Digital twinborn system for monitoring state of electric power tower in real time and identifying common faults

Abstract

The invention provides an electric power tower state real-time monitoring and common fault identification system based on a digital twinborn technology. According to the system, a high-precision digital twinborn model module is constructed, and a sensor network module is integrated so as to acquire key data of a tower in real time. The data processing center module utilizes an advanced clock synchronization technology to ensure data precision, cooperates with the common fault identification system module and the health state evaluation module, and adopts finite element analysis and Paris laws to accurately evaluate the health condition of the tower. The system further comprises an intelligent maintenance decision support module which provides a scientific maintenance strategy. In addition, the data visualization platform module and the user interaction interface module enhance the user experience, so that the system is not limited to monitoring and evaluation, and optimization of maintenance decision and individuation of user interaction are also realized. The overall system design pays attention to real-time performance, accuracy and intelligence, and aims at improving the operation and maintenance efficiency and safety of the electric power tower.

Description

Technical Field

[0001] The present invention relates to the technical field of power system monitoring, and in particular to a power tower state real-time monitoring and common fault identification system based on digital twin technology. Background Art

[0002] As an important part of the transmission line, the stability of power towers is directly related to the reliability and safety of the power grid. Traditional power tower monitoring methods mainly rely on manual inspections, which have many limitations, such as long monitoring cycles, inability to respond in real time, high labor intensity, and low efficiency. With the development of sensor technology and information technology, some automated monitoring solutions have emerged, but these solutions often have problems such as low monitoring accuracy, limited fault identification capabilities, and insufficient intelligence in maintenance decisions. Summary of the invention

[0003] In order to solve the above problems, a real-time monitoring and common fault identification system for power towers based on digital twin technology has been developed. The system has the following features: A high-precision digital twin model is created for each power tower, which can accurately reflect the physical characteristics, environmental conditions and operating status of the tower. The system can receive and integrate data from tower sensors in real time, including but not limited to stress, displacement, vibration, ambient temperature and humidity, etc. Develop a set of algorithms to analyze real-time data and automatically identify common problems of towers, such as corrosion, cracks, and unstable foundations. Based on the results of common fault identification, the system can evaluate the health of the tower and predict potential failure risks. The system provides intelligent maintenance decision support, including maintenance time, required materials, personnel arrangements, etc.

[0004] To achieve the above objectives, this application adopts the following technical solutions:

[0005] The digital twin system of power towers includes:

[0006] -High-precision digital twin model module, which is used to receive the design parameters, material properties, construction records and historical maintenance data of the tower, and build a digital twin model containing geometric models and physical properties;

[0007] -Sensor network module, which is used to deploy sensors according to the location information of key monitoring points in the digital twin model, collect the force, displacement and vibration data of the tower in real time, and generate sensor data sets;

[0008] -Data processing center module, used to receive sensor data sets, perform data synchronization, preprocessing and anomaly detection, and output preprocessed data sets;

[0009] - Common fault identification system module, used to receive preprocessed data sets and digital twin models, perform feature extraction, model training and fault identification, and output fault status sets;

[0010] -Health status assessment module, which is used to receive the fault status set and digital twin model, perform fault severity analysis and risk prediction, and output the health index;

[0011] -Maintenance decision support module, which is used to receive health index and digital twin model, determine maintenance priority and optimize resource allocation, and generate maintenance plan.

[0012] The high-precision digital twin model modules include:

[0013] -Environmental simulation unit, used to simulate the mechanism of the impact of environmental factors such as wind, temperature, humidity, etc. on the tower;

[0014] -Real-time update unit, used to synchronously update the digital twin model through wireless or wired communication to reflect the real-time status of the tower.

[0015] The sensor network module includes:

[0016] -An adaptive algorithm unit for adjusting sensor positions based on real-time feedback to enhance the accuracy of the sensor dataset;

[0017] -Data synchronization algorithm unit to handle communication delays or data loss and ensure the spatiotemporal consistency of sensor data sets.

[0018] The data processing center modules include:

[0019] -Data preprocessing unit for processing sensor data sets using low-pass filter, wavelet transform, Z-score or Min-Max normalization method;

[0020] -Outlier detection unit, used to automatically identify and remove abnormal data points in the preprocessed data set based on the isolation forest or One-Class SVM algorithm.

[0021] The common disease identification system modules include:

[0022] -Feature extraction unit, used to extract key features from the preprocessed data set based on PCA and ICA techniques;

[0023] -Model training unit, used to perform fault pattern recognition using SVM, random forest, GBM and deep neural network algorithms to generate fault state sets.

[0024] The health status assessment module includes:

[0025] -Failure severity analysis unit, used to quantify the stress increase caused by concentrated cracks or notches in the failure state using stress concentration factors;

[0026] -Risk prediction unit, used to predict the probability of fault state transition using Markov chain analysis and determine the remaining service life of the tower.

[0027] Maintenance decision support modules include:

[0028] -Maintenance priority determination unit based on risk matrix, which is used to determine maintenance priority by combining health index, tower importance and maintenance resource availability;

[0029] -Time-cost optimization unit, used to balance the time and cost of completing tasks in the maintenance plan.

[0030] The system also includes a data visualization platform module, which includes:

[0031] -Interactive 3D visualization unit, which allows users to observe the digital twin model from different angles and scales;

[0032] - Custom data report generation unit for generating reports based on fault status sets, health indices and maintenance plans.

[0033] The system also includes a user interaction interface module, which includes:

[0034] -Personalized recommendation unit, used to recommend maintenance strategies in the maintenance plan based on the user's historical selections and preferences;

[0035] -Maintenance log tracking unit to record maintenance activities and results and update the digital twin model.

[0036] The system also includes:

[0037] -High-performance computing cluster unit, used to process the pre-processed data set using MapReduce parallel processing technology;

[0038] -Real-time stream processing unit for processing sensor data streams of sensor datasets.

[0039] Through the above technical solution, a real-time monitoring and common fault identification system for power tower status based on digital twin technology is provided. The system uses high-precision digital twin models, real-time data integration, common fault identification algorithms, health status assessment and maintenance decision support technology to achieve accurate monitoring and maintenance of power tower status. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to make the embodiments, technical solutions and technical advantages of the present application clearer, the following drawings provide non-limiting views of the present application:

[0041] Figure 1 The overall architecture diagram of the power tower digital twin system is shown.

[0042] Figure 2 The flowchart for building a high-precision digital twin model is shown.

[0043] Figure 3 A diagram showing the working principle of the environmental simulation unit.

[0044] Figure 4 A schematic diagram of the process of real-time data acquisition and synchronization is shown.

[0045] Figure 5 The data preprocessing and outlier detection flow chart of the data processing center is shown.

[0046] Figure 6 The workflow diagram of the common disease identification algorithm is shown.

[0047] Figure 7 The assessment flow chart of the health status assessment module is presented.

[0048] Figure 8 A decision flow chart of the maintenance decision support module is presented. DETAILED DESCRIPTION

[0049] In order to elaborate on the embodiments, technical solutions and technical advantages of the present application, a detailed and complete description will be provided below in conjunction with the accompanying drawings. It should be clear that the described embodiments only represent some application scenarios of the present application, not all possibilities. According to the embodiments of the present disclosure, those skilled in the art can obviously obtain other possible embodiments without performing creative work, and these embodiments also belong to the scope of protection of the present application.

[0050] 1. Construction of high-precision digital twin model

[0051] The construction of a high-precision digital twin model is a key step in achieving the innovative goal of real-time monitoring of power tower status and common fault identification system. This section describes in detail the construction process, technical details, and integration and embodiment of the digital twin model in actual application scenarios.

[0052] The first step in building a high-precision digital twin model is to comprehensively collect the design parameters, material properties, construction records, and historical maintenance data of the power tower. These data provide basic information for the model, including but not limited to the size, material grade, stress characteristics, and historical maintenance and failure records of the tower. After data collection, preprocessing is performed to ensure data consistency and accuracy, including data cleaning, format unification, and outlier removal.

[0053] Professional 3D modeling software is used to build the geometric model of the tower based on the collected data. This process requires detailed consideration of each component of the tower, such as the base, pole, crossarm, insulator, conductor, etc., as well as the connection relationship between them. Using software such as SolidWorks or Autodesk 3ds Max, a 1:1 scale 3D model is created to ensure that the geometric size and relative position of each component are consistent with the physical tower.

[0054] Based on the geometric model, each component is assigned physical properties, such as density, elastic modulus, Poisson's ratio, yield strength, etc. These properties are based on material science and engineering principles and can be obtained through material mechanics testing or set according to material standards and specifications. For example, for a concrete tower, its elastic modulus E and Poisson's ratio v are determined according to the material grade and used in subsequent force analysis.

[0055] The environmental simulation unit simulates the influences on the tower in the actual environment, such as wind, temperature, humidity, etc. Through environmental simulation software or algorithms, these environmental factors are incorporated into the model to reflect the stress state and performance changes of the tower under different environmental conditions. For example, temperature changes may affect material properties, while wind forces directly act on the tower to produce bending moments and vibrations.

[0056] The wind load can be calculated using the following formula:

[0057] F w =0.5·ρ·A·C d ·V 2

[0058] Among them, Fw is the wind load, ρ is the air density, A is the frontal area, Cd is the drag coefficient, and V is the wind speed.

[0059] The real-time update unit is a key component of the digital twin model. By deploying high-precision sensors such as strain gauges, displacement sensors, and vibration sensors on the tower, the force, displacement, and vibration data of the tower are collected in real time. These data are transmitted to the data processing center via wireless or wired communication and updated synchronously with the digital twin model to ensure that the model reflects the real-time status of the tower.

[0060] To ensure the accuracy of the digital twin model, model validation and calibration are required. By comparing the model prediction with the actual measurement data, the accuracy of the model can be evaluated and necessary adjustments can be made. For example, if the stress distribution predicted by the model deviates from the actual measurement, calibration can be performed by adjusting the material properties or boundary conditions.

[0061] To further enhance the accuracy of the digital twin model, the present invention integrates finite element analysis (FEA). By discretizing the three-dimensional model of the tower into a finite number of elements and nodes, the stress, strain and displacement of the tower under various loads can be analyzed in detail. The general formula of the finite element model is as follows:

[0062] K·U=F

[0063] Where K is the stiffness matrix, U is the node displacement vector, and F is the node load vector. The stiffness matrix of each element can be calculated from the elastic modulus E of the material and the geometric properties.

[0064] The digital twin model simulates the various physical field effects on power towers in actual operation through multi-physics field coupling analysis, such as thermal expansion, electrical corrosion, etc. For example, thermal expansion can be expressed as:

[0065] ΔL=α·L0·ΔT

[0066] Where ΔL is the length change, α is the thermal expansion coefficient, L0 is the initial length, and ΔT is the temperature change.

[0067] Through the above steps, we built a high-precision, dynamically updated digital twin model, laying the foundation for subsequent status monitoring and common problem identification.

[0068] 2. Real-time Data Integration

[0069] The real-time data integration system of the present invention is realized based on the sensor network module and the data processing center module. The two modules work together to ensure the efficiency and accuracy of the whole process from data collection to processing.

[0070] The sensor network module is the core of real-time data collection and is responsible for generating sensor data sets. Its specific implementation is as follows:

[0071] Based on the location information of key monitoring points in the digital twin model, the system uses an optimization algorithm to determine the best layout of sensors. For example, strain gauges are preferentially deployed in the area with the expected maximum bending moment, and vibration sensors are installed on the top of the tower that is susceptible to wind vibration. Through genetic algorithms or particle swarm optimization algorithms, we can find the most cost-effective sensor deployment plan.

[0072] In order to cope with the specific environment and structural characteristics of the tower, the adaptive algorithm unit adjusts the sensor position based on real-time feedback to enhance the accuracy of the sensor data set. For example, if the stress data in a certain area is abnormal, the system automatically adjusts the parameters of the adjacent sensors or adds sensors to enhance the data accuracy in that area.

[0073] To ensure the spatiotemporal consistency of sensor datasets, the data synchronization algorithm uses high-precision clock synchronization technology, such as the Network Time Protocol (NTP), to ensure that all sensor data are collected under a unified time reference. In addition, we also developed a time window-based data synchronization algorithm to handle possible communication delays or data loss.

[0074] The improved PTP (Precision Time Protocol) is used to ensure nanosecond-level time synchronization accuracy. In addition, a spatial calibration algorithm is introduced to ensure that the spatial position errors of different sensors are within an acceptable range. The formula is as follows:

[0075]

[0076] Among them, Δx, Δy, and Δz represent the positioning errors of the sensor on the three coordinate axes respectively.

[0077] Considering that pole towers are usually located in outdoor environments where wiring is difficult, the system uses wireless communication technology to transmit sensor data. According to the data volume, transmission distance and energy consumption requirements, a suitable communication method is selected, such as cellular network, Wi-Fi, LoRa, etc. The data is compressed and encrypted before transmission to reduce bandwidth requirements and ensure data security. The system also develops an adaptive transmission power control algorithm based on channel quality to maximize transmission efficiency and reduce energy consumption. The transmission power PP is dynamically adjusted according to the channel state information (CSI), and the formula is as follows:

[0078] P = min(P max ,k·CSI)

[0079] Among them, P max is the maximum transmission power, and k is a proportional constant determined by energy consumption and coverage range.

[0080] The data processing center module is responsible for receiving sensor data sets, performing data synchronization, preprocessing and anomaly detection, and outputting preprocessed data sets. Its specific implementation is as follows:

[0081] The data preprocessing unit uses a series of algorithms to improve data quality, including: using low-pass filters to remove high-frequency noise, using wavelet transforms for signal denoising, and using Z-score or Min-Max normalization methods to standardize data. These algorithms can automatically adjust parameters to adapt to different data characteristics and environmental changes. For example, the Z-score standardization formula is as follows:

[0082] z=(x-μ) / σ

[0083] Here, x is the original data point, μ is the mean of the data set, and σ is the standard deviation.

[0084] The outlier detection unit automatically identifies and removes abnormal data points in the preprocessed data set based on methods such as isolation forest or One-Class SVM to improve data quality. Its decision function can be expressed as:

[0085] f(x)=w·x+b

[0086] Among them, w and b are model parameters, x is a data point, and the value of f(x) is used to determine whether the data point is an anomaly.

[0087] Data fusion is the integration of data from different sensors into a unified reference frame. The system uses a Kalman filter-based fusion algorithm that estimates and updates the state of the tower in real time, taking into account the uncertainty and correlation of different sensors.

[0088] In order to process large-scale real-time data streams, the system uses real-time stream processing frameworks such as Apache Kafka and Apache Storm. These technologies can efficiently process and analyze sensor data streams of sensor data sets, ensuring the real-time and reliability of data processing.

[0089] Through the above steps, the system can effectively collect, transmit and process the real-time status data of power towers, providing high-quality data support for subsequent common fault identification and health status assessment.

[0090] The data processing center module is responsible for receiving the raw data sensor data set from the sensor network module, performing data synchronization, preprocessing and anomaly detection, and finally outputting the preprocessed data set. This module includes the following main units:

[0091] 1) Data preprocessing unit:

[0092] - Use a low-pass filter to remove high-frequency noise.

[0093] -Use wavelet transform to denoise the signal. The wavelet transform formula is as follows:

[0094]

[0095] Among them, a is the scale parameter, b is the translation parameter, and ψ is the wavelet function.

[0096] - Apply Z-score or Min-Max normalization method to normalize the data.

[0097] 2) Outlier detection unit:

[0098] -Identify abnormal data points based on the Isolation Forest algorithm.

[0099] -One-Class SVM algorithm is used for anomaly detection, and its optimization objective function is:

[0100]

[0101] st(w T φ(x i ))≥p-ξ i ,ξ i ≥0

[0102] Among them, w is the weight vector, φ(x) is the kernel function that maps data to the high-dimensional feature space, and ξ i is the slack variable, ρ is the decision threshold, and l is the total number of data points.

[0103] The common fault identification system module receives the preprocessed data set and the digital twin model for feature extraction, model training and fault identification, and outputs the fault state set. This module includes the following main units:

[0104] 1) Feature extraction unit:

[0105] - Extract key features based on principal component analysis (PCA) technology;

[0106] C=(1 / n)∑(Xi-μ)(Xi-μ)^T

[0107] [U,S,V]=SVD(C)

[0108] Y=U^T·X

[0109] Among them, C is the covariance matrix, U is the eigenvector matrix, S is the singular value matrix, and Y is the feature after dimensionality reduction.

[0110] - Use independent component analysis (ICA) technology to separate independent signal sources:

[0111] max J(y)=H(y)-I(s,y)

[0112] Among them, H(y) is the output entropy and I(s,y) is the mutual information.

[0113] 2) Model training unit:

[0114] -Support Vector Machine (SVM) algorithm:

[0115] min(1 / 2)||w||^2+C∑ξi

[0116] styi(w·xi+b)≥1-ξi,ξi≥0

[0117] Among them, w is the weight vector, b is the bias term, C is the penalty parameter, and ξi is the slack variable

[0118] -Random forest algorithm: construct multiple decision trees and use voting mechanism for classification.

[0119] - Gradient Boosting Machine (GBM) algorithm: optimizes the loss function through iteration.

[0120] F(x) = ∑γm·h(x; am)

[0121] Among them, γm is the step size and h(x;am) is the base learner.

[0122] -Deep neural network: uses a multi-layer perceptron structure and is trained using a back-propagation algorithm.

[0123] 3) Fault identification unit:

[0124] -Combining the outputs of multiple models, an ensemble learning approach is used to improve recognition accuracy.

[0125] -Real-time update of fault status set, including fault type, location and severity.

[0126] Through the collaborative work of the above modules, the system can effectively identify various common problems of power towers and provide a reliable basis for subsequent health status assessment and maintenance decisions.

[0127] 4. Health Status Assessment Module

[0128] The health status assessment module is responsible for receiving the fault status set and the digital twin model, performing fault severity analysis and risk prediction, and finally outputting the health index. This module includes the following main units:

[0129] 1) Failure Severity Analysis Unit: Use the Stress Concentration Factor (SCF) to quantify the stress increase caused by a crack or notch:

[0130]

[0131] Among them, σ max is the maximum stress, σ nom is the nominal stress.

[0132] Finite Element Analysis (FEA) is used to simulate the impact of failures on the structural integrity of the tower:

[0133]

[0134] Among them, σ feais the stress simulated by FEA, N is the force, and A is the effective cross-sectional area of ​​the tower. The Markov chain model is used to predict the transition probability of the fault state:

[0135]

[0136] Among them, P ij is the probability of transitioning from state i to state j, T ij is the number of transitions from i to j in the state transition matrix.

[0137] The Cox proportional hazards model was used to analyze the time to failure:

[0138] h(t)=h0(t)exp(βX)

[0139] Among them, h(t) is the risk function at time t, h0(t) is the benchmark risk function, β is the risk coefficient, and X is the influencing factor vector.

[0140] Remaining Useful Life (RUL) predictions can be done using physics-based degradation models or data-driven methods such as the Kaplan-Meier estimator:

[0141]

[0142] in, is the survival probability at time t, n i is the number of units that failed at time i, N i is the total number of units at risk at start time i.

[0143] The health index calculation takes into account the structural integrity of the tower, historical maintenance records and real-time monitoring data. The health index of the tower is calculated using a weighted comprehensive method:

[0144] HI=w1·S1+w2·S H +w3·S M

[0145] Among them, HI is the health index, S I , S H and S M are the standardized scores of structural integrity, historical maintenance and real-time monitoring, respectively, and w1, w2 and w3 are the corresponding weights.

[0146] 5. Maintenance Decision Support Module

[0147] The maintenance decision support module receives the health index and digital twin model, determines the maintenance priority and optimizes resource allocation, and finally generates a maintenance plan. This module includes the following main units:

[0148] 1) Maintenance priority determination unit based on risk matrix:

[0149] -Construct a risk matrix with the horizontal axis representing the probability of failure and the vertical axis representing the severity of the consequences of the failure.

[0150] -Determine the maintenance priority for each tower by combining health index, tower importance and maintenance resource availability.

[0151] 2) Time-cost optimization unit:

[0152] -Establish multi-objective optimization model:

[0153] minf1(x)=∑ i c i ·x i (Total maintenance cost)

[0154] minf2(x)=max(t i ·x i )(Maximum completion time)

[0155]

[0156] Among them, c i is the cost of maintaining task i, t i is the completion time of maintenance task i, x i is the decision variable.

[0157] 3) Resource Scheduling Unit:

[0158] -Resource allocation based on the Constraint Satisfaction Problem (CSP) model:

[0159] V = {v1, v2, ..., vn} (variable set)

[0160] D={D1,D2,...,Dn}(value range set)

[0161] C = {c1, c2, ..., cm} (constraint set)

[0162] -Use the backtracking search algorithm to solve the CSP and obtain a resource allocation solution that satisfies all constraints

[0163] 4) Maintenance plan generation unit:

[0164] - Comprehensively consider maintenance priorities, time-cost optimization results and resource allocation plans

[0165] -Generate a detailed maintenance plan including a list of maintenance tasks, execution time, required resources, etc.

[0166] Through the collaborative work of the above modules, the system can make scientific maintenance decisions based on the health status of the towers, realize preventive maintenance of power towers and efficient use of resources.

[0167] 6. Data visualization platform module, user interaction interface module, high-performance computing and real-time data processing capabilities

[0168] The data visualization platform module is designed to intuitively and efficiently display various data and analysis results generated by the system. This module mainly includes the following units:

[0169] 1) Interactive 3D visualization unit:

[0170] This unit uses WebGL technology to implement browser-based 3D rendering. The specific implementation steps are as follows:

[0171] a. Create a scene, camera, and renderer, and set appropriate viewing angles and lighting conditions.

[0172] b. Build a 3D model of the tower based on the digital twin model, including geometric shape and material properties.

[0173] c. Implement interactive control of the model, allowing users to rotate, zoom, and translate.

[0174] d. Use color coding to display the stress distribution of each part of the tower. The mapping relationship between stress value σ and color can be expressed as:

[0175] color = f(σ) = {

[0176] Red, ifσ>σmax

[0177] Yellow, if σthreshold<σ≤σmax

[0178] Green, ifσ≤σthreshold

[0179] }

[0180] Among them, σmax is the maximum allowable stress of the material, and σthreshold is the warning threshold.

[0181] 2) Custom data report generation unit:

[0182] This unit creates interactive data visualization charts and generates comprehensive reports. The specific implementation is as follows:

[0183] a. Health trend chart: Use time series analysis methods, such as exponential smoothing, to predict future health index trends.

[0184] Prediction formula: H(t+1)=α*H(t)+(1-α)*H'(t)

[0185] Among them, H(t) is the actual health index at time t, H'(t) is the predicted value at time t, and α is the smoothing coefficient.

[0186] b. Maintenance Gantt chart: Generate a task timeline based on the maintenance plan and use different colors to indicate task priorities.

[0187] c. Fault distribution map: Use a heat map to display the fault frequency of each part of the tower, and the color depth indicates the probability of fault occurrence.

[0188] The user interaction interface module provides an intuitive and friendly operating environment for system operators. This module mainly includes the following units:

[0189] 1) Personalized recommendation unit:

[0190] This unit uses collaborative filtering algorithms and reinforcement learning methods to recommend maintenance strategies for users. The specific implementation is as follows:

[0191] a. Collaborative filtering algorithm: Based on the user-item matrix, calculate the similarity between users and predict the user's interest in unrated items.

[0192] The similarity is calculated using cosine similarity: sim(a,b)=cos(θ)=(A·B) / (||A||||B||), where A and B represent the rating vectors of users a and b, respectively.

[0193] b. Reinforcement learning algorithm: Use the Q-learning method to optimize the recommendation strategy. The Q value update formula is:

[0194] Q(s,a)←Q(s,a)+α[r+γmax Q(s',a')-Q(s,a)]

[0195] Among them, s is the current state, a is the selected action, r is the immediate reward, γ is the discount factor, and α is the learning rate.

[0196] 2)Maintenance log tracking unit:

[0197] This unit uses blockchain technology to ensure that maintenance records cannot be tampered with. The specific implementation is as follows:

[0198] a. Block structure: contains index, timestamp, data, hash value of the previous block and hash value of the current block.

[0199] b. Hash calculation: Use SHA-256 algorithm to ensure data integrity.

[0200] c. Consensus mechanism: Proof of Work (PoW) mechanism is adopted, and the difficulty value is dynamically adjusted to ensure the stability of block generation speed.

[0201] To handle large amounts of sensor data and support complex analytical models, the system also includes the following features:

[0202] 1) High-performance computing cluster unit:

[0203] This unit uses a distributed computing framework to achieve large-scale data processing. The main features include:

[0204] a. Data partitioning: Use consistent hashing algorithm to evenly distribute data to cluster nodes.

[0205] b. Parallel computing: Using the MapReduce model, the Map function processes a single data point, and the Reduce function aggregates the processing results.

[0206] c. Fault-tolerant mechanism: implement data replication and task rescheduling to ensure system reliability.

[0207] 2) Real-time stream processing unit:

[0208] This unit processes sensor data streams of sensor datasets. The main features include:

[0209] a. Sliding window model: Use time window W and sliding step size S to group and process data.

[0210] b. Anomaly detection: Use the Z-score method to identify outliers in real time, Z = (x-μ) / σ

[0211] Among them, x is the observed value, μ is the historical average, and σ is the standard deviation.

[0212] c. Real-time aggregation: Use approximate algorithms (such as HyperLogLog) to calculate aggregate statistics of large-scale data streams.

[0213] Through the collaborative work of the above modules and functions, the system can efficiently process and analyze large amounts of real-time data, providing strong support for the condition monitoring and maintenance decisions of power towers.

Claims

1. A real-time monitoring and common fault identification system for power tower status based on digital twin technology, characterized by ,include: -High-precision digital twin model module, which is used to receive the design parameters, material properties, construction records and historical maintenance data of the tower, and build a digital twin model containing geometric models and physical properties; -Sensor network module, which is used to deploy sensors according to the location information of key monitoring points in the digital twin model, collect the force, displacement and vibration data of the tower in real time, and generate sensor data sets; -Data processing center module, used to receive sensor data sets, perform data synchronization, preprocessing and anomaly detection, and output preprocessed data sets; - Common fault identification system module, used to receive pre-processed data sets and digital twin models, perform feature extraction, model training and fault identification, and output fault status sets; -Health status assessment module, which is used to receive the fault status set and digital twin model, perform fault severity analysis and risk prediction, and output the health index; -Maintenance decision support module, which is used to receive health index and digital twin model, determine maintenance priority and optimize resource allocation, and generate maintenance plan.

2. The system according to claim 1, characterized in that ,The high-precision digital twin model module includes: -Environmental simulation unit, used to simulate the mechanism of the impact of environmental factors such as wind, temperature, humidity, etc. on the tower; -Real-time update unit, used to synchronously update the digital twin model through wireless or wired communication to reflect the real-time status of the tower.

3. The system according to claim 1, characterized in that ,The sensor network module includes: -An adaptive algorithm unit for adjusting sensor positions based on real-time feedback to enhance the accuracy of the sensor dataset; -Data synchronization algorithm unit to handle communication delays or data loss and ensure the spatiotemporal consistency of sensor data sets.

4. The system according to claim 1, characterized in that ,The data processing center module includes: -Data preprocessing unit for processing sensor data sets using low-pass filter, wavelet transform, Z-score or Min-Max normalization method; -Outlier detection unit, used to automatically identify and remove abnormal data points in the preprocessed data set based on the isolation forest or One-Class SVM algorithm.

5. The system according to claim 1, characterized in that ,The common disease identification system module includes: -Feature extraction unit, used to extract key features from the preprocessed data set based on PCA and ICA techniques; -Model training unit, used to perform fault pattern recognition using SVM, random forest, GBM and deep neural network algorithms to generate fault state sets.

6. The system according to claim 1, characterized in that ,The health status assessment module includes: -Failure severity analysis unit, used to quantify the stress increase caused by concentrated cracks or notches in the failure state using stress concentration factors; -Risk prediction unit, used to predict the probability of fault state transition using Markov chain analysis and determine the remaining service life of the tower.

7. The system according to claim 1, characterized in that ,The maintenance decision support module includes: -Maintenance priority determination unit based on risk matrix, which is used to determine maintenance priority by combining health index, tower importance and maintenance resource availability; -Time-cost optimization unit, used to balance the time and cost of completing tasks in the maintenance plan.

8. The system according to claim 1, characterized in that , also includes a data visualization platform module, the data visualization platform module includes: -Interactive 3D visualization unit, which allows users to observe the digital twin model from different angles and scales; - Custom data report generation unit for generating reports based on fault status sets, health indices and maintenance plans.

9. The system according to claim 1, characterized in that , further comprising a user interaction interface module, the user interaction interface module comprising: -Personalized recommendation unit, used to recommend maintenance strategies in the maintenance plan based on the user's historical selections and preferences; -Maintenance log tracking unit to record maintenance activities and results and update the digital twin model.

10. The system according to claim 1, characterized in that , also includes: -High-performance computing cluster unit, used to process the pre-processed data set using MapReduce parallel processing technology; -Real-time stream processing unit for processing sensor data streams of sensor datasets.

Images