Insulator health state detection method and system

Abstract

The invention relates to an insulator health state detection method and system, and belongs to the technical field of insulator detection. The method comprises the following steps: according to a sensor array formed by carrying out ad hoc network on more than two of a camera, a vibration sensor, an ultrasonic sensor, an ultraviolet sensor and an infrared sensor, acquiring data acquired by each sensor at each part of an insulator in real time, in each detection period, feature extraction is performed on data acquired by each sensor at each part of the insulator in the current detection period, then feature fusion is performed on the extracted features, and the health state of each part of the insulator is obtained according to the feature fusion result of each part. And obtaining the overall health state of the insulator according to the health state of each part of the insulator. The obtained target information is more reliable and accurate, the influence of external adverse factors on the insulator health state detection result can be reduced, and the detection result is more accurate.

Description

Technical Field

[0001] The invention relates to a method and a system for detecting the health status of an insulator, and belongs to the technical field of insulator detection. Background Art

[0002] The open-type switch (AIS) is mainly used to connect different electrical equipment in the substation to transmit electrical energy, and plays an important role in the distribution network. As the voltage level increases, the equipment tends to become larger and higher to ensure a safe distance, making manual detection increasingly difficult.

[0003] During the opening and closing operation of the switch, insulators (such as porcelain insulators) are subjected to various forces, and typical faults such as tilting and breaking often occur in actual operation. Statistics show that 60%-70% of circuit breaker failures are caused by insulator damage, which seriously threatens the safe and stable operation of the converter station. There are several reasons for insulator failures. First, it is damaged due to poor manufacturing process, inadequate installation and commissioning, second, the existence of mechanical stress during operation, and third, the influence of extreme weather conditions in outdoor operation, such as sharp temperature differences and acid rain. Defective insulators often change their inherent frequency and decrease their insulation resistance, and even cause local discharge and local overheating. Therefore, it is necessary to use a reasonable method to detect the health status of the insulators, predict possible faults of the insulators, and avoid affecting the safe and stable operation of the converter station.

[0004] At present, many institutions and scholars at home and abroad have carried out relevant research and proposed different methods for detecting the health status of insulators, such as insulation resistance measurement, ultrasonic detection, vibration acoustic detection, partial discharge detection, ultraviolet imaging detection, infrared detection and visual inspection, etc. However, in practice, these methods may have problems with accuracy and reliability of the target information obtained due to the accuracy of the sensor itself and interference from the external environment (such as the influence of factors such as altitude and ambient light), resulting in low accuracy of the insulator health status detection results. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for detecting the health status of an insulator, so as to solve the problem that the accuracy and reliability of target information obtained by the current method for detecting the health status of an insulator are low, thereby resulting in low accuracy of the insulator health status detection results.

[0006] In order to solve the above technical problems, the present invention provides a method for detecting the health status of an insulator, the method comprising: according to a sensor array formed by self-organizing a network of two or more sensors selected from cameras, vibration sensors, ultrasonic sensors, ultraviolet sensors and infrared sensors, real-time acquisition of data collected by sensors at various parts of the insulator is carried out, in each detection cycle, feature extraction is performed on the data collected by sensors at various parts of the insulator in the current detection cycle, and then feature fusion is performed on the extracted features, the health status of each part of the insulator is obtained according to the feature fusion results of each part, and the overall health status of the insulator is obtained according to the health status of each part of the insulator.

[0007] Furthermore, the method of self-organizing two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor includes: using the LoRa self-organizing network method to self-organize two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor.

[0008] Furthermore, the power supply method for the sensors in the formed sensor array includes: using electromagnetic induction power supply method to power the sensors in the formed sensor array through the lines of the substation where the insulators are located.

[0009] Furthermore, it also includes: when the health status of a certain part of the insulator obtained according to the feature fusion result of a certain part is abnormal, the corresponding position of the part is marked and displayed accordingly on the three-dimensional model of the insulator established according to the digital twin technology.

[0010] Furthermore, the method of extracting features from the data collected by the two or more sensors at various positions of the insulator includes: preprocessing the data collected by the two or more sensors at various positions of the insulator and then extracting features.

[0011] Beneficial effects: The present invention provides a new method for detecting the health status of an insulator. The present invention uses a sensor array formed by self-organizing a network of two or more sensors among cameras, vibration sensors, ultrasonic sensors, ultraviolet sensors and infrared sensors to collect sensor data from each part of the insulator, so that the collected data is more accurate and reliable. In each detection cycle, feature extraction is performed on the data collected by each sensor of each part of the insulator in the current detection cycle, and then the extracted features are fused. The health status of each part of the insulator is obtained according to the feature fusion results of each part, and the overall health status of the insulator is obtained according to the health status of each part of the insulator. Since different sensors have differences in principles and physical properties, the health status of each part of the insulator can be obtained by fusing the features obtained by various sensors to make the data collected by each sensor complementary. Therefore, the target information obtained by this method is more reliable and accurate, thereby reducing the impact of adverse factors in the outside world on the insulator health status detection results, making the detection results more accurate.

[0012] The present invention also provides an insulator health status detection system, which includes a processor, and the processor is used to execute a computer program to implement the steps of the above-mentioned insulator health status detection method.

[0013] The insulator health status detection system can achieve the same beneficial effects as the above-mentioned insulator health status detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flowchart of a method for detecting the health status of an insulator in an embodiment of the method for detecting the health status of an insulator of the present invention; Figure 2 is an example diagram of sensor installation or operation in an embodiment of the insulator health status detection method of the present invention; Figure 3 It is an example diagram of a LoRa self-organizing network in an embodiment of the method for detecting the health status of an insulator of the present invention. DETAILED DESCRIPTION

[0015] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings.

[0016] Embodiment of the method for detecting the health status of an insulator This embodiment provides a technical solution for a method for detecting the health status of an insulator. The method forms a sensor array by self-organizing different sensors, and then performs data fusion analysis on the data of various parts of the insulator collected by the sensors in the sensor array to obtain a health status detection result of the insulator.

[0017] The method specifically includes: according to a sensor array formed by self-organizing a network of two or more sensors among cameras, vibration sensors, ultrasonic sensors, ultraviolet sensors (also called ultraviolet cameras) and infrared sensors (also called infrared cameras), real-time acquisition of data collected by each sensor at each part of the insulator, feature extraction of data collected by each sensor at each part of the insulator in the current detection cycle in each detection cycle, and then feature fusion of the extracted features, obtaining the health status of each part of the insulator according to the feature fusion results of each part, and obtaining the overall health status of the insulator according to the health status of each part of the insulator. Due to the differences in the principles and physical properties of different sensors, the health status of each part of the insulator can be obtained by fusing the features obtained by various sensors to make the data collected by each sensor complementary, so the target information obtained by this method is more reliable and accurate, and the influence of adverse factors in the outside world on the detection results of the insulator health status can be reduced, making the detection results more accurate.

[0018] The flow chart of the insulator health status detection method is as follows: Figure 1 As shown below, combined Figure 1 The method for detecting the health status of insulators is described in detail: According to the sensor array formed by self-organizing two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor, the sensor installation or operation example diagram is as follows Figure 2As shown, this embodiment adopts a camera (i.e., a camera on a drone and other cameras arranged at fixed positions. Since the fixed cameras on the construction site may be blocked or there are blind spots in the viewing angle, in this case, the camera of the drone can be used for supplementary shooting. In other embodiments, only the camera on the drone or other cameras arranged at fixed positions can be used. In this embodiment, both are used), an infrared camera (infrared sensor), an ultraviolet camera (ultraviolet sensor), a vibration sensor and an ultrasonic sensor are self-organized to form a sensor array. Each sensor is respectively arranged at each part of the insulator. Specifically, cameras are arranged on the drone and other fixed positions for taking images of each part of the insulator. An infrared camera (infrared sensor) The external sensor) is used to take infrared images of various parts of the insulator to detect abnormal temperatures; and the ultraviolet camera (ultraviolet sensor) is used to take ultraviolet images of various parts of the insulator to detect discharge phenomena; among them, the ultrasonic sensor is suitable for detecting cracks inside the flange of the insulator, but cannot perform live detection, and the vibration sensor determines the health status of the part to be tested by applying vibration excitation to the part to be tested, and the vibration sensor can perform live detection. Therefore, in this embodiment, a sensor array including vibration sensors and ultrasonic sensors is used to detect different aspects of various parts of the insulator, which can achieve complementary effects of different detection quantities to reduce the impact of adverse factors in the environment on the detection results, thereby improving the accuracy of health status detection.

[0019] In this embodiment, the method of self-organizing two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor is specifically as follows: using the LoRa self-organizing network method to self-organize two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor.

[0020] An example of a LoRa self-organizing network is shown below Figure 3 As shown ( Figure 3 The drone in the description refers to the camera on the drone, and the camera refers to other cameras arranged at fixed positions). The two LoRa gateways are respectively connected to a sensor array consisting of a camera (camera), a vibration sensor, an ultrasonic sensor, an ultraviolet sensor, and an infrared sensor, which are used to obtain data collected by each sensor at each part of the insulator. The two LoRa gateways communicate with the intranet and send the collected data to the intranet. In other embodiments, the sensor self-organizing network can also be realized through other existing wired communication sensor self-organizing network methods (such as optical fiber or CAN line connected sensor self-organizing network methods, etc.) or wireless communication sensor self-organizing network methods (such as WIFI or Bluetooth connected sensor self-organizing network methods, etc.), and the specific contents and principles of these self-organizing network methods are not repeated here.

[0021] In this embodiment, data collected by various sensors at various parts of the insulator are obtained in real time, and features are extracted from the data collected by various sensors at various parts of the insulator in the current detection cycle in each detection cycle. The extracted features are then fused, and the health status of each part of the insulator is obtained based on the feature fusion results of each part. The overall health status of the insulator is obtained based on the health status of each part of the insulator.

[0022] Among them, the method of extracting features from the data collected by more than two sensors at various parts of the insulator includes: preprocessing the data collected by more than two sensors at various parts of the insulator and then extracting features. The preprocessing method is the preprocessing inside each sensor. Specifically, the method of preprocessing the data collected by the camera is CNN, the method of preprocessing the data collected by the vibration sensor is wavelet analysis to extract the defect vibration abnormal signal, the method of preprocessing the data collected by the ultrasonic sensor is wavelet analysis to extract the defect echo signal, the method of preprocessing the data collected by the ultraviolet sensor includes grayscale processing, opening and closing operation filtering, small area elimination, contour extraction, and the method of preprocessing the data collected by the infrared sensor includes grayscale processing, opening and closing operation filtering, small area elimination, and contour extraction; wherein, the data collected by each sensor that undergoes preprocessing is the collected original data.

[0023] In this embodiment, the feature extraction adopts the random forest method. In the process of compressing the features after feature extraction and then fusing the features, the compression adopts PCA or SVD. Compression can reduce the processing amount of original data and improve the processing speed and real-time performance of the system. The method of feature fusion is specifically as follows: the feature vectors of the features after feature extraction and compression are jointly judged and processed to obtain the final fusion result. In other embodiments, feature compression may not be performed, and the extracted features may be directly fused. The method of feature fusion is specifically as follows: the feature vectors of the features after feature extraction are jointly judged and processed to obtain the final fusion result. In this embodiment, the feature fusion adopts the wavelet packet transform algorithm. In other embodiments, other feature fusion algorithms may also be used. Feature fusion improves the system's capacity and fault tolerance for multi-source heterogeneous sensors and reduces the data transmission and storage volume.

[0024] The result after feature fusion is subjected to decision analysis to output the health status of each part of the insulator. The method adopted in the decision analysis belongs to the existing technology, such as the CNN decision method, which will not be described here.

[0025] The power supply method of the sensors in the sensor array formed in this embodiment is specifically: the sensors in the sensor array are powered by electromagnetic induction through the lines of the substation where the insulator is located. Among them, the lines of different voltage levels in the substation can power the sensors in the sensor array. The electromagnetic induction power supply method of this embodiment avoids the need for cumbersome wiring or setting up heavy battery packs in the existing power supply method, and improves the efficiency of insulator health status detection.

[0026] In this embodiment, the insulator health status detection method further includes: when the health status of a certain part of the insulator obtained according to the feature fusion result of a certain part is abnormal, the position corresponding to the part is marked and displayed accordingly on the three-dimensional model of the insulator established according to the digital twin technology. Specifically, the corresponding position of the part can be marked and displayed in various existing marking methods such as highlighting, text marking, pattern marking or frame marking, which will not be repeated here.

[0027] Embodiment of the insulator health status detection system This embodiment provides a technical solution for a system for detecting the health status of an insulator. The system includes a processor, and the processor is used to implement the steps of the method for detecting the health status of an insulator in an embodiment of the method for detecting the health status of an insulator.

[0028] The specific implementation process and principle of the insulator health status detection system have been described in detail in the insulator health status detection method embodiment, and will not be repeated here.

Claims

1. A method for detecting the health status of an insulator, characterized in that: include: According to the sensor array formed by self-organizing networking of two or more sensors among cameras, vibration sensors, ultrasonic sensors, ultraviolet sensors and infrared sensors, data collected by each sensor at each part of the insulator is obtained in real time, and features are extracted from the data collected by each sensor at each part of the insulator in the current detection cycle in each detection cycle, and then the extracted features are fused. The health status of each part of the insulator is obtained according to the feature fusion results of each part, and the overall health status of the insulator is obtained according to the health status of each part of the insulator.

2. The method for detecting the health status of an insulator according to claim 1, characterized in that: The method of self-organizing two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor includes: using the LoRa self-organizing network method to self-organize two or more sensors among the camera, vibration sensor, ultrasonic sensor, ultraviolet sensor and infrared sensor.

3. The method for detecting the health status of an insulator according to claim 1 or 2, characterized in that: The power supply method for the sensors in the formed sensor array includes: using an electromagnetic induction power supply method to power the sensors in the formed sensor array through the lines of the substation where the insulators are located.

4. The method for detecting the health status of an insulator according to claim 1 or 2, characterized in that: Also includes: When the health status of a certain part of the insulator obtained based on the feature fusion result of a certain part is abnormal, the corresponding position of the part is marked and displayed accordingly on the three-dimensional model of the insulator established based on the digital twin technology.

5. The method for detecting the health status of an insulator according to claim 1 or 2, characterized in that: The method of extracting features from data collected by more than two sensors at various positions of the insulator includes: pre-processing the data collected by more than two sensors at various positions of the insulator and then extracting features.

6. An insulator health status detection system, comprising a processor, characterized in that: The processor is used to execute a computer program to implement the steps of the insulator health status detection method according to any one of claims 1 to 5.

Images