Characteristic Parameters The formula is as follows: In the formula: represents the state quality index when the power grid equipment is in a normal state, y is the number of iterations, and ζ k is the combined weight of the k-th characteristic parameter.

9. The intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 8, wherein: Modify the dynamic adjustment status quality index Z by combining health factors a L(y) The formula is as follows: Where: v(·) is the health factor. When determining the health factor based on the equipment operation duration t and the material degradation rate γ, the modified quality index is expressed as: Among them, the more severely aged the device is, the greater the exponential decay amplitude.

10. An intelligent operation and maintenance method for power transmission network equipment based on digital twin dynamic coupling according to claim 9, characterized in that: In the time series non-linear prediction model, a polynomial fitting and LSTM network fusion algorithm is adopted to output the future state quality index The formula is as follows: In the formula: capture the periodic trend through polynomial terms, ζ0 is the fitting coefficient, n is the polynomial degree, and LSTM(·) is to learn non-linear mutation characteristics.

Images