A Method and System for Risk Assessment of Misoperation of Disconnect Switches Based on Multi-Source Data Fusion

Abstract

This invention discloses a method and system for assessing the risk of disconnector switch malfunction based on multi-source data fusion, belonging to the field of disconnector switch malfunction risk management technology. The method includes the following steps: disconnector switch malfunction interference feedback monitoring; malfunction and accuracy monitoring; and abnormal superimposed interference optimization judgment. This invention improves the accuracy of disconnector switch malfunction risk assessment management through multi-source data fusion, solving the problem of low accuracy in existing multi-source data fusion disconnector switch malfunction risk assessment management.

Description

Technical Field

[0001] This invention relates to the field of risk management technology for disconnecting switch malfunctions, and in particular to a method and system for assessing the risk of disconnecting switch malfunctions based on multi-source data fusion. Background Technology

[0002] Risk assessment for disconnector switch maloperation plays a crucial role in power system operation. By integrating information from various data sources, it provides a comprehensive and accurate risk assessment for the safe operation of disconnectors, effectively preventing maloperation accidents. The specific assessment process is as follows: First, multi-source data acquisition is performed. Existing technology uses multiple cameras deployed in the operating area to collect real-time visual recognition data of the disconnector switch, including maloperation behaviors (pulling or closing the switch under load, etc.), equipment status (real-time identification of contact closing gaps, insulator surface damage, etc.), and environmental factors. Simultaneously, limit feature data is collected, including switch status signals (opening, closing, intermediate state) collected by limit sensors, trigger timing signals (opening trigger time, closing completion time), and mechanical travel data (moving contact travel length, arrival feedback signal), etc. Next, multi-source data preprocessing is performed. After spatiotemporal alignment and feature extraction of the multi-source data, the visual recognition results (contact opening distance, safety distance, etc.) and limit features (opening and closing delays, current fluctuations, etc.) are fused based on a Bayesian neural network. Finally, the fused results are input into a dynamic risk assessment model, such as ST-GCN (Spatial Temporal Graph). Convolutional Networks (Spatial-Time Graph Convolutional Networks) can be used to analyze the malfunction of disconnect switches and output the behavioral risk level. Additionally, LSTM (Long Short-Term Memory) networks can be used to detect equipment status faults and output the equipment risk level.

[0003] For example, Chinese invention patent CN111639795B discloses a method for intelligent inspection task planning for substation robots, which includes: setting inspection tasks according to the type and quantity of equipment to be inspected in the substation, and planning inspection routes according to different inspection tasks; the inspection tasks include visible light diagnostic task points, visual inspection task points, infrared diagnostic task points, and sound diagnostic task points. Visible light diagnostic task points include reading and identifying meters on inspection equipment with instruments and checking the opening and closing positions of disconnecting switches on inspection equipment with disconnecting switches. Visual inspection task points include visual inspection of inspection equipment with external insulation, shell, and respirator. Infrared diagnostic task points include point temperature measurement, line temperature measurement, area maximum temperature measurement, and area temperature distribution measurement. The inspection route is an optimal inspection path that starts from the robot's charging position and includes all inspection task points in the inspection task.

[0004] For example, Chinese invention patent CN105321039B discloses an online monitoring data management system and method for disconnecting switches, which includes: a data storage module for classifying and storing various raw data collected by the monitoring system; a data processing module for processing various raw data in the data storage module according to user-defined calculation methods or existing preset methods, and comparing the processed data with the raw data; and a data diagnosis module for analyzing the raw data and processed data in the data storage module according to standard parameters, determining whether the disconnecting switch is faulty, and providing fault handling opinions based on the fault fingerprint database.

[0005] The above-mentioned technology has at least the following technical problems:

[0006] During multi-source data acquisition, outdoor disconnect switches may become loose due to long-term environmental erosion and mechanical wear, exacerbating maloperation. For example, the impact force generated during the opening or closing of the disconnect switch can cause the camera to be unable to stably capture the target outline, resulting in blurred images and reduced feature extraction quality. This adds more noise to the visual recognition output, such as contact distance and safety distance, directly interfering with the accuracy of the visual recognition results. At the same time, limit signals also have the risk of error or failure. Looseness can indirectly cause contact jitter and hysteresis of the limit sensor, leading to deviations in key timing features such as the opening and closing trigger time and the position feedback signal. This results in opening and closing delays or timing abnormalities in the limit features, which are superimposed with visual feature noise. This further leads to deviations in the accuracy of the analysis results when analyzing disconnect switch maloperation based on fused visual recognition data and limit feature data (such as opening delay). Therefore, there is a problem of low accuracy in the risk assessment and management of disconnect switch maloperation based on multi-source data fusion. Summary of the Invention

[0007] To address the low accuracy of risk assessment and management for malfunctions of disconnecting switches using multi-source data fusion in existing technologies, this invention provides a method and system for assessing malfunctions of disconnecting switches using multi-source data fusion. The technical solution is as follows:

[0008] On the one hand, a method for assessing the risk of disconnector switch malfunction through multi-source data fusion is provided. This method includes: during multi-source data acquisition, providing disconnector switch malfunction interference feedback and outputting a feedback result reflecting the interference of disconnector switch malfunction on the data fusion process; determining whether to provide malfunction and accuracy feedback, which is used to assess the impact of the disconnector switch malfunction interference feedback result on the acquisition qualification of visual recognition data and limit feature data; if malfunction and accuracy feedback is provided, then based on the malfunction and accuracy feedback results reflecting the accuracy of visual recognition and limit feature acquisition, accuracy feedback and optimization are performed to improve the anti-interference capability of visual recognition and limit feature acquisition against disconnector switch malfunction; if malfunction and accuracy feedback is not provided, then fusion data accuracy analysis is performed, and based on the output fusion data accuracy analysis results reflecting the qualification of data fusion, determining whether to perform abnormal superposition interference optimization to reduce input noise in visual recognition data and limit feature data.

[0009] On the other hand, the multi-source data fusion disconnector malfunction risk assessment system applies a multi-source data fusion disconnector malfunction risk assessment method, including: a disconnector malfunction interference feedback monitoring module, a malfunction and accuracy monitoring module, and an abnormal superposition interference optimization judgment module. The disconnector malfunction interference feedback monitoring module is used to provide disconnector malfunction interference feedback during multi-source data acquisition and output a disconnector malfunction interference feedback result reflecting the interference of disconnector malfunction on the data fusion process, determining whether to perform malfunction and accuracy feedback. The malfunction and accuracy monitoring module, if malfunction and accuracy feedback is performed, performs accuracy feedback and optimization based on the malfunction and accuracy feedback results reflecting visual recognition accuracy and limit feature accuracy. The abnormal superposition interference optimization judgment module, if malfunction and accuracy feedback is not performed, performs fusion data accuracy analysis and, based on the output fusion data accuracy analysis results reflecting the data fusion qualification status, determines whether to perform abnormal superposition interference optimization to reduce input noise in visual recognition data and limit feature data.

[0010] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0011] 1. By providing feedback on interference from disconnector switch maloperation and determining whether to provide feedback on maloperation and accuracy, it helps to accurately locate the interference points and the degree of impact of disconnector switch maloperation on the data fusion process. This provides a clear judgment benchmark for subsequent targeted handling of data fusion anomalies, avoiding resource waste caused by blind optimization. If feedback on maloperation and accuracy is provided, then accuracy feedback and optimization are performed, which helps to improve the accuracy of visual recognition and the reliability of limit feature judgment during disconnector switch data fusion. If no feedback on maloperation and accuracy is provided, then the accuracy of the fused data is analyzed, and based on the output of the fused data accuracy analysis results, it is determined whether to optimize for abnormal superposition interference. This helps to promptly identify and handle possible abnormal superposition interference problems during the data fusion process, providing reliable data support for disconnector switch operation status monitoring and maintenance decisions. In turn, it helps to improve the accuracy of disconnector switch maloperation risk assessment and management in multi-source data fusion, solving the problem of low accuracy in the existing technology for multi-source data fusion disconnector switch maloperation risk assessment and management.

[0012] 2. By first comparing whether the interference factor of the unqualified opening misoperation is greater than that of the unqualified closing misoperation, and then deciding whether to use the unqualified opening misoperation interference value, the unqualified closing misoperation interference value, or both the unqualified opening misoperation interference value and the unqualified closing misoperation interference value as the harmonic average with the visual recognition accuracy value or the limit feature accuracy value, compared with the single judgment and analysis based on a single parameter in the existing technology, it helps to improve the comprehensiveness of coverage of different types of unqualified misoperation interference of disconnecting switches, avoid the analysis bias caused by relying on a single parameter in the existing technology, and avoid the problems of unclear optimization direction and waste of resources caused by the single analysis dimension, thus improving the overall reliability of disconnecting switch data fusion.

[0013] 3. By determining whether the interference data from disconnector switch maloperation meets the acceptable conditions for disconnector switch maloperation interference, if it does not meet the conditions, feedback on maloperation and accuracy is provided. If it does meet the conditions, the corresponding acceptable visual recognition data and acceptable limit feature data are obtained, and the accuracy of the fused data is analyzed. This helps to directly eliminate the influence of unacceptable interference data on the core data (visual recognition data and limit feature data) by pre-judging the acceptableness of the disconnector switch maloperation interference data, ensuring that the marked acceptable visual recognition data and acceptable limit feature data have a reliable analytical basis, avoiding interference data from polluting the subsequent fusion process, and thus helping to reduce the probability of data fusion anomalies from the source.

[0014] 4. First, determine whether the monitored disconnector switch maloperation interference data meets the qualified conditions for disconnector switch maloperation interference. If it does not meet the qualified conditions, provide feedback on the maloperation and accuracy. If it meets the conditions, continue monitoring the number of maloperations. If the number of monitored maloperations exceeds the preset maximum total number of maloperations, send a maintenance prompt to the preset personnel. Otherwise, perform data accuracy analysis based on visual recognition data and limit switch feature data fusion. Since the system may experience faults that indirectly cause the monitored disconnector switch maloperation interference data to be unqualified, precise processing and early warning based on different scenarios help avoid the lag of traditional operation and maintenance, shorten the fault diagnosis time, improve operation and maintenance efficiency, ensure that the visual recognition and limit switch feature data used for fusion analysis have high reliability, and avoid the waste of efficiency caused by indiscriminate processing. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of the method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion provided in this embodiment of the invention;

[0017] Figure 2 This is a general overview diagram of the multi-source data fusion-based risk assessment method for disconnector malfunction provided in this embodiment of the invention;

[0018] Figure 3 This is a schematic diagram of the error and accuracy feedback process of the multi-source data fusion disconnector malfunction risk assessment method provided in this embodiment of the invention;

[0019] Figure 4 This is a schematic diagram of the structure of the multi-source data fusion-based disconnector malfunction risk assessment system provided in this embodiment of the invention;

[0020] Figure 5 This application provides an operational risk interface 1 for its embodiments.

[0021] Figure 6 The second type of operational risk interface provided in the embodiments of this application. Detailed Implementation

[0022] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0023] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0024] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0025] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0026] Example 1: This embodiment of the invention provides a method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion. For example... Figure 1 The flowchart shown is for a multi-source data fusion-based risk assessment method for disconnector malfunctions. The processing flow of this method may include the following steps:

[0027] Disconnect switch maloperation interference feedback monitoring: During multi-source data acquisition, disconnect switch maloperation interference feedback is performed, and the feedback results are output to reflect the interference of disconnect switch maloperation on the data fusion process. Based on the disconnect switch maloperation interference feedback results, it is determined whether to perform maloperation and accuracy feedback. The maloperation and accuracy feedback is used to assess the impact of the disconnect switch maloperation interference feedback results on the qualification of visual recognition data and limit feature data acquisition. By monitoring disconnect switch maloperation interference feedback, it is helpful to promptly capture the interference of disconnect switch maloperation on the multi-source data fusion process, clarify the interference effect, and provide a basis for determining whether to initiate maloperation and accuracy assessment in the subsequent process, ensuring that interference in the early stage of data fusion is perceptible and quantifiable.

[0028] Misoperation and accuracy monitoring: If misoperation and accuracy feedback is performed, the accuracy feedback results, which reflect the accuracy of visual recognition and limit feature acquisition, are used to improve the anti-interference ability of visual recognition and limit feature acquisition against misoperation of disconnecting switches. Through misoperation and accuracy monitoring, the anti-interference ability of visual recognition and limit feature acquisition against misoperation of disconnecting switches is improved, ensuring the qualification of core acquisition data.

[0029] Anomaly superposition interference optimization judgment: If no error operation and accuracy feedback are performed, the accuracy of the fused data is analyzed. Based on the output of the fused data accuracy analysis results, which reflects the qualified status of the data fusion, it is determined whether to perform anomaly superposition interference optimization to reduce the input noise of visual recognition data and limit feature data. The anomaly superposition interference optimization judgment helps to accurately determine whether anomaly superposition interference optimization needs to be carried out, thereby effectively reducing the input noise of visual recognition and limit feature data and ensuring the stability of data fusion.

[0030] like Figure 2 The diagram shown is a general overview of the multi-source data fusion-based risk assessment method for disconnector malfunctions provided in this application embodiment. Figure 2 It can be seen that: when the monitored disconnector switch maloperation interference data meets the qualified conditions for disconnector switch maloperation interference, the accuracy analysis of the fused data is performed; otherwise, maloperation and accuracy feedback is performed to obtain the unqualified opening maloperation interference factor and the unqualified closing maloperation interference factor. Based on the unqualified opening maloperation interference value and the unqualified closing maloperation interference value, accuracy feedback and optimization are performed. After the accuracy feedback and optimization are completed, the accuracy analysis of the fused data is performed to obtain the fused noise anomaly value. When the fused noise anomaly value is greater than the preset fused noise anomaly value, the abnormal superposition interference optimization is performed; otherwise, the corresponding disconnector switch maloperation interference data is marked as qualified risk assessment management data, and the behavior risk level is output based on the qualified risk assessment management data.

[0031] Before designing the multi-source data fusion disconnector maloperation risk assessment method provided in this application, a database storing various settings data was established. The database includes, but is not limited to, preset opening maloperation interference values, preset closing maloperation interference values, preset maximum total number of maloperations, etc., and the various values ​​are directly set by technical personnel.

[0032] In this embodiment, the combined effects of disconnector switch malfunction interference feedback monitoring, malfunction and accuracy monitoring, and abnormal superposition interference optimization judgment are interconnected, which helps to achieve full-process multi-source data acquisition and fusion guarantee from interference monitoring to accuracy optimization and noise control. This comprehensively improves the reliability and anti-interference capability of disconnector switch related monitoring data, and provides high-quality data support for accurate judgment of disconnector switch operating status.

[0033] Furthermore, the specific process for feedback on disconnector switch maloperation interference is as follows: Obtain disconnector switch maloperation interference data reflecting the interference level of disconnector switch maloperation within a specified maloperation monitoring period; determine whether the disconnector switch maloperation interference data meets the qualified conditions for disconnector switch maloperation interference: if not, mark the corresponding disconnector switch maloperation interference data as unqualified disconnector switch maloperation interference data and perform maloperation and accuracy feedback; if compliant, mark the visual recognition data corresponding to the disconnector switch maloperation interference data that meets the qualified conditions as qualified visual recognition data, mark the corresponding limit feature data as qualified limit feature data, and perform fusion data accuracy analysis; unqualified disconnector switch maloperation interference data includes unqualified opening maloperation interference values ​​and unqualified closing maloperation interference values; disconnector switch maloperation interference... The interference qualification condition indicates that the tripping misoperation interference value is less than the preset tripping misoperation interference value, and the closing misoperation interference value is less than the preset closing misoperation interference value. The preset tripping misoperation interference value is represented by the average of the tripping misoperation interference values ​​over a historical time period, and the preset closing misoperation interference value is represented by the average of the closing misoperation interference values ​​over a historical time period. The specified misoperation monitoring time period refers to the preset time period corresponding to the feedback of disconnecting switch misoperation interference. The disconnecting switch misoperation interference data includes tripping misoperation interference values ​​reflecting the degree of deviation in tripping completion and closing misoperation interference values ​​reflecting the degree of deviation in closing completion. The tripping time of the disconnecting switch is monitored by a timer, and the difference between this timer and the preset tripping time is used as the tripping misoperation interference value. The closing time of the disconnecting switch is monitored by a timer, and the difference between this timer and the preset closing time is used as the closing misoperation interference value.

[0034] In this embodiment, by marking the visual recognition data corresponding to disconnector maloperation interference data that meets the qualified conditions for disconnector maloperation interference as qualified visual recognition data, and marking the corresponding limit feature data as qualified limit feature data, and performing data fusion accuracy analysis, it is helpful to accurately screen out high-quality collected data (qualified visual recognition data and qualified limit feature data) that are less affected by maloperation interference. This provides a reliable data foundation for subsequent data fusion. At the same time, the fusion accuracy analysis verifies the fusion effectiveness of qualified data, ensuring the credibility of the data fusion results. When the disconnector maloperation interference data does not meet the qualified conditions for disconnector maloperation interference, feedback on maloperation and accuracy is provided. This helps to promptly locate the unqualified collection problems of visual recognition data and limit feature data caused by maloperation interference, providing accurate feedback basis for subsequent targeted optimization of the collection mechanism and improvement of data collection anti-interference capability.

[0035] Furthermore, the error operation and accuracy feedback includes the assessment of the interference level of disconnecting switch errors and accuracy feedback and optimization. The assessment of the interference level of disconnecting switch errors is used to evaluate the impact of unqualified disconnecting switch error interference data on accuracy feedback and optimization. The specific process is as follows: unqualified disconnecting switch error interference data is input into the error interference level mapping table, and the corresponding error interference level factor is read. The error interference level factor includes unqualified opening error interference factor and unqualified closing error interference factor, which reflect the impact of unqualified disconnecting switch error interference data on accuracy feedback and optimization. Based on the error interference level factor, the following judgment is made: when the unqualified opening error interference factor is greater than the unqualified closing error interference factor, accuracy feedback and optimization are performed based on the unqualified opening error interference value; when the unqualified opening error interference factor is equal to the unqualified closing error interference factor, accuracy feedback and optimization are performed based on the unqualified disconnecting switch error interference data; when the unqualified opening error interference factor is less than the unqualified closing error interference factor, accuracy feedback and optimization are performed based on the unqualified closing error interference value.

[0036] like Figure 3 The diagram shown illustrates the error and accuracy feedback process of the multi-source data fusion-based disconnector malfunction risk assessment method provided in this application embodiment. Figure 3It can be seen that: by inputting the interference data of the unqualified disconnector switch maloperation into the maloperation interference level mapping table, the interference factors of the unqualified opening and closing maloperation are obtained. When the interference factor of the unqualified opening maloperation is greater than the interference factor of the unqualified closing maloperation, the unqualified opening maloperation interference value is used for accuracy feedback and optimization. When the interference factor of the unqualified opening maloperation is equal to the interference factor of the unqualified closing maloperation, the unqualified disconnector switch maloperation interference data is used for accuracy feedback and optimization. When the interference factor of the unqualified opening maloperation is less than the interference factor of the unqualified closing maloperation, the unqualified closing maloperation interference value is used for accuracy feedback and optimization.

[0037] In the embodiments of this application, several sets of mapping tables and mapping sets pre-constructed by designated personnel are retrieved from the database, including a mapping table for the degree of maloperation interference, an exposure time adjustment mapping table, an aperture adjustment mapping table, a sampling window length mapping table, a mapping table for the accuracy of fused data, a photosensitivity mapping set, a sensor gain mapping set, etc. Their mapping relationships are dynamic, supporting both one-to-one correspondence between single and single parameters and many-to-one correspondence between many and single parameters. For example, by inputting the maloperation interference data of unqualified disconnect switches, the performance reflection value of metering pumps, the visual recognition accuracy feedback value and the combination of the exposure frequency of the maloperation behavior camera, the combination of the visual recognition feedback deviation value and the camera focal length, the combination of the limit feature feedback deviation value and the number of sampling points of the opening and closing signal, the qualified fusion accuracy result, the combination of the fusion noise anomaly value and the ambient brightness, and the combination of the fusion noise anomaly value and the number of pixels of the maloperation behavior image into a machine learning model based on feature importance, such as a decision tree model, the corresponding weights or data, i.e., the maloperation interference degree factor, the exposure time adjustment value, etc., are obtained through feature splitting, and the mapping tables and mapping sets are obtained by correspondingly matching the data of the historical time period with the corresponding weights or data.

[0038] Specifically, the maloperation interference level factor provided in this embodiment is determined based on the proportion of maloperation interference data of the corresponding unqualified disconnecting switches. It is used to reflect the degree of influence of the maloperation interference data of unqualified disconnecting switches on accuracy feedback and optimization. A one-to-one or many-to-one mapping relationship can be established between the maloperation interference data of unqualified disconnecting switches and the corresponding maloperation interference level factor. By inputting the maloperation interference data of unqualified disconnecting switches acquired in real time into the corresponding maloperation interference level mapping table, the corresponding maloperation interference level factor is output according to the preset mapping relationship, and the value range is limited to the interval between 0 and 1.

[0039] Specifically, the accuracy feedback and optimization process is as follows: When the interference factor of the unqualified tripping misoperation is detected to be greater than that of the unqualified closing misoperation, the result of harmonic averaging of the unqualified tripping misoperation interference value (used to reflect the degree of interference of the tripping misoperation) and the visual recognition accuracy value (used to measure the degree of interference with visual recognition accuracy) is used as the visual recognition accuracy feedback value. Similarly, the result of harmonic averaging of the unqualified tripping misoperation interference value and the limit feature accuracy value (used to measure the accuracy of the limit feature) is used as the limit feature accuracy feedback value. The larger the unqualified tripping misoperation interference value, the more severe the tripping interference, and the higher the impact of tripping interference on visual recognition accuracy and limit feature accuracy. By harmonic averaging the unqualified tripping misoperation interference value (used to reflect the degree of interference of the tripping misoperation) and the visual recognition accuracy value (used to measure the degree of interference with visual recognition accuracy), the result of harmonic averaging of the unqualified tripping misoperation interference value and the visual recognition accuracy value (used to measure the degree of interference with visual recognition accuracy) is used as the limit feature accuracy feedback value. Harmonic averaging of the limit feature accuracy values, which measures the accuracy of the limit features, helps to target strong interference scenarios of misoperation during circuit breaker tripping. This makes the accuracy feedback values ​​of visual recognition and limit features more closely match the actual impact of tripping interference on data acquisition, providing a precise quantitative basis for subsequent optimization of camera imaging qualification and reduction of limit feature delay deviation during tripping. By monitoring the imaging duration of misoperation behavior images through a timer, the difference between this and the preset imaging duration of misoperation behavior images is used as the visual recognition accuracy value. This reflects the qualification level of the images captured by the camera when acquiring images of the disconnector switch tripping or closing process. The larger the visual recognition accuracy value, the greater the interference on the accuracy of the images captured by the camera when acquiring images of the disconnector switch tripping or closing process. By monitoring the time difference between the limit feature being triggered and the detection of the trigger signal and the response through a timer, the limit feature accuracy value is used to reflect the limit feature delay deviation.

[0040] When the interference factor of the unqualified tripping misoperation is equal to that of the unqualified closing misoperation, the result of harmonic averaging the unqualified tripping misoperation interference value, the unqualified closing misoperation interference value, and the visual recognition accuracy value is used as the visual recognition accuracy feedback value. Similarly, the result of harmonic averaging the unqualified tripping misoperation interference value, the unqualified closing misoperation interference value, and the limit feature accuracy value is used as the limit feature accuracy feedback value. Furthermore, the harmonic averaging of the unqualified tripping misoperation interference value and the unqualified closing misoperation interference value indicates that tripping interference and closing interference are similar, and the visual recognition accuracy and limit feature accuracy are affected by both tripping and closing interference. For similar degrees, the larger the interference values ​​of unqualified tripping and closing misoperations, the greater the potential increase in both visual recognition accuracy feedback values ​​and feature accuracy feedback values. By harmonizing and averaging the interference values ​​of unqualified tripping and closing misoperations with the limit feature accuracy values, it is possible to balance the interference scenarios of tripping and closing misoperations. This allows the accuracy feedback values ​​to fully reflect the combined impact of the two types of misoperation interference on visual recognition and limit feature acquisition, providing comprehensive quantitative support for simultaneously optimizing the data acquisition quality of the tripping and closing processes.

[0041] When the interference factor of the unqualified tripping misoperation is less than that of the unqualified closing misoperation, the result of harmonic averaging the unqualified closing misoperation interference value (which reflects the degree of interference in closing misoperation) and the visual recognition accuracy value is used as the visual recognition accuracy feedback value. Similarly, the result of harmonic averaging the unqualified closing misoperation interference value and the limit feature accuracy value is used as the limit feature accuracy feedback value. This helps to accurately focus on strong interference scenarios of closing misoperation, ensuring that the accuracy feedback value precisely matches the impact of closing interference on data acquisition. This provides a clear quantitative direction for targeted optimization of camera image qualification during closing and reduction of limit feature delay deviation. The larger the unqualified closing misoperation interference value, the more severe the closing interference, and the higher the impact of closing interference on visual recognition accuracy and limit feature accuracy.

[0042] It should be added that accuracy feedback and optimization also includes judging the accuracy feedback of visual recognition and limit features. The specific process is as follows: Based on the visual recognition accuracy feedback value and the limit feature accuracy feedback value, a judgment is made: If the visual recognition accuracy feedback value is less than the preset visual recognition accuracy feedback value, visual recognition accuracy optimization is performed to improve the anti-interference capability of visual recognition against misoperation of the disconnector switch; otherwise, fusion data accuracy analysis is performed. The preset visual recognition accuracy feedback value is represented by the average value of visual recognition accuracy feedback values ​​over a historical time period. If the limit feature accuracy feedback value is less than the preset limit feature accuracy feedback value, limit feature accuracy optimization is performed to improve the timing accuracy of the limit signal; otherwise, fusion data accuracy analysis is performed. The preset limit feature accuracy feedback value is represented by the average value of limit feature accuracy feedback values ​​over a historical time period.

[0043] In this embodiment, based on the visual recognition accuracy feedback value and limit feature accuracy feedback value obtained from the assessment of the degree of interference from disconnector switch maloperation, specific accuracy feedback and optimization measures are determined. This helps to transform the quantitative interference assessment results into targeted optimization actions, avoid blind optimization measures, and efficiently improve the resistance of the two types of data collection to maloperation interference. Through the interaction between the disconnector switch maloperation interference assessment and accuracy feedback and optimization, a closed-loop management system is formed from quantitative interference assessment to precise optimization and improvement. The disconnector switch maloperation interference assessment provides a clear problem orientation and quantitative basis for accuracy feedback and optimization, while accuracy feedback and optimization address the shortcomings found in the assessment. The two work together to ensure the continuous qualification of visual recognition and limit feature data collection, thereby stabilizing the quality of multi-source data fusion and laying a solid data foundation for the accurate monitoring of the disconnector switch's operating status.

[0044] Further, the specific process for optimizing visual recognition accuracy is as follows: The visual recognition accuracy feedback value and the camera exposure frequency corresponding to the erroneous behavior directly read by the camera are input into the exposure time adjustment mapping table, and the exposure time adjustment value is read; the exposure time is gradually reduced by using the amplitude corresponding to the exposure time adjustment value as the adjustment step size; Limit feature accuracy optimization specifically involves: filtering the limit feature data based on a moving average filtering algorithm to suppress high-frequency noise and instantaneous hysteresis in the sensor signal, improving the timing accuracy and stability of the limit signal; Visual recognition accuracy optimization and limit feature accuracy optimization also include performing visual recognition-limit feature qualification judgment, the specific process of which is as follows: After the visual recognition accuracy optimization or limit feature accuracy optimization is completed, it is determined whether the re-acquired visual recognition accuracy feedback value is still less than the preset visual recognition accuracy feedback value, and whether the limit feature accuracy feedback value is still less than the preset limit feature accuracy feedback value. If the visual recognition accuracy feedback value is still less than the preset visual recognition accuracy feedback value, a value reflecting the visual recognition accuracy is obtained. The system calculates the visual recognition feedback deviation value to reflect the degree of accuracy deviation and performs deviation-visual recognition accuracy optimization; otherwise, it performs fusion data accuracy analysis. If the limit feature accuracy feedback value is still less than the preset limit feature accuracy feedback value, it obtains the limit feature feedback deviation value to reflect the degree of accuracy deviation and performs deviation-limit feature accuracy optimization; otherwise, it performs fusion data accuracy analysis. Deviation-visual recognition accuracy optimization means that dynamic aperture adjustment provides reliable visual input for multi-source data fusion, thereby improving the accuracy of misoperation risk assessment. Deviation-limit feature accuracy optimization means that dynamic adjustment of the sampling window length suppresses transient noise interference and extracts more stable limit features, thereby improving the accuracy of misoperation risk assessment. The visual recognition feedback deviation value is represented by the difference between the preset visual recognition accuracy feedback value and the visual recognition accuracy feedback value obtained after visual recognition accuracy optimization. The limit feature feedback deviation value is represented by the difference between the preset limit feature accuracy feedback value and the limit feature accuracy feedback value obtained after limit feature accuracy optimization.

[0045] It should be added that the combination of visual recognition accuracy feedback value and exposure frequency of camera for erroneous behavior provided in this embodiment establishes a one-to-one or many-to-one mapping relationship with the corresponding exposure time adjustment value; by inputting the combination of visual recognition accuracy feedback value and exposure frequency of camera for erroneous behavior collected in real time into the corresponding exposure time adjustment mapping table, the corresponding exposure time adjustment value is output according to the preset mapping relationship, and the value range is limited to the 0-1 interval.

[0046] In this embodiment, by using the magnitude corresponding to the exposure time adjustment value as the adjustment step size and gradually reducing the exposure time, it helps to achieve stable and controllable optimization of camera imaging parameters, avoids fluctuations in image quality caused by sudden changes in exposure time, and accurately improves the imaging qualification of visual recognition data under the interference of misoperation (such as reducing overexposure and improving image clarity), providing a stable image foundation for subsequent visual recognition accuracy. By performing visual recognition-limit feature qualification judgment to determine whether to continue deviation-visual recognition accuracy optimization and deviation-limit feature accuracy optimization, it avoids over-optimization of qualified data (such as causing resource waste or new deviations), and only performs precise optimization on unqualified data, ensuring the necessity and efficiency of optimization measures. By using the visual recognition feedback deviation value and the limit feature feedback deviation value as feedback values ​​for deviation-visual recognition accuracy optimization and deviation-limit feature accuracy optimization, it helps to accurately locate the degree of problems in visual recognition and limit features, allowing optimization actions to target the root cause of deviations and efficiently improve the accuracy and anti-interference ability of the two types of data acquisition.

[0047] Furthermore, the specific process of deviation-visual recognition accuracy optimization is as follows: The visual recognition feedback deviation value and the camera focal length directly read from the camera are input into the aperture adjustment mapping table to read the aperture adjustment value; the aperture is gradually reduced step by step, using the magnitude corresponding to the aperture adjustment value as the adjustment step size; if the visual recognition accuracy feedback value obtained after deviation-visual recognition accuracy optimization is not less than the preset visual recognition accuracy feedback value, then fusion data accuracy analysis is performed; otherwise, a visual recognition anomaly alarm is sent. By gradually reducing the aperture step by step, using the magnitude corresponding to the aperture adjustment value as the adjustment step size, it helps to smoothly and controllably optimize the camera's optical parameters, avoiding drastic fluctuations in image depth or light intake caused by sudden aperture changes, and accurately improving the clarity of visual recognition images under erroneous operation interference (such as enhancing image details and reducing stray light interference), providing a high-quality imaging foundation for visual recognition accuracy optimization.

[0048] The specific process for optimizing the accuracy of the deviation-limit feature is as follows: The deviation value of the limit feature feedback and the number of sampling points for the opening and closing signals monitored by the oscilloscope and spectrum analyzer are input into the sampling window length mapping table to read the sampling window length adjustment value; the sampling window length is gradually increased step by step, using the amplitude corresponding to the sampling window length adjustment value as the adjustment step size; if the re-acquired limit feature accuracy feedback value is not less than the preset limit feature accuracy feedback value after the deviation-limit feature accuracy optimization is completed, then the fusion data accuracy analysis is performed; otherwise, a limit feature abnormality alarm is sent. By gradually increasing the sampling window length step by step, using the amplitude corresponding to the sampling window length adjustment value as the adjustment step size, it helps to smoothly expand the sampling range of the limit feature data, reduce the impact of instantaneous noise on the sampling results, more fully capture the effective signal of the limit feature, improve the stability and accuracy of limit feature data acquisition, and provide reliable data support for subsequent limit feature accuracy optimization.

[0049] Specifically, the process for analyzing the accuracy of fused data is as follows: Obtain fusion noise anomaly values ​​to reflect the degree of noise anomalies during the fusion of visual recognition data and limit feature data; make judgments based on the fusion noise anomaly values: if the fusion noise anomaly value is greater than the preset fusion noise anomaly value, perform anomaly superposition interference optimization; otherwise, mark the corresponding disconnector switch misoperation interference data as qualified risk assessment management data; the qualified risk assessment management data is used to input into the preset risk assessment model to obtain the disconnector switch misoperation analysis output behavior risk level. For example, inputting the qualified risk assessment management data into the ST-GCN model outputs the behavior risk level; monitor the noise intensity of the corresponding fused signal during the fusion of visual recognition data and limit feature data using a spectrum analyzer and a sound level meter, and use the difference between this and the preset fusion noise intensity as the fusion noise anomaly value.

[0050] It should be added that the combination of visual recognition feedback deviation value and camera focal length, as well as the combination of limit feature feedback deviation value and the number of sampling points for opening and closing signals provided in this embodiment, establish a one-to-one or many-to-one mapping relationship with the corresponding aperture adjustment value and sampling window length adjustment value. By inputting the combination of visual recognition feedback deviation value and camera focal length, as well as the combination of limit feature feedback deviation value and the number of sampling points for opening and closing signals, which are collected in real time, into the corresponding aperture adjustment mapping table and sampling window length mapping table, the corresponding aperture adjustment value and sampling window length adjustment value are output according to the preset mapping relationship, and the value range is limited to the interval between 0 and 1.

[0051] In this embodiment, if the visual recognition accuracy feedback value obtained after deviation-visual recognition accuracy optimization and deviation-limit feature accuracy optimization is not less than the preset visual recognition accuracy feedback value, or the limit feature accuracy feedback value is not less than the preset limit feature accuracy feedback value, then the accuracy analysis of the fused data is performed to obtain qualified risk assessment management data. This helps to verify the actual effect of deviation-visual recognition accuracy optimization and deviation-limit feature accuracy optimization, and further evaluates the qualified risk of the qualified visual recognition data and limit feature data in the fusion process, identifies potential problems in the fusion process in advance, provides a quantitative basis for risk control of multi-source data fusion quality, and ensures the reliability of the final fusion result.

[0052] Furthermore, the specific process for optimizing abnormal superposition interference is as follows: The qualified fusion accuracy result is input into the fusion data accuracy mapping table to obtain a fusion-accuracy factor that reflects the impact of disconnector switch maloperation on the fusion process of visual recognition data and limit feature data. The qualified fusion accuracy result includes disconnector switch maloperation interference data and fusion noise anomalies. The fusion-accuracy factor includes a fusion-visual recognition accuracy factor and a fusion-limit feature accuracy factor. Based on the fusion-accuracy factor, a judgment is made: if the fusion-visual recognition accuracy factor is greater than the fusion-limit feature accuracy factor, only fusion-visual recognition optimization to reduce the abnormal superposition of noise in the visual recognition data is performed. This helps to target and focus on the interference problem of disconnector switch maloperation in the visual recognition data, concentrate resources to reduce its abnormal noise superposition, avoid over-optimization of the limit feature data with less interference impact, and efficiently improve... To enhance the anti-interference capability of visual recognition data during the fusion process; if the fusion-visual recognition degree factor is less than the fusion-limit feature degree factor, only the fusion-limit feature optimization, which is used to reduce the noise abnormal superposition of limit feature data, is performed. This helps to accurately locate the weakness of limit feature data affected by the misoperation of disconnecting switches, and specifically reduce its noise abnormal superposition, avoiding wasting resources on visual recognition data with less interference, and effectively improving the stability of limit feature data during the fusion process; if the fusion-visual recognition degree factor is equal to the fusion-limit feature degree factor, both fusion-visual recognition optimization and fusion-limit feature optimization are performed simultaneously. This helps to balance the interference problem of visual recognition data and limit feature data affected by the misoperation of disconnecting switches, and simultaneously reduce the noise abnormal superposition of the two types of data, ensuring that the anti-interference capability of both is improved in sync during the fusion process, and guaranteeing the balance and reliability of the overall fused data quality.

[0053] Specifically, the fusion-accuracy factor provided in this embodiment is determined based on the proportion of the corresponding qualified fusion accuracy result. It is used to reflect the degree of influence of the qualified fusion accuracy result on the fusion process of visual recognition data and limit feature data due to the misoperation of the disconnecting switch. A one-to-one or many-to-one mapping relationship can be established between the qualified fusion accuracy result and the corresponding fusion-accuracy factor. By inputting the qualified fusion accuracy result obtained in real time into the corresponding fusion data accuracy mapping table, the corresponding fusion-accuracy factor is output according to the preset mapping relationship, and the value range is limited to the interval between 0 and 1.

[0054] The fusion-visual recognition optimization specifically involves: obtaining a sensitivity adjustment value by mapping the fusion noise anomaly value and the ambient brightness monitored by the illuminance meter into the sensitivity input. The corresponding amplitude is used as the adjustment step size to gradually reduce the sensitivity. This helps to smoothly and accurately reduce the camera sensitivity, avoid image noise superposition caused by high sensitivity, and adapt to ambient light conditions, improving the cleanliness of the visual recognition image and reducing noise interference for subsequent data fusion. If the fusion noise anomaly value is still not greater than the preset fusion noise anomaly value after the fusion-visual recognition optimization is completed, the corresponding disconnect switch misoperation interference data is marked as qualified risk assessment management data; otherwise, a fusion-visual recognition optimization anomaly prompt is sent.

[0055] Specifically, this embodiment provides a one-to-one or many-to-one mapping relationship between the combination of fused noise anomalies and ambient brightness, and the combination of fused noise anomalies and the number of pixels in the erroneous behavior image, and the corresponding photosensitivity adjustment value and sensor gain adjustment value. By inputting the real-time acquired combination of fused noise anomalies and ambient brightness, and the combination of fused noise anomalies and the number of pixels in the erroneous behavior image, into the corresponding photosensitivity mapping set and sensor gain mapping set, the corresponding photosensitivity adjustment value and sensor gain adjustment value are output according to the preset mapping relationship, and the value range is limited to the 0-1 interval.

[0056] The fusion-limit feature optimization specifically involves inputting the fusion noise anomaly value and the number of pixels in the maloperation image monitored by image analysis software such as Adobe Photoshop into the sensor gain mapping set to obtain the sensor gain adjustment value. The corresponding amplitude is used as the adjustment step size to gradually increase the sensor gain, which helps to smoothly amplify the effective signal (such as the detailed signal of maloperation behavior). While ensuring the clarity of the pixel features of the maloperation image, it also combats fusion noise interference, improves the accuracy of visual recognition data in capturing maloperation behavior, and provides a more reliable image signal foundation for data fusion. When the fusion noise anomaly value is still not greater than the preset fusion noise anomaly value after the fusion-limit feature optimization is completed, the corresponding disconnect switch maloperation interference data is marked as qualified risk assessment management data; otherwise, a fusion-limit feature optimization anomaly prompt is sent.

[0057] like Figure 4 The diagram shown is a structural schematic of the multi-source data fusion disconnector malfunction risk assessment system provided in this application embodiment. It applies a multi-source data fusion disconnector malfunction risk assessment method, characterized by including: a disconnector malfunction interference feedback monitoring module, a malfunction and accuracy monitoring module, and an abnormal superposition interference optimization judgment module. The disconnector malfunction interference feedback monitoring module is used to provide disconnector malfunction interference feedback during multi-source data acquisition and output a disconnector malfunction interference feedback result reflecting the interference of disconnector malfunction on the data fusion process. Based on the disconnector malfunction interference feedback result, it determines whether to perform malfunction and accuracy feedback. The malfunction and accuracy monitoring module is used to perform accuracy feedback and optimization based on the malfunction and accuracy feedback results reflecting visual recognition accuracy and limit feature accuracy, if malfunction and accuracy feedback is performed, based on the malfunction and accuracy feedback results. The abnormal superposition interference optimization judgment module is used to perform fusion data accuracy analysis if malfunction and accuracy feedback is not performed, and based on the output fusion data accuracy analysis results reflecting the data fusion qualification status, it determines whether to perform abnormal superposition interference optimization to reduce input noise in visual recognition data and limit feature data.

[0058] In this embodiment, a judgment is first made based on the fusion-visual recognition degree factor and the fusion-limiting feature degree factor, and then the specific measures for fusion-visual recognition optimization and fusion-limiting feature optimization are determined. The two are interconnected, which helps to accurately locate the main interference shortcomings of visual recognition data and limiting feature data in the fusion process based on the quantified fusion-visual recognition degree factor and fusion-limiting feature degree factor, avoid the waste of resources caused by indiscriminate optimization, and ensure that the optimization action is strongly correlated with the actual interference situation, effectively reduce the abnormal superposition of noise between the two types of data, and ultimately ensure the accuracy and stability of the multi-source data fusion results.

[0059] like Figure 5 As shown, this is an example of an operational risk interface provided in this application embodiment. Figure 6 The image shown is a second operational risk interface provided in this application embodiment. This application is used to implement the operational risk function in the risk type management module within the power risk management system. The power risk management system includes a risk type management module, a risk identification module, a risk assessment module, and a system management module. The risk type management module includes equipment risks, operational risks, power grid operation risks, and environmental and external risks. The risk identification module includes risk data collection and automatic risk identification. The risk assessment module includes risk level assessment, risk trend prediction, and risk alarm. The system management module includes user and permission management, system configuration, and database.

[0060] Depend on Figure 5 It is known that the currently monitored power equipment is a disconnect switch (number 17). The current error operation feedback progress is 50%. Both the visual recognition status light and the limit feature status light are red, indicating an abnormal status. Clicking "Modify Configuration" will take you to the second operation-related risk interface. Figure 6 As can be seen, the current optimization progress is 50%, and the progress of error feedback has been updated to 80%.

[0061] Example 2: Based on Example 1, the specific process for feedback on disconnector switch maloperation interference is as follows: If the monitored disconnector switch maloperation interference data does not meet the qualified conditions for disconnector switch maloperation interference, feedback on maloperation and accuracy is provided; if it does meet the conditions, it is determined whether the number of monitored maloperations is greater than the preset maximum total number of maloperations. If so, a maintenance prompt is sent to the preset personnel; otherwise, a data accuracy analysis is performed based on visual recognition data and limit feature data. The total number of disconnector switch opening and closing operations within a specified maloperation monitoring period is monitored by a counter as the number of maloperations, and the preset maximum total number of maloperations is represented by the maximum value of the number of maloperations in the historical time period.

[0062] In this embodiment, Example 2 adds a secondary judgment on the number of misoperations. Even if the interference data meets the qualified conditions, if the number of misoperations exceeds the preset maximum total number of misoperations, additional processing is triggered. This avoids the potential risk of high-frequency misoperations despite a single interference data being qualified (such as hidden problems such as stuck operating mechanism or slow sensing of the disconnecting switch; frequent occurrences of interference even if a single interference does not exceed the standard will accumulate risks). By sending maintenance prompts to preset personnel, high-frequency misoperations are often not simply data acquisition problems, but may point to mechanical faults in the disconnecting switch itself (such as abnormal opening and closing mechanism) or loopholes in the operating procedure. Sending maintenance prompts in advance can encourage maintenance personnel to promptly investigate potential problems in the hardware operation of the equipment, avoid the accumulation of faults due to ignoring high-frequency misoperations, and thus reduce subsequent more serious operational accidents (such as line faults caused by misoperations).

[0063] In summary, this application embodiment, by providing feedback on interference from disconnector switch maloperation and determining whether to provide feedback on maloperation and accuracy, helps to accurately locate the interference points and degree of impact of disconnector switch maloperation on the data fusion process. This provides a clear judgment benchmark for subsequent targeted handling of data fusion anomalies, avoiding resource waste caused by blind optimization. If feedback on maloperation and accuracy is provided, then accuracy feedback and optimization are performed, which helps to improve the accuracy of visual recognition and the reliability of limit feature judgment during disconnector switch data fusion. If no feedback on maloperation and accuracy is provided, then the accuracy of the fused data is analyzed, and based on the output of the fused data accuracy analysis results, it is determined whether to perform abnormal superposition interference optimization. This helps to promptly identify and handle possible abnormal superposition interference problems during the data fusion process, providing reliable data support for disconnector switch operation status monitoring and maintenance decisions. In turn, it helps to improve the accuracy of disconnector switch maloperation risk assessment and management based on multi-source data fusion, solving the problem of low accuracy in the existing technology for multi-source data fusion disconnector switch maloperation risk assessment and management.

[0064] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0065] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0066] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0067] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for assessing the risk of misoperation of disconnect switches based on multi-source data fusion, characterized in that, The method includes: During the multi-source data acquisition process, feedback on interference from disconnect switch malfunction is performed, and the feedback result of disconnect switch malfunction is output to reflect the interference of disconnect switch malfunction on the data fusion process, so as to determine whether to perform malfunction and accuracy feedback. If error and accuracy feedback is provided, then accuracy feedback and optimization are performed based on the error and accuracy feedback results that reflect the accuracy of visual recognition and the accuracy of limit features. If no error or accuracy feedback is provided, then the accuracy of the fused data is analyzed. Based on the output of the fused data accuracy analysis results, which reflects the qualification of the data fusion, it is determined whether to perform abnormal superposition interference optimization to reduce the input noise of visual recognition data and limit feature data. The specific process for feedback on interference caused by misoperation of the disconnecting switch is as follows: Acquire interference data of disconnecting switch maloperation, wherein the interference data of disconnecting switch maloperation includes interference values ​​of opening maloperation and closing maloperation; Determine whether the interference data from disconnector switch malfunction meets the acceptable criteria for disconnector switch malfunction interference: If it does not meet the requirements, the corresponding disconnect switch misoperation interference data will be marked as unqualified disconnect switch misoperation interference data, and feedback on misoperation and accuracy will be provided. If it meets the requirements, the visual recognition data corresponding to the disconnect switch misoperation interference data that meets the qualified conditions for disconnect switch misoperation interference will be marked as qualified visual recognition data, the corresponding limit feature data will be marked as qualified limit feature data, and the accuracy analysis of the fused data will be performed. The interference data of the unqualified disconnecting switch misoperation includes the interference value of unqualified opening misoperation and the interference value of unqualified closing misoperation; The qualified condition for interference of disconnecting switch misoperation means that the interference value of opening misoperation is less than the preset interference value of opening misoperation, and the interference value of closing misoperation is less than the preset interference value of closing misoperation. The specific process of accuracy feedback and optimization is as follows: Input the unqualified tripping misoperation interference value and the unqualified closing misoperation interference value into the pre-constructed misoperation interference degree mapping table, and read the corresponding unqualified tripping misoperation interference factor and unqualified closing misoperation interference factor. Based on the magnitude relationship between the interference factors of the unqualified opening and closing operations, the interference values ​​of the unqualified opening and closing operations are selectively harmonic averaged with the visual recognition accuracy value and the limit feature accuracy value to obtain the visual recognition accuracy feedback value and the limit feature accuracy feedback value. The accuracy feedback and optimization also includes performing visual recognition-limit feature accuracy feedback judgment, the specific process of which is as follows: Judgment is made based on the accuracy feedback values ​​of visual recognition and the accuracy feedback values ​​of positioning features: If the visual recognition accuracy feedback value is less than the preset visual recognition accuracy feedback value, visual recognition accuracy optimization is performed; otherwise, data fusion accuracy analysis is performed. If the feedback value of the limit feature accuracy is less than the preset limit feature accuracy feedback value, the limit feature accuracy is optimized; otherwise, the accuracy of the fused data is analyzed. The specific process for performing accuracy analysis of the fused data is as follows: Obtain fusion noise anomalies; If the fusion noise anomaly value is greater than the preset fusion noise anomaly value, abnormal superposition interference optimization is performed; otherwise, the corresponding disconnect switch misoperation interference data is marked as qualified risk assessment management data. The fusion noise anomaly value is represented by the difference between the noise intensity of the corresponding fusion signal and the preset fusion noise intensity during the fusion process of visual recognition data and limit feature data. The qualified risk assessment management data is used to input into the preset risk assessment model to obtain the risk level of the behavior output by the analysis of misoperation of the disconnecting switch.

2. The method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion according to claim 1, characterized in that, The specific process for providing feedback on interference from misoperation of the disconnecting switch is as follows: If the monitored data on the interference of disconnecting switch maloperation does not meet the qualified conditions for interference of disconnecting switch maloperation, feedback on maloperation and accuracy shall be provided. If the conditions are met, determine whether the number of monitored misoperations is greater than the preset maximum total number of misoperations. If so, send a maintenance prompt to the preset personnel; otherwise, perform a data accuracy analysis based on the fusion of visual recognition data and limit feature data. The number of misoperations refers to the total number of times the disconnector switch is opened and closed within a specified misoperation monitoring period.

3. The method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion according to claim 1 or 2, characterized in that, The error and accuracy feedback includes the assessment of the interference level of disconnector switch malfunctions and the accuracy feedback and optimization. The assessment of the degree of interference from misoperation of disconnecting switches is used to evaluate the impact of misoperation interference data from unqualified disconnecting switches on accuracy feedback and optimization. The specific process is as follows: Input the interference data of the unqualified disconnect switch misoperation into the misoperation interference level mapping table, and read the corresponding misoperation interference level factor. The misoperation interference level factor includes the unqualified opening misoperation interference factor and the unqualified closing misoperation interference factor. When the interference factor of the unqualified tripping misoperation is detected to be greater than the interference factor of the unqualified closing misoperation, the result of the harmonic averaging of the unqualified tripping misoperation interference value and the visual recognition accuracy value is used as the visual recognition accuracy feedback value, and the result of the harmonic averaging of the unqualified tripping misoperation interference value and the limit feature accuracy value is used as the limit feature accuracy feedback value. When the interference factor of the unqualified tripping operation is equal to the interference factor of the unqualified closing operation, the result of the harmonic averaging of the unqualified tripping operation interference value, the unqualified closing operation interference value and the visual recognition accuracy value is used as the visual recognition accuracy feedback value, and the result of the harmonic averaging of the unqualified tripping operation interference value, the unqualified closing operation interference value and the limit feature accuracy value is used as the limit feature accuracy feedback value. When the interference factor of the unqualified opening misoperation is less than that of the unqualified closing misoperation, the result of harmonic averaging of the unqualified closing misoperation interference value (which reflects the degree of interference of closing misoperation) and the visual recognition accuracy value is used as the visual recognition accuracy feedback value, and the result of harmonic averaging of the unqualified closing misoperation interference value and the limit feature accuracy value is used as the limit feature accuracy feedback value.

4. The method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion according to claim 1, characterized in that, The specific process for optimizing the accuracy of visual recognition is as follows: Input the visual recognition accuracy feedback value and the camera exposure frequency of erroneous behavior into the exposure time adjustment mapping table, and read the exposure time adjustment value; The exposure time is gradually reduced by using the magnitude corresponding to the exposure time adjustment value as the adjustment step size. The accuracy optimization of the limiting feature is specifically achieved by filtering the limiting feature data based on a moving average filtering algorithm to suppress high-frequency noise and instantaneous hysteresis in the sensor signal, thereby improving the timing accuracy and stability of the limiting signal. The optimization of visual recognition accuracy and the optimization of the positioning feature accuracy also include performing a visual recognition-positioning feature qualification judgment, the specific process of which is as follows: After the visual recognition accuracy optimization or limit feature accuracy optimization is completed, it is determined whether the re-acquired visual recognition accuracy feedback value is still less than the preset visual recognition accuracy feedback value, and whether the limit feature accuracy feedback value is still less than the preset limit feature accuracy feedback value. If the visual recognition accuracy feedback value is still less than the preset visual recognition accuracy feedback value, the visual recognition feedback deviation value is obtained and deviation-visual recognition accuracy optimization is performed; otherwise, fusion data accuracy analysis is performed. If the limit feature accuracy feedback value is still less than the preset limit feature accuracy feedback value, the limit feature feedback deviation value is obtained and deviation-limit feature accuracy optimization is performed; otherwise, fusion data accuracy analysis is performed. The aforementioned deviation-visual recognition accuracy optimization means improving the accuracy of error risk assessment by providing reliable visual input for multi-source data fusion through dynamic aperture adjustment; The aforementioned deviation-limit feature accuracy optimization means that by dynamically adjusting the sampling window length to suppress transient noise interference and extract more stable limit features, the accuracy of misoperation risk assessment is improved.

5. The method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion according to claim 4, characterized in that, The specific process for optimizing the accuracy of the deviation-visual recognition is as follows: Input the visual recognition feedback deviation value and the camera focal length into the aperture adjustment mapping table, and read the aperture adjustment value; The aperture is gradually decreased step by step, with the adjustment step size corresponding to the aperture adjustment value. If the visual recognition accuracy feedback value is not less than the preset visual recognition accuracy feedback value after the deviation-visual recognition accuracy optimization is completed, then the fusion data accuracy analysis is performed; otherwise, a visual recognition anomaly alarm is sent. The specific process for optimizing the accuracy of the deviation-limiting feature is as follows: Input the limit feature feedback deviation value and the number of sampling points for the opening and closing signals into the sampling window length mapping table to read the sampling window length adjustment value; The sampling window length is increased step by step, using the amplitude corresponding to the sampling window length adjustment value as the adjustment step size. If the accuracy feedback value of the new limit feature is not less than the preset accuracy feedback value after the deviation-limit feature accuracy optimization is completed, then the accuracy analysis of the fused data will be performed; otherwise, an alarm for abnormal limit feature will be sent.

6. The method for assessing the risk of misoperation of disconnecting switches based on multi-source data fusion according to claim 1, characterized in that, The specific process for optimizing abnormal superposition interference is as follows: The qualified fusion accuracy results are input into the fusion data accuracy mapping table to obtain the fusion-accuracy factor. The qualified fusion accuracy results include disconnect switch misoperation interference data and fusion noise anomalies. The fusion-accuracy factor includes the fusion-visual recognition degree factor and the fusion-limiting feature degree factor. If the fusion-visual recognition degree factor is greater than the fusion-limiting feature degree factor, only the fusion-visual recognition optimization is performed; If the fusion-visual recognition degree factor is less than the fusion-constraint feature degree factor, only the fusion-constraint feature optimization is performed; If the fusion-visual recognition degree factor is equal to the fusion-constraint feature degree factor, perform both fusion-visual recognition optimization and fusion-constraint feature optimization simultaneously. The fusion-visual recognition optimization is specifically as follows: the fusion noise anomaly value and the ambient brightness input photosensitivity are mapped together to obtain the photosensitivity adjustment value. The corresponding amplitude is used as the adjustment step size to gradually reduce the photosensitivity. When the fusion-visual recognition optimization is completed and the fusion noise anomaly value is still not greater than the preset fusion noise anomaly value, the corresponding disconnect switch misoperation interference data is marked as qualified risk assessment and management data. Otherwise, a fusion-visual recognition optimization anomaly prompt is sent. The fusion-limit feature optimization specifically involves: inputting the fusion noise anomaly value and the number of pixels in the malfunction image into the sensor gain mapping set to obtain the sensor gain adjustment value, using its corresponding amplitude as the adjustment step size, and gradually increasing the sensor gain. When the fusion-limit feature optimization is completed and the fusion noise anomaly value is still not greater than the preset fusion noise anomaly value, the corresponding disconnect switch malfunction interference data is marked as qualified risk assessment management data; otherwise, a fusion-limit feature optimization anomaly prompt is sent.

7. A multi-source data fusion-based risk assessment system for disconnecting switch malfunctions, employing the multi-source data fusion-based risk assessment method for disconnecting switch malfunctions as described in any one of claims 1-6, characterized in that... include: The module includes a disconnector switch maloperation interference feedback monitoring module, a maloperation and accuracy monitoring module, and an abnormal superposition interference optimization and judgment module. The disconnector switch maloperation interference feedback monitoring module is used to provide disconnector switch maloperation interference feedback during multi-source data acquisition and output disconnector switch maloperation interference feedback results to reflect the interference of disconnector switch maloperation on the data fusion process, and to determine whether to provide maloperation and accuracy feedback. The error operation and accuracy monitoring module is used to perform accuracy feedback and optimization based on the error operation and accuracy feedback results that reflect the accuracy of visual recognition and the accuracy of limit features, if error operation and accuracy feedback is performed. The abnormal superposition interference optimization judgment module is used to perform fusion data accuracy analysis if no erroneous operation and accuracy feedback are performed, and based on the output fusion data accuracy analysis results that reflect the qualified status of data fusion, it determines whether to perform abnormal superposition interference optimization to reduce the input noise of visual recognition data and limit feature data.

Images