Electric power equipment operation state evaluation system and evaluation method

Abstract

The invention discloses a power equipment operation state evaluation system and evaluation method, and relates to the technical field of power equipment state monitoring and evaluation, the system comprises a data acquisition module, a data integration module, a data analysis module, a data fusion module, a state evaluation module and a decision support module; by integrating various data acquisition means, including traditional sensor data acquisition, equipment appearance image and video information acquisition and a terahertz nondestructive testing technology, comprehensive and real-time monitoring of the operation state of the power equipment is realized, and the diversified data acquisition mode has the advantages of high accuracy and high reliability. The accuracy and integrity of data collection are improved, an information basis is provided for subsequent data analysis and state evaluation, the introduction of a terahertz detection technology makes it possible to detect an internal insulating material and structure of equipment, the blank of a traditional detection means in the aspect is filled, and the detection efficiency is improved. And the accuracy and reliability of state evaluation of the power equipment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment status monitoring and evaluation, and in particular to a power equipment operating status evaluation system and evaluation method.

[0002] In the electric power industry, the monitoring and evaluation of the operating status of power equipment is a key link to ensure the safe and stable operation of the power grid. With the continuous development of science and technology, power equipment has become increasingly complex, and higher requirements have been put forward for the monitoring of its operating status. At present, the power industry is actively exploring and applying new technologies and methods to improve the accuracy and efficiency of equipment monitoring. In this context, a power equipment operating status evaluation system that integrates multiple data collection and analysis methods has emerged, aiming to achieve comprehensive, real-time and accurate monitoring of the operating status of power equipment.

[0003] Traditional power equipment operating status monitoring technology mainly relies on single sensor data collection and analysis, such as voltage transformers, current transformers, and temperature sensors. Although these technologies can reflect the operating status of the equipment to a certain extent, they have many shortcomings. On the one hand, a single data source often cannot fully reflect the overall operating status of the equipment and is prone to missing key information; on the other hand, traditional technologies have blind spots in monitoring the internal insulation materials and structures of the equipment, making it difficult to discover potential safety hazards. In addition, traditional technologies have limited data analysis and processing capabilities and are difficult to adapt to large-scale and complex power equipment monitoring needs. Summary of the invention

[0004] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide a power equipment operating status evaluation system and evaluation method. The present invention integrates data collection, integration, analysis, fusion, evaluation and decision support, and realizes comprehensive, real-time and accurate monitoring of the operating status of power equipment. By adopting a variety of sensors, cameras and terahertz detection equipment, it can obtain multi-dimensional information inside and outside the equipment, and use advanced data analysis algorithms and fusion technologies to construct a feature vector reflecting the current operating status of the equipment. Combined with historical data and comprehensive evaluation functions, it can accurately judge the operating status of the equipment, provide scientific decision support for operation and maintenance personnel, and effectively improve the safe operation level and maintenance efficiency of power equipment.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a power equipment operation status evaluation system includes a data acquisition module, a data integration module, a data analysis module, a data fusion module, a status evaluation module and a decision support module;

[0006] The data acquisition module is equipped with a variety of sensors for collecting the operating data of the power equipment; equipped with cameras to obtain images and video information of the equipment appearance from different angles; equipped with terahertz detection equipment to use terahertz waves to perform non-destructive detection of the internal insulation materials and structures of the equipment. All acquisition devices transmit the collected raw data to the data integration module in real time through wireless communication;

[0007] The data integration module receives data from the data acquisition module, performs format verification and cleaning operations on the data, removes invalid data and noise data, synchronizes the timestamp of the cleaned data, and aligns the data of different acquisition devices on the time axis;

[0008] The data analysis module: uses different algorithms to analyze different types of data respectively. For traditional sensor data, conventional numerical analysis and trend judgment are performed. For image data, the module identifies whether the equipment has external damage, signs of oil leakage, or foreign matter adhesion. For terahertz data, the module performs spectrum analysis on the terahertz detection data based on the propagation characteristics and reflection and refraction laws of terahertz waves in different media to determine the moisture content of the insulation material inside the equipment, the location and size of tiny defects, and the aging degree of the insulation material. The analysis results are transmitted to the data fusion module in the form of structured data.

[0009] The data fusion module: after receiving various analysis results from the data analysis module, uses information entropy to determine the weights of various types of data in the fusion process, fuses data from different sources through a data fusion algorithm, comprehensively considers various factors, constructs a feature vector that can reflect the current operating state of the transformer, and outputs it to the state assessment module;

[0010] The state evaluation module collects various data samples of power equipment under different operating conditions in the past, including normal operation, abnormal operation before failure, and data when failure occurs, and labels and preprocesses these data. The power equipment operation state evaluation model is constructed through a comprehensive evaluation function. The equipment state feature vector output by the data fusion module is set as Where n is the number of features, and the health baseline vector is defined as The fault feature vector is defined as Calculate the deviation S between the equipment status feature vector and the healthy baseline vector d , the formula is: A1 i is the weight of the i-th feature, and calculates the proximity S between the device state feature vector and the fault feature vector c , the formula is: Among them B1 iis another weight parameter, e is the base of the natural logarithm, and the operating status of the equipment is judged by the comprehensive evaluation function s. The calculation formula is: The model is trained with historical data. After receiving the device status feature vector output by the data fusion module, it is input into the evaluation model to calculate the comprehensive evaluation score of the output device. The corresponding status grade is given according to the comprehensive evaluation score, and the evaluation result is transmitted to the decision support module.

[0011] The decision support module provides decision support information to the operation and maintenance personnel based on the equipment operation status evaluation results given by the status evaluation module. When the equipment is operating normally, it provides optimization operation suggestions. When the equipment is in an abnormal state, it provides specific troubleshooting steps and maintenance plan suggestions. At the same time, it interacts with the external system.

[0012] Furthermore, the sensors used in the data acquisition module are: a voltage transformer, a current transformer and a pressure sensor.

[0013] Furthermore, the data integration module processes the received data:

[0014] For sensor data, check whether the data value range and data type meet the preset standards, and repair or remove abnormal data points;

[0015] For image and video data, check the file integrity and encoding format, detect the integrity of the image; for video data, check the frame rate stability and picture continuity, mark and repair the clips with frame loss, screen distortion and damage;

[0016] For the data collected by the terahertz detection equipment, verify whether its signal strength and frequency range parameters are within the normal detection range, check the integrity of the data's phase information, identify abnormal signal fluctuation data points, and correct or exclude data that does not conform to physical laws.

[0017] Furthermore, the data analysis module analyzes the sensor data and calculates its statistical characteristics using a dynamic weighted statistical method. Suppose the data sequence is x 1 , x 2 , …, x n , introduce the time decay factor λ, and 0<λ<1, calculate the weighted mean to understand the real-time trend of the data, the formula is: Calculate the weighted variance to measure the discreteness of the data and its stability changes. The formula is: Adopt adaptive fitting algorithm to draw the trend curve of data changing with time to capture the nonlinear change trend of data. Assume that the time series is t 1 , t 2 ,…,t n, the corresponding sensor data is x 1 , x 2 , …, x n , the formula is: where t 0 is the starting time, m is the polynomial order, a k is the coefficient to be determined.

[0018] Furthermore, the analysis of the device image data in the data analysis module is performed through the appearance damage detection formula. Assuming that the pixel gray value matrix of the device area in the image is G, the size is m×n, and a two-dimensional discrete wavelet transform is performed to obtain the low-frequency coefficient matrix LL and the high-frequency coefficient matrix LH, HL, HH, and define the damage feature quantity as D, the formula is: Set a threshold T d , when D>T d When the equipment is damaged, it is determined that the equipment has external damage; the oil leakage sign detection formula is used to detect the oil leakage sign, and the image is converted from the RGB color space to the HSV color space to obtain the saturation channel matrix S, and the average saturation is calculated. The formula is: Where m and n are the number of rows and columns of the image, and the threshold T is set. When the equipment is in the state of leakage, it is judged that there is a sign of oil leakage; the foreign body adhesion detection formula is used to detect foreign body adhesion, and the edge detection of the image is performed to obtain the edge image E, whose pixel value is e(x, y), the edge is 1, and the non-edge is 0. At the same time, the normal contour of the equipment is modeled to obtain the contour template TL, whose pixel value is t(x, y), the inside of the contour is 1, and the outside of the contour is 0. The foreign body adhesion feature quantity is defined as F. The calculation formula is: Set the threshold T f , when F>T f When the sensor is turned on, it is determined that there is foreign matter attached.

[0019] Furthermore, the analysis of the terahertz detection data in the data analysis module determines the moisture content of the insulating material inside the device through a formula. Assuming that the propagation speed of the terahertz wave in the insulating material is v, the frequency is f, and the phase change after passing through the material is Δφ, it is known that the standard phase change of the material in a dry state is Δφ 0 , then the calculation formula for moisture content M is: where k 1 is the proportionality coefficient; the position and size of the tiny defect are determined by the formula. For the reflected signal intensity of the terahertz wave I(x, y), the average reflection intensity of the defect-free area is I 0 , let the relative intensity deviation function D(x, y) be calculated as: When D(x, y)>T1, T1 is the preset threshold, and it is determined that there is a small defect at the (x, y) position. The size of the defect is estimated by integrating the area that satisfies D(x, y)>T1. The formula is: S=∑ x，y：D(x，y)＞T 1×ΔxΔy, where Δx and Δy are the spatial sampling intervals; the aging degree of the insulating material is determined by the formula, and the energy integral value of the terahertz spectrum in the low frequency band is set to be E l f, the energy integral value in the high frequency band is E h f, the high-frequency and low-frequency energy ratio of the new device is: The current detected high-frequency and low-frequency energy ratio is: The calculation formula for the aging degree Ad of the insulating material is: where k 2 is the aging coefficient.

[0020] Furthermore, the data fusion module uses information entropy to determine the weights of various types of data in the fusion process. Let p ij is the probability of the jth data point appearing in the i-th type of data. For discrete data, it is obtained by statistical frequency. The information entropy H(x i ) is calculated as: H(x i )=-∑ j p ij log(p ij ), then the weight w of the i-th type of data i for: Where n is the total number of categories in the data.

[0021] Furthermore, in the data fusion module, the different types of data that have been normalized and assigned weights are organically fused through a multi-source data fusion formula. Suppose the data after normalization and weight assignment is and w i , the formula is: where α ij is the adaptive coefficient of the jth data point in the i-th class of data.

[0022] Furthermore, the status assessment module gives a corresponding status grade assessment based on the comprehensive assessment score. When S is close to 0, it indicates that the equipment is close to normal operating status; when S is close to 1, it indicates that the equipment is close to a fault state. When S<0.3, it is assessed as a normal state; when 0.3≤S≤0.7, it is assessed as an abnormal state; and when S≥0.7, it is assessed as a fault state.

[0023] On the other hand, a method for evaluating the operating status of electric power equipment is provided, wherein the specific steps of the evaluation method are:

[0024] S100, data collection: Use a variety of sensors to collect power equipment operation data, use cameras to obtain equipment appearance images and videos, use terahertz detection equipment to detect the internal insulation materials and structural information of the equipment, and transmit the raw data collected by the collection equipment to the data integration module in real time through wireless communication;

[0025] S200, data integration: after receiving the collected data, format verification and cleaning are performed on different types of data, including checking the value range of sensor data, checking the integrity of image and video data, and verifying the signal parameters of terahertz detection data. At the same time, the data is synchronized and aligned with the timestamp, and the processed data is transmitted to the data analysis module;

[0026] S300, data analysis: Use different algorithms to analyze different types of data, and perform conventional numerical analysis and trend judgment on sensor data; image data is tested to determine whether there is any appearance damage, oil leakage, or foreign matter attachment; terahertz detection data determines the moisture content, minor defects, and aging degree of insulating materials based on relevant characteristics and rules, and simultaneously transmits the analysis results to the data fusion module;

[0027] S400, data fusion: after receiving the analysis results, the information entropy is used to determine the weights of various types of data, and then the different types of data are fused through the multi-source data fusion formula to construct a feature vector reflecting the equipment operation status, and output it to the status assessment module;

[0028] S500, status evaluation: collect and pre-process equipment data samples under different operating conditions, build a power equipment operating status evaluation model through a comprehensive evaluation function, use the power equipment operating status evaluation model trained with historical data to give a status grade assessment for the received real-time data, and transmit the result to the decision support module;

[0029] S600, decision support: Provide decision support for operation and maintenance personnel based on status assessment results, give optimization suggestions during normal operation, give troubleshooting and maintenance plan suggestions during abnormal conditions, and interact with external systems to achieve information sharing and collaborative work.

[0030] Compared with the prior art, the power equipment operation status evaluation system and evaluation method of the present invention have the following beneficial effects:

[0031] 1. The present invention realizes comprehensive and real-time monitoring of the operating status of power equipment by integrating multiple data collection methods, including traditional sensor data collection, image and video information collection of equipment appearance, and terahertz non-destructive testing technology. This diversified data collection method not only improves the accuracy and completeness of data collection, but also provides an information basis for subsequent data analysis and status evaluation. The introduction of terahertz detection technology makes it possible to detect the internal insulating materials and structures of the equipment, filling the gap in traditional detection methods in this regard and improving the accuracy and reliability of power equipment status evaluation.

[0032] 2. The present invention uses the data integration module to perform format verification, cleaning and timestamp synchronization on the original data, thereby ensuring the accuracy and consistency of the data. The data analysis module uses different algorithms to analyze various types of data and extract key information. This refined data analysis makes the evaluation of the operating status of power equipment more accurate and can provide more specific and targeted decision-making support for operation and maintenance personnel. At the same time, the data fusion module uses information entropy to determine the weights of various types of data in the fusion process, and organically integrates different types of data through a multi-source data fusion algorithm to construct a feature vector that can reflect the current operating status of the equipment, further improving the accuracy and reliability of status assessment.

[0033] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a process framework diagram of the state evaluation system of the present invention;

[0035] Figure 2 The figure is a flow chart of the state evaluation method of the present invention. DETAILED DESCRIPTION

[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0037] Embodiment 1

[0038] Evaluation of the operating status of the main transformer in a 110kV substation

[0039] In a 110kV substation, for the main transformer, the voltage transformer and current transformer in the data acquisition module are installed on the high-voltage and low-voltage sides of the transformer respectively to collect voltage and current data in real time; the temperature sensor is installed in the transformer oil tank and winding to monitor temperature changes; the pressure sensor is installed inside the oil tank to detect the oil pressure. The camera is installed at a suitable position around the transformer to capture the overall appearance, bushing, and radiator of the transformer. It has optical zoom, low-light shooting, and wide dynamic range functions to obtain images and video information of the equipment appearance from different angles. During the transformer power outage and maintenance, the terahertz detection equipment performs non-destructive detection of the internal insulation materials and structures of the transformer. All acquisition equipment transmits the collected raw data to the data integration module in real time through wireless communication.

[0040] After receiving the data, the data integration module checks the sensor data to see whether the value range and data type meet the preset standards. For example, the voltage value collected by the voltage transformer should be within the normal operating voltage range. If there are data points that are out of range, they will be repaired or eliminated. For image and video data, the file integrity and encoding format are checked, and the image integrity is detected. For video data, the frame rate stability and picture continuity are checked. If it is found that the transformer appearance image is partially blurred (similar to the situation of flower screen damage), it will be marked and repaired. For the data collected by the terahertz detection equipment, its signal strength and frequency range parameters are verified to be within the normal detection range, the integrity of the data phase information is checked, abnormal signal fluctuation data points are identified, and data that does not conform to physical laws is corrected or excluded. Then, the data of different acquisition devices are aligned on the time axis through the timestamp synchronization algorithm, and the processed data is transmitted to the data analysis module.

[0041] For sensor data, the dynamic weighted statistical method is used to calculate its statistical characteristics. Suppose the collected temperature data sequence of a certain phase winding is x 1 , x 2 , …, x n , introduce the time decay factor λ, calculate the weighted mean to understand the real-time trend of the data, the formula is: Calculate the weighted variance to measure the discreteness of the data and its stability changes. The formula is: Adaptive fitting algorithm is used to draw the trend curve of data changing over time to capture the nonlinear change trend of data.

[0042] For the equipment appearance image data, the appearance damage detection formula is used for detection. Assume that the pixel gray value matrix of the transformer box area in the image is G, with a size of m×n. Perform a two-dimensional discrete wavelet transform on it to obtain the low-frequency coefficient matrix LL and the high-frequency coefficient matrix LH, HL, HH. Calculate the damage feature value. The formula is: Set a threshold T d, when D>T d When the equipment is judged to have external damage, the image is converted from RGB color space to HSV color space through the oil leakage sign detection formula, and the saturation channel matrix is ​​obtained as S, and the average saturation is calculated. The calculation formula is: Set the threshold T, when When the oil leakage is detected, the edge image E is obtained by edge detection through the foreign body adhesion detection formula. The pixel value is e(x, y), the edge is 1, and the non-edge is 0. At the same time, the normal contour of the transformer is modeled to obtain the contour template TL, whose pixel value is t(x, y), the inside of the contour is 1, and the outside of the contour is 0. The foreign body adhesion feature quantity is defined as F. The calculation formula is: Set the threshold T f , when F>T f When the sensor is turned on, it is determined that there is foreign matter attached.

[0043] For terahertz detection data, the moisture content of the insulating material inside the equipment is determined by the formula. Assume that the propagation speed of the terahertz wave in the insulating material is v, the frequency is f, and the phase change after passing through the material is Δφ. It is known that the standard phase change of the material in a dry state is Δφ 0 , then the calculation formula for moisture content M is: where k 1 is the proportional coefficient obtained through experimental calibration. The position and size of the tiny defect are determined by the formula. For the reflected signal intensity of the terahertz wave I(x, y), the average reflection intensity of the defect-free area is assumed to be I 0 , let the relative intensity deviation function D(x, y) be calculated as: When D(x, y)>T1, T1 is the preset threshold, and it is determined that there is a small defect at the (x, y) position. The size of the defect is estimated by integrating the area that satisfies D(x, y)>T1. The formula is: S=∑ x，y：D(x，y)＞T 1×ΔxΔy, where Δx and Δy are the spatial sampling intervals. The aging degree of the insulating material is determined by the formula. Assume that the energy integral value of the terahertz spectrum in the low frequency band is E l f, the energy integral value in the high frequency band is E h f, the high-frequency and low-frequency energy ratio of the new device is: The current detected high-frequency and low-frequency energy ratio is: The calculation formula for the aging degree Ad of the insulating material is: where k 2 is the aging coefficient, and then the analysis results are transmitted to the data fusion module.

[0044] After receiving various analysis results, the data fusion module uses information entropy to determine the weights of various data in the fusion process. Let pij is the probability of the jth data point appearing in the i-th type of data. For discrete data, it is obtained by statistical frequency. The information entropy H(x i ) is calculated as: H(x i )=-∑ j p ij log(p ij ), then the weight w of the i-th type of data i for: Where n is the total number of data categories. Then, the normalized and weighted data of different types are organically fused through the multi-source data fusion formula. The formula is: where α ij It is the adaptive coefficient of the jth data point in the i-th category data, and constructs a feature vector that can reflect the current operating state of the transformer and outputs it to the state assessment module.

[0045] The status assessment module collects various data samples of the main transformer in normal operation, abnormal operation before fault, and when fault occurs, such as data of oil temperature, winding temperature, voltage, and current during normal operation, data of abnormal increase in oil temperature and increase in partial discharge before fault, and data of sharp increase in short-circuit current and oil temperature when fault occurs. The data are annotated and preprocessed, and a power equipment operation status assessment model is constructed through a comprehensive evaluation function. The equipment status feature vector output by the data fusion module is set to Where n is the number of features, and the health baseline vector is defined as The fault feature vector is defined as Calculate the deviation S between the equipment status feature vector and the healthy baseline vector d , the formula is: A1 i is the weight of the i-th feature, and the closeness between the device status feature vector and the fault feature vector is calculated using the formula: A1 i is the weight of the i-th feature, and calculates the proximity S between the device state feature vector and the fault feature vector c , the formula is: Among them B1 i is another weight parameter, e is the base of the natural logarithm, and the operating status of the equipment is judged by the comprehensive evaluation function s. The calculation formula is: The model is trained using historical data. After receiving the device status feature vector output by the data fusion module, it is input into the evaluation model to calculate the comprehensive evaluation score of the output device. Assume that S=0.2 is obtained from this calculation. According to the status grade assessment rule, when S<0.3, it is assessed as a normal state; when 0.3≤S<0.7, it is assessed as an abnormal state; when S≥0.7, it is assessed as a fault state. The main transformer is assessed as a normal state, and the assessment result is transmitted to the decision support module.

[0046] Based on the equipment operation status assessment results given by the status assessment module, if the switch cabinet is in an abnormal state, the decision support module provides specific troubleshooting steps. For example, based on the relevant data of the abnormal temperature part in the data analysis, first check whether the connection of the part is loose, and then check whether the insulating parts show signs of damage; give maintenance plan suggestions, such as tightening loose connections and replacing insulating parts with signs of aging. At the same time, it interacts with external systems to share switch cabinet status information and realize collaborative work, such as notifying maintenance personnel to carry out maintenance in a timely manner and feeding back the switch cabinet status information to the operation and maintenance management platform, so as to comprehensively manage the equipment operation status of the entire distribution room.

[0047] In summary, the present invention applies the power equipment operation status evaluation system on the 110kV substation main transformer, uses a variety of sensors, cameras and terahertz detection equipment to comprehensively collect data, and through integration, analysis, fusion and evaluation, can accurately judge the operation status of the main transformer, provide decision support for operation and maintenance, and effectively ensure the safe and stable operation of the substation main transformer, reflecting the efficiency and practicality of the invention in large-scale power equipment status monitoring and management.

[0048] Embodiment 2:

[0049] Evaluation of the operating status of the switch cabinet in the 35kV distribution room

[0050] In the 35kV distribution room, for the switch cabinet, the voltage transformer and current transformer in the data acquisition module are installed on the incoming and outgoing line sides of the switch cabinet to collect voltage and current data; the temperature sensor is installed in the busbar joints, circuit breaker contacts, and cable connections in the switch cabinet that are prone to heat to monitor temperature changes; the pressure sensor is installed in the gas chamber of the switch cabinet (if it is a gas-insulated switch cabinet) to detect the air pressure, and the camera is installed at the appropriate position on the front and side of the switch cabinet to obtain the appearance image and video information of the equipment. When the switch cabinet is shut down for maintenance, the terahertz detection equipment performs non-destructive detection of the insulating materials and structures inside the switch cabinet. The acquisition equipment transmits the original data to the data integration module in real time through wireless communication.

[0051] After receiving the data, the data integration module checks the value range and data type of the sensor data. For example, the current value collected by the current transformer should be within the normal operating current range. If any abnormality occurs, it will be processed. For image and video data, its integrity and format are checked. If it is found that part of the image taken by the camera is missing (similar to the case of incomplete files), it will be marked and repaired. For terahertz detection data, the signal strength and frequency range parameters are verified, the phase information is checked, the abnormal data is corrected or excluded, and then the timestamp is synchronized. The processed data is passed to the data analysis module. For the equipment appearance image data, the appearance damage detection formula is used to detect whether the switch cabinet door and cabinet body are damaged. The oil leakage sign detection formula is used to detect whether there is oil leakage (for oil-immersed switch cabinets). The foreign matter attachment detection formula detects whether there is foreign matter attached to the surface or internal components of the switch cabinet.

[0052] For sensor data, a dynamic weighted statistical method is used, assuming that the busbar joint temperature data sequence is x 1 , x 2 , …, x n , introduce the time decay factor λ, calculate the weighted mean to understand the real-time trend of the data, the formula is: Calculate the weighted variance to measure the discreteness of the data and its stability changes. The formula is: Adaptive fitting algorithm is used to draw the trend curve of temperature changing with time. Let the time series be t 1 , t 2 ,…,t n , the corresponding sensor data is x 1 , x 2 , …, x n , the formula is: where t 0 is the starting time, m is the polynomial order, a k is the coefficient to be determined.

[0053] For the equipment appearance image data, the appearance damage detection formula is used for detection. Assume that the pixel gray value matrix of the switch cabinet area in the image is G, with a size of m×n. Perform a two-dimensional discrete wavelet transform on it to obtain the low-frequency coefficient matrix LL and the high-frequency coefficient matrix LH, HL, HH. Calculate the damage feature value. The formula is: Set the threshold T d , when D>T d When the equipment is judged to have external damage, the image is converted from RGB color space to HSV color space through the oil leakage sign detection formula, and the saturation channel matrix is ​​obtained as S, and the average saturation is calculated. The calculation formula is: Set the threshold T, when When the oil leakage is detected, the edge image E is obtained by using the foreign body adhesion detection formula. The pixel value is e(x, y), which is 1 at the edge and 0 at the non-edge. At the same time, the normal contour of the switch cabinet is modeled to obtain the contour template TL, whose pixel value is t(x, y), which is 1 inside the contour and 0 outside the contour. The foreign body adhesion feature is defined as F and the calculation formula is: Set the threshold T f , when F>T f When the sensor is turned on, it is determined that there is foreign matter attached.

[0054] For terahertz detection data, the moisture content of the insulating material inside the equipment is determined by the formula. Assume that the propagation speed of the terahertz wave in the insulating material is v, the frequency is f, and the phase change after passing through the material is Δφ. It is known that the standard phase change of the material in a dry state is Δφ 0 , then the calculation formula for moisture content M is: where k 1 is the proportional coefficient obtained through experimental calibration. The position and size of the tiny defect are determined by the formula. For the reflected signal intensity of the terahertz wave I(x, y), the average reflection intensity of the defect-free area is assumed to be I 0 , let the relative intensity deviation function D(x, y) be calculated as: When D(x, y)>T1, T1 is the preset threshold, and it is determined that there is a small defect at the (x, y) position. The size of the defect is estimated by integrating the area that satisfies D(x, y)>T1. The formula is: S=∑ x，y：D(x，y)＞T 1×ΔxΔy, where Δx and Δy are the spatial sampling intervals. The aging degree of the insulating material is determined by the formula. Assume that the energy integral value of the terahertz spectrum in the low frequency band is E l f, the energy integral value in the high frequency band is E h f, the high-frequency and low-frequency energy ratio of the new device is: The current detected high-frequency and low-frequency energy ratio is: The calculation formula for the aging degree Ad of the insulating material is: where k 2 is the aging coefficient, and then the analysis results are transmitted to the data fusion module.

[0055] After receiving various analysis results, the data fusion module uses information entropy to determine the weights of various data in the fusion process. Let p ij is the probability of the jth data point appearing in the i-th type of data. For discrete data, it is obtained by statistical frequency. The information entropy H(x i ) is calculated as: H(x i )=-∑ j p ij log(p ij), then the weight w of the i-th type of data i for: Where n is the total number of data categories. Then, the normalized and weighted data of different types are organically fused through the multi-source data fusion formula. The formula is: where α ij It is the adaptive coefficient of the jth data point in the i-th category of data, and constructs a feature vector that can reflect the current operating status of the switch cabinet and outputs it to the status evaluation module.

[0056] The status assessment module collects various data samples of the switch cabinet in normal operation, abnormal operation before fault, and when fault occurs, such as temperature, voltage, and current data during normal operation, data on increased partial discharge and abnormal temperature rise before fault, and data on short-circuit current and insulation breakdown when fault occurs. The data are annotated and preprocessed, and a power equipment operation status assessment model is constructed through a comprehensive evaluation function. The equipment status feature vector output by the data fusion module is set to Where n is the number of features, and the health baseline vector is defined as The fault feature vector is defined as Calculate the deviation S between the equipment status feature vector and the healthy baseline vector d , the formula is: A1 i is the weight of the i-th feature, and calculates the proximity S between the device state feature vector and the fault feature vector c , the formula is: Among them B1 i is another weight parameter, e is the base of the natural logarithm, and the operating status of the equipment is judged by the comprehensive evaluation function s. The calculation formula is: The model is trained using historical data. After receiving the device status feature vector output by the data fusion module, it is input into the evaluation model to calculate the comprehensive evaluation score of the output device. Assume that the value of S is obtained in this calculation. According to the status grade assessment rule, when S < 0.3, it is assessed as a normal state; when 0.3 ≤ S < 0.7, it is assessed as an abnormal state; when S ≥ 0.7, it is assessed as a fault state. The status of the switch cabinet is determined, and the assessment result is transmitted to the decision support module.

[0057] Based on the equipment operation status assessment results given by the status assessment module, if the switch cabinet is in an abnormal state, the decision support module provides specific troubleshooting steps. For example, based on the relevant data of the abnormal temperature part in the data analysis, first check whether the connection of the part is loose, and then check whether the insulating parts show signs of damage; give maintenance plan suggestions, such as tightening loose connections and replacing insulating parts with signs of aging. At the same time, it interacts with the external system of the management system of the distribution room to share the switch cabinet status information and realize collaborative work, such as notifying maintenance personnel to carry out maintenance in time and feeding back the switch cabinet status information to the operation and maintenance management platform, so as to comprehensively manage the equipment operation status of the entire distribution room.

[0058] In summary, in the evaluation of the operating status of the switch cabinet in the 35kV distribution room, with the help of the evaluation system of the present invention, from data collection to final decision support, each module works together, and through the processing and analysis of various aspects of the switch cabinet data, abnormal conditions can be discovered in time and targeted measures can be given, while interaction with the external system can be achieved. This shows that the present invention can effectively improve the equipment operation and maintenance level in the operation management of medium and low voltage power equipment, ensure the reliable operation of the switch cabinet in the distribution room, and has important practical application value.

[0059] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A power equipment operating status evaluation system, characterized in that: The system includes a data acquisition module, a data integration module, a data analysis module, a data fusion module, a state assessment module and a decision support module; The data acquisition module is equipped with a variety of sensors for collecting the operating data of the power equipment; equipped with cameras to obtain images and video information of the equipment appearance from different angles; equipped with terahertz detection equipment to use terahertz waves to perform non-destructive detection of the internal insulation materials and structures of the equipment. All acquisition devices transmit the collected raw data to the data integration module in real time through wireless communication; The data integration module receives data from the data acquisition module, performs format verification and cleaning operations on the data, removes invalid data and noise data, synchronizes the timestamp of the cleaned data, and aligns the data of different acquisition devices on the time axis; The data analysis module: uses different algorithms to analyze different types of data respectively. For traditional sensor data, conventional numerical analysis and trend judgment are performed. For image data, the module identifies whether the equipment has external damage, signs of oil leakage, or foreign matter adhesion. For terahertz data, the module performs spectrum analysis on the terahertz detection data based on the propagation characteristics and reflection and refraction laws of terahertz waves in different media to determine the moisture content of the insulation material inside the equipment, the location and size of tiny defects, and the aging degree of the insulation material. The analysis results are transmitted to the data fusion module in the form of structured data. The data fusion module: after receiving various analysis results from the data analysis module, uses information entropy to determine the weights of various types of data in the fusion process, fuses data from different sources through a data fusion algorithm, comprehensively considers various factors, constructs a feature vector that can reflect the current operating state of the transformer, and outputs it to the state assessment module; The state evaluation module collects various data samples of power equipment under different operating conditions in the past, including normal operation, abnormal operation before failure, and data when failure occurs, and labels and preprocesses these data. The power equipment operation state evaluation model is constructed through a comprehensive evaluation function. The equipment state feature vector output by the data fusion module is set as Where n is the number of features, and the health baseline vector is defined as The fault feature vector is defined as Calculate the deviation S between the equipment status feature vector and the healthy baseline vector d , the formula is: A1 i is the weight of the i-th feature, and calculates the proximity S between the device state feature vector and the fault feature vector c , the formula is: Among them B1 i is another weight parameter, e is the base of the natural logarithm, and the operating status of the equipment is judged by the comprehensive evaluation function s. The calculation formula is: The model is trained with historical data. After receiving the device status feature vector output by the data fusion module, it is input into the evaluation model to calculate the comprehensive evaluation score of the output device. The corresponding status grade is given according to the comprehensive evaluation score, and the evaluation result is transmitted to the decision support module. The decision support module provides decision support information to the operation and maintenance personnel based on the equipment operation status evaluation results given by the status evaluation module. When the equipment is operating normally, it provides optimization operation suggestions. When the equipment is in an abnormal state, it provides specific troubleshooting steps and maintenance plan suggestions. At the same time, it interacts with the external system.

2. A power equipment operating status evaluation system according to claim 1, characterized in that: The sensors used in the data acquisition module are: voltage transformer, current transformer, temperature sensor and pressure sensor.

3. The power equipment operation status evaluation system according to claim 1, characterized in that: The processing of the received data in the data integration module includes: For sensor data, check whether the data value range and data type meet the preset standards, and repair or remove abnormal data points; For image and video data, check the file integrity and encoding format, detect the integrity of the image; for video data, check the frame rate stability and picture continuity, mark and repair the clips with frame loss, screen distortion and damage; For the data collected by the terahertz detection equipment, verify whether its signal strength and frequency range parameters are within the normal detection range, check the integrity of the data's phase information, identify abnormal signal fluctuation data points, and correct or exclude data that does not conform to physical laws.

4. The power equipment operation status evaluation system according to claim 1, characterized in that: The data analysis module analyzes the sensor data by using a dynamic weighted statistical method to calculate its statistical characteristics. Suppose the data sequence is x1, x2, ..., x n , introduce the time decay factor λ, and 0<λ<1, calculate the weighted mean to understand the real-time trend of the data, the formula is: Calculate the weighted variance to measure the discreteness of the data and its stability changes. The formula is: Adaptive fitting algorithm is used to draw the trend curve of data changing over time to capture the nonlinear change trend of data. Assume that the time series is t1, t2, ..., t n , the corresponding sensor data are x1, x2, ..., x n , the formula is: Where t0 is the starting time, m is the polynomial order, a k is the coefficient to be determined.

5. The power equipment operation status evaluation system according to claim 1, characterized in that: The analysis of the equipment image data in the data analysis module is performed through the appearance damage detection formula. Assuming that the pixel gray value matrix of the equipment area in the image is G, the size is m×n, and a two-dimensional discrete wavelet transform is performed to obtain the low-frequency coefficient matrix LL and the high-frequency coefficient matrix LH, HL, HH, and define the damage feature quantity as D, the formula is: Set a threshold T d , when D>T d When the equipment is damaged, it is determined that the equipment has external damage; the oil leakage sign detection formula is used to detect the oil leakage sign, and the image is converted from the RGB color space to the HSV color space to obtain the saturation channel matrix S, and the average saturation is calculated. The formula is: Where m and n are the number of rows and columns of the image, and the threshold T is set. When the equipment is in the state of leakage, it is judged that there is a sign of oil leakage; the foreign body adhesion detection formula is used to detect foreign body adhesion, and the edge detection of the image is performed to obtain the edge image E, whose pixel value is e(x, y), the edge is 1, and the non-edge is 0. At the same time, the normal contour of the equipment is modeled to obtain the contour template TL, whose pixel value is t(x, y), the inside of the contour is 1, and the outside of the contour is 0. The foreign body adhesion feature quantity is defined as F. The calculation formula is: Set the threshold T f , when F>T f When the sensor is turned on, it is determined that there is foreign matter attached.

6. The power equipment operation status evaluation system according to claim 1, characterized in that: The data analysis module analyzes the terahertz detection data and determines the moisture content of the insulating material inside the device through a formula. Assuming that the propagation speed of the terahertz wave in the insulating material is v, the frequency is f, and the phase change after passing through the material is Δφ, it is known that the standard phase change of the material in a dry state is Δφ0, then the calculation formula for the moisture content M is: Where k1 is the proportionality coefficient; the position and size of the tiny defect are determined by the formula. For the reflected signal intensity of the terahertz wave I(x, y), the average reflection intensity of the defect-free area is set to I0, and the relative intensity deviation function is set to D(x, y). The calculation formula is: When D(x, y)>T1, T1 is the preset threshold, and it is determined that there is a small defect at the (x, y) position. The size of the defect is estimated by integrating the area that satisfies D(x, y)>T1. The formula is: S=∑ x，y：D (x, y)>T 1 ×ΔxΔy , where Δx and Δy are the spatial sampling intervals; the aging degree of the insulating material is determined by the formula, assuming that the energy integral value of the terahertz spectrum in the low frequency band is E l f, the energy integral value in the high frequency band is E h f, the high-frequency and low-frequency energy ratio of the new device is: The current detected high-frequency and low-frequency energy ratio is: The calculation formula for the aging degree Ad of the insulating material is: Where k2 is the aging coefficient.

7. The power equipment operation status evaluation system according to claim 1, characterized in that: The data fusion module uses information entropy to determine the weights of various types of data in the fusion process. Let p ij is the probability of the jth data point appearing in the i-th type of data. For discrete data, it is obtained by statistical frequency. The information entropy H(x i ) is calculated as: H(xi) = -∑ j p ij log(p ij ), then the weight w of the i-th type of data i for: Where n is the total number of categories in the data.

8. The power equipment operation status evaluation system according to claim 1, characterized in that: In the data fusion module, the different types of data that have been normalized and assigned weights are organically fused through a multi-source data fusion formula. Suppose the data after normalization and weight assignment is and w i , the formula is: where α ij is the adaptive coefficient of the jth data point in the i-th class of data.

9. The power equipment operation status evaluation system according to claim 1, characterized in that: The status assessment module gives a corresponding status grade assessment based on the comprehensive assessment score. When S is close to 0, it means that the equipment is close to normal operating status; when S is close to 1, it means that the equipment is close to fault status. When S<0.3, it is assessed as a normal status; when 0.3≤S<0.7, it is assessed as an abnormal status; when S≥0.7, it is assessed as a fault status.

10. A method for evaluating the operating status of electric power equipment, characterized in that: The specific steps of this evaluation method are: S100, data collection: Use a variety of sensors to collect power equipment operation data, use cameras to obtain equipment appearance images and videos, use terahertz detection equipment to detect the internal insulation materials and structural information of the equipment, and transmit the raw data collected by the collection equipment to the data integration module in real time through wireless communication; S200, data integration: after receiving the collected data, format verification and cleaning are performed on different types of data, including checking the value range of sensor data, checking the integrity of image and video data, and verifying the signal parameters of terahertz detection data. At the same time, the data is synchronized and aligned with the timestamp, and the processed data is transmitted to the data analysis module; S300, data analysis: Use different algorithms to analyze different types of data, and perform conventional numerical analysis and trend judgment on sensor data; image data is tested to determine whether there is any appearance damage, oil leakage, or foreign matter attachment; terahertz detection data determines the moisture content, minor defects, and aging degree of insulating materials based on relevant characteristics and rules, and simultaneously transmits the analysis results to the data fusion module; S400, data fusion: after receiving the analysis results, the information entropy is used to determine the weights of various types of data, and then the different types of data are fused through the multi-source data fusion formula to construct a feature vector reflecting the equipment operation status, and output it to the status assessment module; S500, status evaluation: collect and pre-process equipment data samples under different operating conditions, build a power equipment operating status evaluation model through a comprehensive evaluation function, use the power equipment operating status evaluation model trained with historical data to give a status grade assessment for the received real-time data, and transmit the result to the decision support module; S600, decision support: Provide decision support for operation and maintenance personnel based on status assessment results, give optimization suggestions during normal operation, give troubleshooting and maintenance plan suggestions during abnormal conditions, and interact with external systems to achieve information sharing and collaborative work.

Images