Connecting line fault diagnosis method and system combined with intelligent detection

Abstract

The present invention provides a connecting line fault diagnosis method and system combined with intelligent detection. First, a composite detection data set is obtained, which includes continuously collected operating signal data, connection structure image data composed of connection point appearance images from different perspectives, and environmental status data composed of temperature and humidity measurement records at different time periods. Then, the composite detection data set is subjected to time-space alignment processing to generate an integrated data pool with time-space correlation. Then, based on the integrated data pool, fault features are mined to generate a dynamic diagnosis basis including a fault feature combination library. An intelligent diagnosis model is called to jointly analyze multiple types of data in the integrated data pool to generate preliminary diagnosis results. Finally, the preliminary diagnosis results are cross-validated using the composite detection data set, candidate fault types are corrected, and a final diagnosis report including the fault location and type is generated and sent to maintenance equipment, thereby improving the accuracy and reliability of connecting line fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to a connection line fault diagnosis method and system combined with intelligent detection. Background Art

[0002] In many fields, such as power and communications, connecting cables are critical components, and their operating status directly impacts the stability and safety of the entire system. Traditional methods for diagnosing connecting cable faults have numerous limitations. For one thing, single-data-source diagnostic methods are common. For example, relying solely on operating signal data to diagnose faults is common. However, operating signals can be subject to interference from various factors, making it difficult to accurately and comprehensively identify the fault type and location based solely on such data. Using only connecting structure image data can lead to misjudgments of connecting structure deformation due to issues such as image quality and shooting angle. Relying solely on environmental status data can also easily lead to misjudgments, as the correlation between environmental factors and faults is not absolutely direct. Furthermore, existing methods lack comprehensive consideration of the spatiotemporal relationships among multiple data types in data processing. The acquisition time and spatial location of different data types are not effectively integrated, preventing the full exploration of potential connections between the data, which in turn affects the accuracy and reliability of fault diagnosis. Summary of the Invention

[0003] In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a connection line fault diagnosis method combined with intelligent detection, the method comprising:

[0004] Acquire a composite detection data set of the connecting wire, the composite detection data set comprising continuously collected operating signal data, connection structure image data, and environmental status data, wherein the operating signal data comprises voltage fluctuation records at different time points, the connection structure image data comprises appearance images of the connection points at different viewing angles, and the environmental status data comprises temperature and humidity measurement records at different time periods;

[0005] Performing temporal spatial alignment processing on the composite detection data set, establishing a time correspondence by running the acquisition time of the signal data, the shooting time of the connection structure image data, and the measurement time of the environmental status data, and establishing a spatial position correspondence of the connection structure image data according to the physical layout of the connection line, thereby generating an integrated data pool with temporal and spatial correlation, wherein the integrated data pool records the correspondence between multiple types of data in the same time interval and the same spatial position;

[0006] Performing a fault feature mining operation based on the integrated data pool, analyzing the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics corresponding to different fault types in historical diagnostic records, extracting associated feature combinations of the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics, and generating a dynamic diagnostic basis including a fault feature combination library;

[0007] Invoke an intelligent diagnostic model to jointly analyze multiple types of data in the current integrated data pool. The intelligent diagnostic model identifies abnormal fluctuations in operating signal data, appearance deformations in connection structure image data, and abnormal changes in environmental status data based on the combination of associated features in the dynamic diagnostic basis, and generates a preliminary diagnostic result containing candidate fault types.

[0008] The preliminary diagnostic results are cross-validated using multiple types of data in the composite detection data set. By verifying the correspondence between the abnormal operation signal and the deformation of the connection structure, and the matching between the abnormal environmental state and the historical fault environment, the candidate fault type is corrected, and a final diagnostic report containing the fault location and type is generated, and the final diagnostic report is sent to the maintenance equipment.

[0009] On the other hand, an embodiment of the present invention also provides a connection line fault diagnosis system combined with intelligent detection, including a processor and a machine-readable storage medium, the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to run the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0010] Based on the above aspects, the embodiment of the present invention realizes the comprehensive collection of multi-source data by acquiring a composite detection data set including operation signal data, connection structure image data and environmental status data, performs time-space alignment processing on the composite detection data set, generates an integrated data pool with time-space correlation, breaks the time and space barriers between different types of data, and makes the intrinsic connection between the data clearly presented. The fault feature mining operation is performed based on the integrated data pool to generate a dynamic diagnostic basis including a fault feature combination library, which can dynamically adapt to different fault conditions and improve the accuracy and comprehensiveness of fault feature identification. The intelligent diagnosis model is called to jointly analyze the multiple types of data in the current integrated data pool, and the dynamic diagnosis basis is used to identify various types of anomalies and generate preliminary diagnosis results, realizing intelligent preliminary fault judgment. Finally, the preliminary diagnosis results are cross-validated using the multiple types of data in the composite detection data set, the candidate fault type is corrected, and a final diagnosis report including the fault location and type is generated, which greatly improves the accuracy and reliability of fault diagnosis, effectively reduces the misjudgment rate, and ensures the stable operation of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 The figure is a schematic diagram of the execution flow of the connection line fault diagnosis method combined with intelligent detection provided by an embodiment of the present invention.

[0012] Figure 2 Schematic diagram of the hardware architecture of a connection line fault diagnosis system combined with intelligent detection provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0013] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 FIG1 is a flow chart of a method for diagnosing a fault of a connecting line combined with intelligent detection provided by an embodiment of the present invention. The method for diagnosing a fault of a connecting line combined with intelligent detection is introduced in detail below.

[0014] Step S110: Obtain a composite detection data set of the connecting line, wherein the composite detection data set includes continuously collected operation signal data, connection structure image data and environmental status data, wherein the operation signal data is composed of voltage fluctuation records at different time points, the connection structure image data is composed of connection point appearance images from different perspectives, and the environmental status data is composed of temperature and humidity measurement records at different time periods.

[0015] On industrial automation production lines, various devices communicate and power each other via numerous cables to ensure normal operation. To promptly detect potential cable failures, comprehensive cable inspections are necessary. For operational signal data, high-precision voltage sensors are installed on each cable. These sensors monitor the voltage in real time at regular intervals and record voltage fluctuations at different points in time. For example, when a device starts or stops, or when the load changes, the voltage on the cable fluctuates accordingly. These fluctuations are accurately recorded by the sensors, forming operational signal data.

[0016] Image data of the connection structure is collected using high-definition cameras placed at key locations throughout the production line. These cameras capture the connection points of the connecting wires from different perspectives, providing comprehensive information on their appearance. Images from different angles can capture the various aspects of the connection, such as whether the connection is loose, damaged, or corroded. For example, a frontal view reveals the overall appearance of a connection, while a side view may reveal subtle cracks hidden in corners.

[0017] Environmental status data is collected by temperature and humidity sensors installed around the production line. These sensors measure the ambient temperature and humidity at different times and record the corresponding data. The ambient temperature and humidity of the production line have a significant impact on the performance of the connecting cables. For example, high temperatures can accelerate the aging of the insulation material of the connecting cables, while excessive humidity can cause faults such as short circuits. Therefore, continuous monitoring of changes in ambient temperature and humidity is crucial for fault diagnosis.

[0018] Step S120: Perform temporal-spatial alignment processing on the composite detection data set, establish a time correspondence by running the acquisition time of the signal data, the shooting time of the connection structure image data, and the measurement time of the environmental status data, and at the same time establish a spatial position correspondence of the connection structure image data according to the physical layout of the connection line, and generate an integrated data pool with time-space correlation, which records the correspondence between multiple types of data in the same time interval and the same spatial position.

[0019] Step S121: extracting the acquisition time point of each voltage fluctuation record in the operation signal data to generate a signal time series.

[0020] From the operating signal data, the acquisition time points corresponding to each voltage fluctuation record are extracted one by one. Since the operating signal data is recorded continuously in chronological order, each voltage fluctuation record has a unique acquisition time. Arranging these acquisition time points in chronological order forms a signal time series. This time series effectively displays the chronological order of the voltage fluctuation records. For example, the first voltage fluctuation record was acquired at a certain moment after the production line started up, the second record at a later point, and so on. All acquisition time points form a continuous time series.

[0021] Step S122: extracting the shooting time point and shooting angle information of each appearance image in the connection structure image data to generate an image time-space sequence.

[0022] For connected structure image data, it is necessary to accurately extract the shooting time and shooting angle information of each appearance image. The shooting time can be obtained from the camera's shooting record, which accurately indicates when the image was taken. The shooting angle information is related to the camera's installation position and orientation. Cameras in different positions and orientations have different shooting angles. The shooting time and shooting angle information of each image are combined to generate an image time-space sequence. This image time-space sequence not only contains the image's time information, but also reflects the image's spatial angle information. For example, the shooting time of an image is at a specific time period during the production line's operation, and the shooting angle is from the upper left side of the production line. By integrating this information with the corresponding information of other images, a comprehensive image time-space sequence is formed.

[0023] Step S123: extracting the measurement time point of each temperature and humidity measurement record in the environmental state data to generate an environmental time series.

[0024] Each temperature and humidity measurement record in the environmental status data has a corresponding measurement time point. By organizing the environmental status data, the measurement time point of each measurement record is extracted and arranged in chronological order to generate an environmental time series. This time series shows the chronological order of temperature and humidity measurements, along with the signal time series and the image time-space series. For example, the first temperature and humidity measurement record is measured at the beginning of the production line operation, and the second record is measured at a certain time later. All measurement time points are arranged in sequence to form the environmental time series.

[0025] Step S124: uniformly convert the signal time series, image time series and environment time series into time values ​​of the same time unit, divide the time windows based on the minimum time interval, and determine the signal records, image images and environment records contained in each time window.

[0026] Step S1241: Convert the collection time point of the operation signal data into the accumulated time value starting from the initial moment, and generate a signal time value list.

[0027] In order to facilitate the subsequent time alignment processing, the collection time points of the operating signal data need to be uniformly converted into cumulative time values ​​starting from the initial moment. The initial moment can be set as the time point when the production line starts running. For each collection time point, the time interval between it and the initial moment is calculated, and the time interval is used as the cumulative time value corresponding to the collection time point. The cumulative time values ​​corresponding to all collection time points are arranged in sequence to generate a list of signal time values. For example, the time interval between the first collection time point and the initial moment is zero, and the corresponding cumulative time value is also zero; the second collection time point has a certain time interval from the initial moment, and the interval is accumulated to obtain the corresponding cumulative time value, and so on, forming a list containing the cumulative time values ​​of all collection time points.

[0028] Step S1242: Convert the shooting time points of the connection structure image data into time values ​​with the same accumulated time value, and generate an image time value list.

[0029] Similarly, the capture time points of the connected structure image data also need to be converted into cumulative time values ​​starting from the initial moment. Because the capture time points may not be synchronized with the acquisition time points of the operational signal data, a unified conversion process is required. The time interval between the capture time point and the initial moment of each image is calculated, and this time interval is used as the cumulative time value corresponding to that capture time point. The cumulative time values ​​corresponding to all capture time points are arranged in sequence to generate an image time value list. This image time value list and the signal time value list use the same time units and starting base, facilitating time alignment.

[0030] Step S1243: Convert the measurement time points of the environmental status data into time values ​​with the same accumulated time value, and generate an environmental time value list.

[0031] The same process is performed on the measurement time points of the environmental status data, converting them into cumulative time values ​​starting from the initial moment. The time interval between each measurement time point and the initial moment is calculated and used as the cumulative time value corresponding to that measurement time point. The cumulative time values ​​corresponding to all measurement time points are arranged in sequence to generate an environmental time value list. Thus, the signal time value list, the image time value list, and the environmental time value list are all based on the same time unit and initial moment.

[0032] Step S1244: Calculate the intervals between adjacent time points in the signal time value list to obtain the signal time interval; calculate the intervals between adjacent time points in the image time value list to obtain the image time interval; calculate the intervals between adjacent time points in the environment time value list to obtain the environment time interval.

[0033] After obtaining the three time value lists, calculate the intervals between adjacent time points. For the signal time value list, calculate the intervals between the first and second time points, the intervals between the second and third time points, and so on, to obtain a series of signal time intervals. The same method is used to process the image time value list and the environmental time value list to obtain the image time interval and environmental time interval, respectively. These time intervals reflect the differences in the temporal frequency of different data types.

[0034] Step S1245: Compare the signal time interval, the image time interval, and the environment time interval, and determine the minimum time interval as the time window length.

[0035] Compare the calculated signal, image, and environmental time intervals to find the minimum. This minimum interval will be used as the length of the subsequent time windows. Choosing the minimum time interval as the time window length ensures that each time window contains as many data records as possible, improving data relevance and integrity. For example, if the signal time interval is the minimum, then using the signal time interval as the time window length ensures that each time window contains at least one operational signal record, and may also include connection structure image data and environmental status data.

[0036] Step S1246: Starting from the initial time point, continuous time windows are divided with the time window length as a step length, and each time window is represented as an interval from the start time to the end time.

[0037] Starting from the initial moment, successive time windows are divided sequentially according to the determined time window length. Each time window has a specific start and end time. The start time is the end time of the previous time window (the start time of the first time window is the initial moment), and the end time is the start time plus the time window length. This divides the entire data collection period into multiple consecutive time windows. For example, the initial moment is the time when the production line starts running. Using the determined time window length as the step length, the first time window, the second time window, and so on are divided sequentially. Each time window represents a specific time period.

[0038] Step S1247: For each time window, filter out the voltage fluctuation records that fall within the interval in the signal time value list, filter out the appearance images that fall within the interval in the image time value list, and filter out the temperature and humidity measurement records that fall within the interval in the environment time value list, to complete the matching of multiple types of data in the time window.

[0039] For each divided time window, search the signal time value list for the voltage fluctuation record corresponding to the time point within the time window interval. Similarly, search the image time value list for the appearance image corresponding to the time point within the interval, and search the environment time value list for the temperature and humidity measurement record corresponding to the time point within the interval. Through the above screening method, the signal records, image images and environment records in the same time window are matched, so that each time window contains different types of data records, realizing data alignment in time. For example, in a certain time window, the corresponding voltage fluctuation records, appearance images and temperature and humidity measurement records are screened out. These data are temporally correlated and reflect the operating status of the connecting line, the appearance of the connection point and the environmental status during the time period.

[0040] Step S125: establishing a spatial position correspondence relationship of the connection structure image data according to the physical layout of the connection lines.

[0041] On a production line, connecting lines have a specific physical layout, with different connection points located in different spatial locations. By performing a detailed analysis and recording of the physical layout of the connecting lines, the specific spatial location of each connection point is determined. Simultaneously, the spatial location of the connection point corresponding to each piece of connecting structure image data is determined based on the camera's installation position and viewing angle. For example, an image captured by a camera at a certain position on the left side of the production line corresponds to a connection point located in a certain area on the left side of the production line. A correspondence is established between the image and the connection points in that area. This establishes a spatial correspondence between the connecting structure image data, allowing the spatial location of the connection point to be accurately located based on the image information.

[0042] Step S126: The operation signal records, connection structure images and environmental status records within the same time window are associated and stored with the corresponding physical location coordinates to generate an integrated data unit containing a time window identifier, a physical location identifier and multiple types of data, and all integrated data units are aggregated into an integrated data pool with time and space association.

[0043] After completing the matching of multiple types of data within the time window and establishing the spatial position correspondence of the connection structure image data, the operation signal records, connection structure image images and environmental status records within the same time window are associated with the corresponding physical location coordinates. Each time window has its own unique identifier, and the physical location coordinates also have corresponding identifiers. This information is stored together with the multiple types of data to form an integrated data unit. For example, the operation signal records, corresponding connection structure image images and environmental status records within a certain time window are associated with the identifier of the time window and the physical location identifier of the connection point and stored as a data unit. By summarizing all the above integrated data units, an integrated data pool with spatiotemporal association is formed. The integrated data pool records the correspondence between multiple types of data in the same time interval and the same spatial location.

[0044] Step S130: Perform fault feature mining operations based on the integrated data pool, analyze the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental status characteristics corresponding to different fault types in the historical diagnostic records, extract the associated feature combinations of the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental status characteristics, and generate a dynamic diagnostic basis containing a fault feature combination library.

[0045] Step S131: extracting historical diagnostic records from the integrated data pool, wherein the historical diagnostic records include operating signal data of confirmed fault types, connection structure image data, environmental status data, and corresponding fault location information.

[0046] Historical diagnostic records are filtered from the integrated data pool. These records represent confirmed fault cases. Each historical diagnostic record contains operational signal data, connection structure image data, environmental status data, and the corresponding fault location information at the time of the fault. For example, a historical fault case records the voltage fluctuations on the connecting line at the time of the fault (operating signal data), the appearance image of the connection point (connection structure image data), the ambient temperature and humidity at the time (environmental status data), and the specific location of the fault. By extracting and analyzing these historical diagnostic records, characteristic patterns corresponding to different fault types can be summarized.

[0047] Step S132: Statistically analyze the operation signal data in the historical diagnosis records to determine the changing trend, fluctuation frequency and occurrence pattern of abnormal amplitude of voltage fluctuation under different fault types, and generate an operation signal abnormal pattern set.

[0048] Conduct in-depth statistical analysis of operating signal data from historical diagnostic records. For different fault types, carefully observe the corresponding voltage fluctuation trends. Some faults may cause a continuous voltage rise, which may be caused by a short circuit or excessive load. Other faults may cause a sudden voltage drop, which may be caused by a broken line or poor contact. Still other faults may cause periodic voltage fluctuations, which may be related to cyclical equipment operation or interference. The frequency of voltage fluctuations (the number of voltage fluctuations per unit time) is also calculated to understand the frequency of voltage fluctuations. Furthermore, analyze the occurrence patterns of abnormal amplitudes, that is, voltage fluctuations exceeding the normal range. By analyzing a large amount of historical data, identify abnormal operating signal patterns for different fault types and aggregate these patterns into a set of abnormal operating signal patterns. For example, for a specific fault type, summarize the corresponding voltage fluctuation trend, frequency, and abnormal amplitude patterns. This information is added to the set as a set of abnormal operating signal patterns.

[0049] Step S133: Perform image analysis on the connection structure image data in the historical diagnostic records. By comparing the appearance differences between normal connection points and faulty connection points, the deformation features of the connection points are extracted to generate a set of connection structure deformation features. The deformation features include the degree of edge blur, surface damage area, and shape offset.

[0050] Step S1331: Selecting an appearance image of a normal connection point from historical diagnosis records as a reference image, and selecting an appearance image of a faulty connection point as a comparison image.

[0051] From historical diagnostic records, carefully select images of the appearance of normal and faulty connection points. The images of normal connection points serve as baseline images, representing the appearance characteristics of the connection points in their normal state. The images of faulty connection points serve as comparison images, used for comparison with the baseline images to detect changes in the connection point's appearance when it is faulty. For example, from a large number of historical images, select images of connection points that are clear, undamaged, and undeformed as baseline images, and select images of connection points with obvious problems as comparison images.

[0052] Step S1332: grayscale processing is performed on the reference image and the comparison image to generate a grayscale reference image and a grayscale comparison image.

[0053] To facilitate subsequent image analysis, the reference image and comparison image undergo grayscale processing. Grayscale processing converts a color image into a grayscale image, reducing image complexity and highlighting key features. In a grayscale image, each pixel has a single grayscale value, rather than the multiple color channel values ​​found in a color image. Through grayscale processing, a grayscale reference image and a grayscale comparison image are generated. These two images contain only grayscale information, making them easier to perform operations such as edge detection and contour analysis. For example, the color value of each pixel in the reference image and comparison image is converted to a corresponding grayscale value to generate a grayscale image.

[0054] Step S1333: performing edge detection on the grayscale reference image and the grayscale contrast image, extracting the contour edges of the connection points, and generating a reference contour and a contrast contour.

[0055] Use an edge detection algorithm to process the grayscale reference image and grayscale contrast image to extract the contour edges of the tie points. Edge detection algorithms can identify areas within an image with significant grayscale value variations, which typically correspond to object edges. Through edge detection, the contours of the tie points in the reference and contrast images are obtained, referred to as the reference contour and contrast contour, respectively. For example, common edge detection algorithms, such as the Sobel operator and the Canny operator, are used to process the grayscale image and identify the edge pixels of the tie points, thereby generating the contours.

[0056] Step S1334: Calculate the overlap ratio between the edge pixels of the comparison contour and the edge pixels of the reference contour to generate an edge blur degree parameter. The lower the overlap ratio, the higher the edge blur degree.

[0057] The edge pixels of the comparison contour are compared with those of the reference contour, and their overlap ratio is calculated. The overlap ratio reflects the degree of similarity between the two contours. A low overlap ratio indicates that the edges of the comparison contour differ significantly from those of the reference contour, indicating that the edge of the tie point is blurred. For example, the number of overlapping edge pixels in the comparison contour and the reference contour is counted and compared with the total number of edge pixels to obtain the overlap ratio. This overlap ratio is used as an edge blur parameter to measure the blurriness of the tie point edge.

[0058] Step S1335: Count the number of pixels in the abnormal color area of ​​the connection point surface in the comparison image, calculate the ratio of the number of pixels in the abnormal area to the total number of pixels of the connection point, and generate a surface damage area parameter.

[0059] In the comparison image, carefully count the number of pixels in areas with abnormal color on the tie point surface. These areas may indicate damage, oxidation, or other issues on the tie point surface. Compare the number of pixels in these areas to the total number of pixels in the tie point and calculate their ratio. This ratio is the surface damage area parameter, which reflects the extent of the damage. For example, by performing color analysis on the comparison image, identify areas with significantly different colors from normal areas, count the number of pixels in these areas, and then calculate the ratio of this ratio to the total number of pixels in the tie point.

[0060] Step S1336: Calculate the position offset distance between the geometric center of the comparison contour and the geometric center of the reference contour to generate shape offset parameters.

[0061] The geometric centers of the comparison contour and the reference contour are calculated, and then the positional offset distance between these two geometric centers is calculated. This offset distance represents the degree to which the shape of the connection point has shifted relative to its normal state and is used as a shape offset parameter. For example, by analyzing the pixel distribution of the comparison contour and the reference contour, the coordinates of their geometric centers are calculated, and then the distance between the two geometric centers is calculated using a coordinate calculation method.

[0062] Step S1337: The edge blur degree parameter, the surface damage area parameter, and the shape offset parameter are combined into the deformation feature of the connection point, and the deformation features of all historical fault cases are collected to generate a connection structure deformation feature set.

[0063] The calculated edge blur, surface damage area, and shape offset parameters are combined to form the deformation characteristics of the connection point. This process is repeated for all historical fault cases, and the deformation characteristics of each case are collected to ultimately generate a set of connection structure deformation characteristics. This set of connection structure deformation characteristics contains deformation characteristics of connection points under different fault types.

[0064] Step S134: Statistically analyze the environmental status data in the historical diagnosis records to determine the duration range, change rate, and frequency of abnormal values ​​of temperature and humidity under different fault types, and generate an environmental status abnormality feature set.

[0065] Perform a comprehensive statistical analysis of the environmental status data in the historical diagnostic records. For different fault types, analyze the continuous range of temperature and humidity, that is, the range in which the temperature and humidity are maintained during the fault. For example, some faults may be more likely to occur in high temperature and high humidity environments, so it is necessary to determine the continuous range of temperature and humidity under the above fault types. At the same time, calculate the rate of change of temperature and humidity, that is, the amount of change in temperature and humidity per unit time, to understand the changes in temperature and humidity. In addition, count the frequency of occurrence of abnormal values, that is, the number of times the temperature and humidity exceed the normal range. By analyzing a large amount of historical data, determine the abnormal characteristics of the environmental status under different fault types, and summarize these characteristics into a set of abnormal environmental status features. For example, for a specific fault type, summarize its corresponding continuous range of temperature and humidity, rate of change, and frequency of abnormal value occurrence, and add this information as an abnormal environmental status feature to the set.

[0066] Step S135: Analyze the correlation between the abnormal operating signal pattern set, the connection structure deformation feature set, and the abnormal environmental state feature set, count the situations where the abnormal operating signal pattern set, the connection structure deformation feature set, and the abnormal environmental state feature set appear at the same time under the same fault type, extract frequently occurring feature combinations, and generate a fault feature combination library.

[0067] Conduct in-depth correlation analysis on the set of abnormal operating signal patterns, the set of connection structure deformation features, and the set of abnormal environmental state features. Count the instances where these abnormal operating signal patterns, connection structure deformation features, and abnormal environmental state features occur simultaneously under the same fault type. For example, for a certain fault type, analyze whether the abnormal operating signal patterns, connection structure deformation features, and abnormal environmental state features appear simultaneously when the fault occurs. Through statistical analysis of a large number of historical fault cases, frequently occurring feature combinations are extracted. For example, if a specific abnormal operating signal pattern, connection structure deformation feature, and abnormal environmental state feature appear simultaneously in many fault cases of the same type, then the combination of these abnormal operating signal patterns, connection structure deformation features, and abnormal environmental state feature sets is extracted as a frequently occurring feature combination. All of the aforementioned frequent feature combinations are aggregated to generate a fault feature combination library, which records feature combination information corresponding to different fault types.

[0068] Step S136: Summarize the abnormal operation signal pattern set, the connection structure deformation feature set, the abnormal environment state feature set, and the fault feature combination library as a dynamic diagnosis basis.

[0069] The set of abnormal operating signal patterns, the set of connection structure deformation characteristics, the set of abnormal environmental state characteristics, and the fault feature combination library are aggregated to form a dynamic diagnostic basis. This dynamic diagnostic basis includes the abnormal operating signal patterns, connection structure deformation characteristics, abnormal environmental state characteristics corresponding to different fault types, and the associated feature combinations between them. For example, during fault diagnosis, this dynamic diagnostic basis can be used to determine whether the current operating signal, connection structure image, and environmental state meet the feature combination of a certain fault type, thereby determining the possible fault type.

[0070] Step S140: Call the intelligent diagnosis model to perform a joint analysis on multiple types of data in the current integrated data pool. The intelligent diagnosis model identifies abnormal fluctuations in the operating signal data, appearance deformations in the connection structure image data, and abnormal changes in the environmental status data based on the combination of associated features in the dynamic diagnosis basis, and generates a preliminary diagnosis result containing candidate fault types.

[0071] Step S141: extracting operation signal data from the current integrated data pool, analyzing the change trend of voltage fluctuation, fluctuation frequency and the occurrence of abnormal amplitude, and obtaining the current operation signal characteristics.

[0072] Extract the operating signal data from the current integrated data pool and analyze it in detail. Analyze the changing trend of voltage fluctuations to determine whether the voltage is rising, falling, or remaining stable. For example, if the voltage continues to rise, it may indicate problems such as a line short circuit or excessive load; if the voltage suddenly drops, it may be caused by a line break or poor contact. Count the frequency of voltage fluctuations to understand how frequent the voltage fluctuations are. At the same time, observe the occurrence of abnormal amplitudes, that is, situations where the voltage fluctuations exceed the normal range. Through these analyses, the characteristics of the current operating signal are obtained, which reflect the status of the current operating signal. For example, through the analysis of the operating signal data, it is found that the current voltage fluctuation trend is rising, the fluctuation frequency is high, and there are multiple abnormal amplitudes. This information is used as the characteristics of the current operating signal.

[0073] Step S142: extracting connection structure image data from the current integrated data pool, calculating the edge blur, surface damage area and shape offset of the connection point through edge detection and contour comparison, and obtaining the current connection structure features.

[0074] Connected structure image data is extracted from the current integrated data pool and subjected to image analysis. First, edge detection is performed to extract the contours of the connected points. The contours of the current connected points are then compared with those of normal connected points to calculate the degree of edge blur, surface damage area, and shape offset. The degree of edge blur can be calculated by comparing the overlap ratio of the edge pixels of the contour with those of the normal contour; the surface damage area can be calculated by counting the number of pixels in the abnormal color area of ​​the connected point surface to the total number of pixels; and the shape offset can be calculated by calculating the positional offset between the geometric center of the current contour and the geometric center of the normal contour. These calculations yield the features of the current connected structure, which reflect the current appearance of the connected point. For example, analysis of the connected structure image data indicates that the current connected point has a high degree of edge blur, a large surface damage area, and a large shape offset. This information is used as the current connected structure features.

[0075] Step S143: extracting environmental status data from the current integrated data pool, analyzing the continuous range, change rate and occurrence of abnormal values ​​of temperature and humidity, and obtaining the current environmental status characteristics.

[0076] Environmental status data is extracted from the current integrated data pool and analyzed in detail. The persistent range of temperature and humidity is analyzed to determine whether it is within the normal range. The rate of change of temperature and humidity is calculated to understand the dynamics of temperature and humidity. Furthermore, the occurrence of outliers—that is, situations where temperature and humidity exceed the normal range—is observed. Through these analyses, characteristics of the current environmental status are derived, reflecting the current state of the environment. For example, if the analysis of environmental status data reveals that the persistent range of temperature and humidity exceeds the normal range, the rate of change is rapid, and there are multiple outliers, this information is used as a characteristic of the current environmental status.

[0077] Step S144: From the fault feature combination library based on dynamic diagnosis, traverse the feature combination corresponding to each fault type, check whether the current operation signal feature conforms to the operation signal abnormality mode of the fault type, whether the current connection structure feature conforms to the connection structure deformation feature of the fault type, and whether the current environmental state feature conforms to the environmental state abnormality feature of the fault type.

[0078] From the fault feature combination library used as the basis for dynamic diagnosis, the feature combinations corresponding to each fault type are sequentially traversed. For each feature combination, check whether the current operating signal characteristics match the operating signal anomaly pattern for that fault type. For example, check whether the current voltage fluctuation trend, fluctuation frequency, and abnormal amplitude are consistent with the operating signal anomaly pattern for that fault type. Also, check whether the current connection structure characteristics match the connection structure deformation characteristics for that fault type. Specifically, check whether the edge blurring, surface damage area, and shape deviation meet the requirements. Furthermore, check whether the current environmental state characteristics match the environmental state anomaly characteristics for that fault type. Specifically, check whether the temperature and humidity duration, change rate, and abnormal value occurrence are consistent. For example, for a certain fault type, the operating signal anomaly pattern is a continuous voltage increase with a high fluctuation frequency and abnormal amplitude; the connection structure deformation characteristics are high edge blurring, large surface damage area, and large shape deviation; and the environmental state anomaly characteristics are temperature and humidity outside the normal range with a rapid change rate. Check whether the current operating signal characteristics, connection structure characteristics, and environmental state characteristics meet these requirements.

[0079] Step S145: Count the fault types that meet the feature combination conditions, record the number of features that meet the conditions, and generate a feature matching degree.

[0080] Step S1451: For each fault type feature combination in the dynamic diagnosis basis, determine the number of abnormal operation signal patterns, connection structure deformation features, and abnormal environmental state features contained therein as three feature conditions.

[0081] For each fault type's characteristic combination in the dynamic diagnosis basis, the three characteristics it contains are clearly defined: the abnormal operating signal pattern, the connection structure deformation characteristics, and the abnormal environmental state characteristics. These three characteristics are used as the three characteristic conditions. For example, for a certain fault type's characteristic combination, its abnormal operating signal pattern specifies the specific voltage fluctuation trend, fluctuation frequency, and abnormal amplitude; the connection structure deformation characteristics specify the requirements for edge blur, surface damage area, and shape offset; and the abnormal environmental state characteristics specify the continuous range, change rate, and abnormal value occurrence of temperature and humidity. These three characteristics constitute the three characteristic conditions for this fault type.

[0082] Step S1452: Check whether the current operating signal feature is consistent with the operating signal abnormality pattern of the fault type. If they are consistent, a feature condition is met.

[0083] Perform a detailed comparison of the current operating signal characteristics with the operating signal anomaly pattern for this fault type. Check whether the current voltage fluctuation trend, fluctuation frequency, and occurrence of abnormal amplitudes are consistent with the requirements of the operating signal anomaly pattern. If they are consistent, a characteristic condition is considered to be met. For example, if the operating signal anomaly pattern for this fault type requires an upward trend in voltage fluctuations, a high frequency, and the presence of abnormal amplitudes, and the current operating signal characteristics show an upward trend in voltage fluctuations, a high frequency, and multiple occurrences of abnormal amplitudes, then the characteristic condition is met.

[0084] Step S1453: Check whether the current connection structure feature is consistent with the connection structure deformation feature of the fault type. If they are consistent, a feature condition is satisfied.

[0085] Compare the current connection structure characteristics with the connection structure deformation characteristics for this fault type. Check whether the edge blur, surface damage area, and shape offset of the current connection point are consistent with the requirements of the connection structure deformation characteristics. If they are consistent, it is considered that a characteristic condition has been met. For example, if the connection structure deformation characteristics for this fault type require a high degree of edge blur, a large surface damage area, and a large shape offset, and the current connection structure characteristics show that the connection point has a high degree of edge blur, a large surface damage area, and a large shape offset, then the characteristic condition has been met.

[0086] Step S1454: Check whether the current environmental state characteristics are consistent with the abnormal environmental state characteristics of the fault type. If they are consistent, a characteristic condition is met.

[0087] Compare the current environmental state characteristics with the environmental state anomaly characteristics for this fault type. Check whether the current temperature and humidity persistence range, rate of change, and occurrence of abnormal values ​​are consistent with the requirements of the environmental state anomaly characteristics. If they are consistent, it is considered that a characteristic condition has been met. For example, if the environmental state anomaly characteristics for this fault type require that the temperature and humidity exceed the normal range, change at a rapid rate, and have multiple abnormal values, and the current environmental state characteristics show that the persistence range of the temperature and humidity does exceed the normal range, change at a rapid rate, and have multiple abnormal values, then the characteristic condition has been met.

[0088] Step S1455: Count the number of characteristic conditions satisfied by the fault type, calculate the ratio of the number of satisfied characteristic conditions to the total number of characteristic conditions, and generate the characteristic matching degree of the fault type.

[0089] Count the number of characteristic conditions that the fault type meets, specifically the three conditions: abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental conditions. Then, compare the number of characteristic conditions met with the total number of characteristic conditions (three) and calculate the ratio. This ratio represents the characteristic matching degree for the fault type. For example, if a fault type meets two characteristic conditions, its characteristic matching degree is two-thirds.

[0090] Step S1456: Traverse the feature combinations of all fault types, repeat the above steps, and generate the feature matching degree of each fault type.

[0091] Traverse the feature combinations of all fault types in the fault feature combination library used as the basis for dynamic diagnosis, repeat the above inspection and calculation steps, and generate a feature matching degree for each fault type. This method determines the degree of match between each fault type and the current data. For example, traverse all fault types in the fault feature combination library and calculate the feature matching degree for each fault type. The feature matching degree for fault type A is two-thirds, the feature matching degree for fault type B is one-third, and so on.

[0092] Step S146: The fault type with the highest feature matching degree is taken as a candidate fault type, and a preliminary diagnosis result including the candidate fault type and the feature matching degree is generated.

[0093] After generating the feature matching degree for each fault type, compare the feature matching degrees of all fault types to identify the fault type with the highest feature matching degree. This fault type is designated as a candidate fault type, and its feature matching degree is recorded. The candidate fault type and feature matching degree are combined to generate a preliminary diagnosis result containing the candidate fault type and feature matching degree. For example, after comparison, it is found that fault type A has the highest feature matching degree. Fault type A is selected as a candidate fault type, and its feature matching degree is two-thirds. This information is used as the preliminary diagnosis result.

[0094] Step S150: Use the multiple types of data in the composite detection data set to cross-validate the preliminary diagnostic results, and by verifying the correspondence between the abnormal operation signal and the deformation of the connection structure, and the matching between the abnormal environmental state and the historical fault environment, correct the candidate fault type, generate a final diagnostic report including the fault location and type, and send the final diagnostic report to the maintenance equipment.

[0095] Step S151: extracting candidate fault types and their feature matching degrees from the preliminary diagnosis results.

[0096] Extract candidate fault types and their corresponding feature matching degrees from the preliminary diagnosis results. A candidate fault type is the most likely fault type determined by the intelligent diagnosis model analysis. The feature matching degree reflects the degree of match between the fault type and the current data. For example, if the candidate fault type extracted from the preliminary diagnosis results is fault type A, its feature matching degree is two-thirds.

[0097] Step S152: extracting the abnormal operation signal pattern, connection structure deformation characteristics and environmental state abnormal characteristics corresponding to the candidate fault type from the dynamic diagnosis basis.

[0098] The dynamic diagnostic basis is used to identify the abnormal operating signal pattern, connection structure deformation characteristics, and abnormal environmental state characteristics corresponding to the candidate fault type. These characteristics are summarized from historical fault case analysis and represent the typical characteristics of this fault type. For example, for candidate fault type A, the dynamic diagnostic basis extracts the corresponding abnormal operating signal pattern as a continuous voltage increase, a high frequency of fluctuation, and abnormal amplitude; the connection structure deformation characteristics as high edge blur, large surface damage area, and large shape deviation; and the abnormal environmental state characteristics as temperature and humidity outside the normal range, with a rapid rate of change and multiple abnormal values.

[0099] Step S153: Analyze whether the abnormal fluctuation in the current running signal data corresponds to the appearance deformation in the connection structure image data in time and space, that is, whether the time when the abnormal fluctuation occurs is consistent with the shooting time of the appearance deformation image, and whether the connection point position corresponding to the abnormal fluctuation is consistent with the shooting position of the appearance deformation image, and generate a signal and structure correspondence score.

[0100] Step S1531: extracting a list of time points where abnormal fluctuations occur in the current operating signal data.

[0101] Carefully analyze the current operating signal data to identify the time points at which abnormal fluctuations occurred. Organize these time points into a list, recording the chronological order in which the abnormal fluctuations occurred. For example, if the operating signal data analysis reveals multiple time points of abnormal fluctuations, record these time points in sequence to form a list of abnormal fluctuation time points.

[0102] Step S1532: extracting a list of image shooting time points at which appearance deformation occurs in the current connection structure image data.

[0103] Extract the shooting time points of images exhibiting appearance deformation from the current connection structure image data. Organize these shooting time points into a list, recording the shooting time sequence of the appearance-deformed images. For example, by analyzing the connection structure image data, identify images exhibiting appearance deformation, record their shooting time points, and form a list of appearance-deformed image shooting time points.

[0104] Step S1533: Calculate the number of overlaps between the abnormal fluctuation time point and the appearance deformation shooting time point. A greater number of overlaps indicates a better time correspondence.

[0105] Compare the abnormal fluctuation time point list with the appearance deformation image capture time point list and count the number of overlapping time points. A greater number of overlaps indicates a greater temporal correspondence between the abnormal operating signal fluctuations and the connection structure's appearance deformation. For example, if the abnormal fluctuation time point list and the appearance deformation image capture time point list show multiple overlapping time points, count these overlapping time points.

[0106] Step S1534: Extract a list of connection point locations corresponding to abnormal fluctuations in the current operating signal data.

[0107] Analyze the current operating signal data to identify the connection point location corresponding to each abnormal fluctuation. Organize these connection point locations into a list, recording the locations where the abnormal fluctuations occurred. For example, by analyzing the operating signal data, determine the connection point location corresponding to each abnormal fluctuation, and record these locations in sequence to form a list of connection point locations corresponding to the abnormal fluctuations.

[0108] Step S1535: extracting a shooting position list corresponding to the appearance deformation image in the current connection structure image data.

[0109] Extract the shooting locations corresponding to each deformed image from the current connection structure image data. Organize these shooting locations into a list, recording the shooting locations of the deformed images. For example, by analyzing the connection structure image data, determine the connection point locations corresponding to each deformed image, and record these locations sequentially to form a list of shooting locations corresponding to the deformed images.

[0110] Step S1536: Calculate the number of overlaps between the abnormal fluctuation position and the appearance deformation position. A greater number of overlaps indicates better spatial correspondence.

[0111] Compare the list of connection point locations corresponding to the abnormal fluctuations with the list of shooting locations corresponding to the appearance deformation image, and count the number of overlapping locations. A greater number of overlaps indicates a better spatial correspondence between the abnormal fluctuations in the operating signal and the appearance deformation of the connection structure. For example, if the list of connection point locations corresponding to the abnormal fluctuations and the list of shooting locations corresponding to the appearance deformation image are compared and multiple overlapping locations are found, count these overlapping locations.

[0112] Step S1537: Add the temporal correspondence overlap number and the spatial correspondence overlap number to generate a signal-structure correspondence score.

[0113] The sum of the temporal and spatial correspondence overlaps is used as the signal-to-structure correspondence score. This score reflects the comprehensive temporal and spatial correspondence between abnormal fluctuations in the operating signal and the deformation of the connected structure. For example, if the temporal correspondence overlaps are three and the spatial correspondence overlaps are two, the sum of these two scores yields a signal-to-structure correspondence score of five.

[0114] Step S154: Analyze whether the abnormal changes in the current environmental status data are consistent with the abnormal environmental status characteristics of the fault type in the historical diagnostic records, that is, whether the continuous range, change rate and frequency of occurrence of abnormal values ​​of temperature and humidity match, and generate an environment and history matching score.

[0115] The temperature and humidity persistence range, rate of change, and frequency of abnormal values ​​in the current environmental status data are compared in detail with the abnormal environmental status characteristics of the candidate fault type in the historical diagnostic records. If the current temperature and humidity persistence range, rate of change, and frequency of abnormal values ​​are consistent with the historical characteristics, it means that the current environmental status matches the historical fault environment to a high degree; if not, the degree of match is low. Based on the degree of match, an environmental and historical matching score is generated. For example, if the temperature and humidity persistence range, rate of change, and frequency of abnormal values ​​in the current environmental status data are mostly consistent with the abnormal environmental status characteristics of the fault type in the historical diagnostic records, then the environmental and historical matching score is high; if only partially consistent, then the score is low.

[0116] Step S155: Perform weighted calculation on the feature matching degree, the signal-structure correspondence score, and the environment-history matching score to generate a comprehensive confidence level.

[0117] A weighted calculation is performed on the feature matching degree, signal-structure correspondence score, and environment-history matching score. Each score is assigned a weight, determined by its importance in fault diagnosis. Each score is multiplied by its corresponding weight, and the results are summed to obtain the overall confidence level. For example, the feature matching degree has a weight of 0.5, the signal-structure correspondence score has a weight of 0.3, and the environment-history matching score has a weight of 0.2. Assuming a feature matching degree of two-thirds, a signal-structure correspondence score of five, and an environment-history matching score of four, the feature matching degree is converted to a corresponding numerical value and then weighted to obtain the overall confidence level.

[0118] Step S156: If the comprehensive confidence exceeds the preset threshold, the candidate fault type is confirmed to be the final fault type, and the position with the most obvious appearance deformation in the current connection structure image data is extracted as the fault position.

[0119] A preset threshold is set to determine whether the comprehensive confidence level is high enough. If the comprehensive confidence level exceeds the preset threshold, it indicates that the candidate fault type has a high degree of match with the current data, and the candidate fault type can be confirmed as the final fault type. At the same time, the location with the most obvious appearance deformation is found from the current connection structure image data, and this location is used as the fault location. For example, if the preset threshold is a certain value and the calculated comprehensive confidence level exceeds the preset threshold, the candidate fault type is confirmed as the final fault type. Then, by analyzing the connection structure image data, the connection point location with the most severe appearance deformation is found and used as the fault location.

[0120] Step S157: Integrate the final fault type and fault location into a final diagnosis report.

[0121] The final confirmed fault type and location are combined to form a final diagnostic report. This report contains detailed information about the fault type and location, providing important guidance for subsequent maintenance work. For example, the final fault type is Fault Type A and the fault location is a specific connection point.

[0122] Step S158: Send the final diagnosis report to the maintenance equipment.

[0123] Step S1581: Extract the fault type and fault location information in the final diagnosis report.

[0124] Extract the fault type and location from the final diagnostic report. This information is crucial for maintenance personnel to perform fault repairs. For example, the final diagnostic report may identify the fault type as Fault Type A and the fault location as a specific connection point.

[0125] Step S1582: converting the fault type into a preset fault type description text, wherein the fault type description text includes specific types such as poor contact, insulation damage, and line aging.

[0126] The extracted fault type is converted into a preset fault type description. The preset fault type description includes common fault types such as poor contact, insulation damage, and line aging. Based on the characteristics of the fault type, it is mapped to the corresponding description text. For example, if fault type A corresponds to poor contact, it is converted into a description text of poor contact.

[0127] Step S1583: converting the fault location into a coordinate identifier in the physical layout diagram of the connection line, wherein the coordinate identifier corresponds to the location of the appearance deformation in the connection structure image data.

[0128] The fault location is converted to coordinates in the physical layout diagram of the connecting lines. The physical layout diagram of the connecting lines shows the specific location and layout of the connecting lines on the production line. By analyzing the location of the appearance deformation in the connecting structure image data, the fault location is mapped to the coordinates in the physical layout diagram. For example, based on the location of the appearance deformation in the connecting structure image data, the corresponding coordinates in the physical layout diagram of the connecting lines are found and converted to the coordinates.

[0129] Step S1584: Associating the fault type description text with the coordinate identifier to generate a diagnostic report content including text description and location mark.

[0130] The converted fault type description is associated with the coordinate identifier to generate a diagnostic report containing a text description and location annotation. This diagnostic report includes both a detailed description of the fault type and the specific location of the fault, providing clear information for maintenance personnel. For example, the fault type description of "poor contact" can be associated with a specific coordinate identifier to generate a diagnostic report containing the text description of the poor contact and the location of the coordinate identifier.

[0131] Step S1585: Send the diagnostic report content to the designated data receiving interface of the maintenance device and monitor the receiving status of the maintenance device. If no reception confirmation information is received within the specified time, resend the diagnostic report content until the reception confirmation information is received.

[0132] The generated diagnostic report content is sent to the designated data receiving interface of the maintenance device. At the same time, the maintenance device's reception status is monitored, and confirmation information is used to determine whether the maintenance device has successfully received the diagnostic report content. If confirmation information is not received within the specified time, it may indicate a transmission problem or that the maintenance device is not receiving the content properly. In this case, the diagnostic report content needs to be resent until confirmation information is received. For example, after the diagnostic report content is sent to the designated data receiving interface of the maintenance device, a timer starts. If confirmation information is not received within the specified time, the diagnostic report content is resent until confirmation information is received.

[0133] Figure 2 FIG. 1 shows the hardware structure of a connection line fault diagnosis system 100 for implementing the connection line fault diagnosis method combined with intelligent detection provided by an embodiment of the present invention. Figure 2 As shown, the connection line fault diagnosis system 100 combined with intelligent detection may include a processor 110 , a machine-readable storage medium 120 , a bus 130 , and a communication unit 140 .

[0134] In one possible design, the connecting line fault diagnosis system 100 combined with intelligent detection can be a single server or a server group. The server group can be centralized or distributed (for example, the connecting line fault diagnosis system 100 combined with intelligent detection can be a distributed system). In some embodiments, the connecting line fault diagnosis system 100 combined with intelligent detection can be local or remote. For example, the connecting line fault diagnosis system 100 combined with intelligent detection can access information and / or data stored in the machine-readable storage medium 120 via a network. For another example, the connecting line fault diagnosis system 100 combined with intelligent detection can be directly connected to the machine-readable storage medium 120 to access the stored information and / or data.

[0135] The machine-readable storage medium 120 can store data and / or instructions. In some embodiments, the machine-readable storage medium 120 can store data obtained from an external terminal. In some embodiments, the machine-readable storage medium 120 can store data and / or instructions that the connecting line fault diagnosis system 100 combined with intelligent detection uses to execute or perform the exemplary methods described herein.

[0136] During the specific implementation process, one or more processors 110 execute computer executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the connection line fault diagnosis method combined with intelligent detection in the above method embodiment. The processor 110, the machine-readable storage medium 120 and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the sending and receiving actions of the communication unit 140.

[0137] The specific implementation process of the processor 110 can refer to the various method embodiments executed by the above-mentioned connecting line fault diagnosis system 100 combined with intelligent detection. The implementation principles and technical effects are similar and will not be repeated here in this embodiment.

[0138] In addition, an embodiment of the present invention further provides a readable storage medium having computer executable instructions set therein. When a processor runs the computer executable instructions, the connection line fault diagnosis method combined with intelligent detection as described above is implemented.

[0139] It should be noted that, in order to simplify the description of the present invention and thus facilitate understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, various features may sometimes be grouped together into one embodiment, figure, or description thereof. Similarly, it should be noted that, in order to simplify the description of the present invention and thus facilitate understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, various features may sometimes be grouped together into one embodiment, figure, or description thereof.

Claims

1. A method for diagnosing connection line faults combined with intelligent detection, characterized in that: The method comprises: Acquire a composite detection data set of the connecting wire, the composite detection data set comprising continuously collected operating signal data, connection structure image data, and environmental status data, wherein the operating signal data comprises voltage fluctuation records at different time points, the connection structure image data comprises appearance images of the connection points at different viewing angles, and the environmental status data comprises temperature and humidity measurement records at different time periods; Performing temporal spatial alignment processing on the composite detection data set, establishing a time correspondence by running the acquisition time of the signal data, the shooting time of the connection structure image data, and the measurement time of the environmental status data, and establishing a spatial position correspondence of the connection structure image data according to the physical layout of the connection line, thereby generating an integrated data pool with temporal and spatial correlation, wherein the integrated data pool records the correspondence between multiple types of data in the same time interval and the same spatial position; Performing a fault feature mining operation based on the integrated data pool, analyzing the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics corresponding to different fault types in historical diagnostic records, extracting associated feature combinations of the abnormal operating signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics, and generating a dynamic diagnostic basis including a fault feature combination library; Invoke an intelligent diagnostic model to jointly analyze multiple types of data in the current integrated data pool. The intelligent diagnostic model identifies abnormal fluctuations in operating signal data, appearance deformations in connection structure image data, and abnormal changes in environmental status data based on the combination of associated features in the dynamic diagnostic basis, and generates a preliminary diagnostic result containing candidate fault types. The preliminary diagnostic results are cross-validated using multiple types of data in the composite detection data set. By verifying the correspondence between the abnormal operation signal and the deformation of the connection structure, and the matching between the abnormal environmental state and the historical fault environment, the candidate fault type is corrected, and a final diagnostic report containing the fault location and type is generated, and the final diagnostic report is sent to the maintenance equipment.

2. The connecting line fault diagnosis method combined with intelligent detection according to claim 1, characterized in that: The composite detection data set is subjected to temporal spatial alignment processing, and a time correspondence is established by running the acquisition time of the signal data, the shooting time of the connection structure image data, and the measurement time of the environmental status data. At the same time, a spatial position correspondence of the connection structure image data is established according to the physical layout of the connection line, thereby generating an integrated data pool with temporal and spatial correlation, including: Extract the acquisition time point of each voltage fluctuation record in the operating signal data to generate a signal time series; Extract the shooting time and viewing angle information of each appearance image in the connection structure image data to generate an image time-space sequence; Extract the measurement time point of each temperature and humidity measurement record in the environmental status data to generate an environmental time series; Convert the signal time series, image time series and environment time series into time values ​​of the same time unit, divide the time windows based on the minimum time interval, and determine the signal records, image images and environment records contained in each time window; According to the physical layout diagram of the connection line, the physical position coordinates of the corresponding connection points are marked for each connection structure image data to establish the spatial correspondence between the image and the physical position; The operation signal records, connection structure images and environmental status records within the same time window are associated and stored with the corresponding physical location coordinates to generate an integrated data unit containing the time window identifier, physical location identifier and multiple types of data. All integrated data units are aggregated into an integrated data pool with time and space association.

3. The connecting line fault diagnosis method combined with intelligent detection according to claim 2, characterized in that: The method of uniformly converting the signal time series, image time series, and environment time series into time values ​​of the same time unit, dividing the time windows based on the minimum time interval, and determining the signal records, image images, and environment records contained in each time window includes: Convert the acquisition time points of the operating signal data into the accumulated seconds from the initial moment to generate a signal time value list; convert the shooting time points of the connection structure image data into time values ​​with the same accumulated seconds to generate an image time value list; convert the measurement time points of the environmental status data into time values ​​with the same accumulated seconds to generate an environmental time value list; Calculate the intervals between adjacent time points in the signal time value list to obtain the signal time interval, calculate the intervals between adjacent time points in the image time value list to obtain the image time interval, and calculate the intervals between adjacent time points in the environment time value list to obtain the environment time interval; Compare the signal time interval, image time interval and environment time interval, and determine the minimum time interval as the time window length; Starting from the initial time point, continuous time windows are divided with the time window length as the step length. Each time window is represented as the interval from the start time to the end time; For each time window, the voltage fluctuation records falling within the interval in the signal time value list are filtered out, the appearance images falling within the interval in the image time value list are filtered out, and the temperature and humidity measurement records falling within the interval in the environment time value list are filtered out to complete the matching of multiple types of data in the time window.

4. The connecting line fault diagnosis method combined with intelligent detection according to claim 1, characterized in that: The fault feature mining operation is performed based on the integrated data pool, analyzing the abnormal operation signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics corresponding to different fault types in the historical diagnosis records, extracting the associated feature combinations of the abnormal operation signal patterns, connection structure deformation characteristics, and abnormal environmental state characteristics, and generating a dynamic diagnosis basis containing a fault feature combination library, including: Extracting historical diagnostic records from the integrated data pool, wherein the historical diagnostic records include operating signal data of confirmed fault types, connection structure image data, environmental status data, and corresponding fault location information; Perform statistical analysis on the operating signal data in historical diagnostic records to determine the changing trend, fluctuation frequency and occurrence pattern of abnormal amplitude of voltage fluctuation under different fault types, and generate a set of abnormal operating signal patterns; Perform image analysis on the connection structure image data in historical diagnostic records. By comparing the appearance differences between normal connection points and faulty connection points, the deformation features of the connection points are extracted to generate a set of connection structure deformation features. The deformation features include edge blur, surface damage area, and shape offset. Perform statistical analysis on environmental status data in historical diagnostic records to determine the duration, change rate, and frequency of abnormal values ​​of temperature and humidity under different fault types, and generate a set of abnormal environmental status features. Analyze the correlation between the abnormal pattern set of operating signals, the deformation feature set of connection structures, and the abnormal feature set of environmental conditions. Count the cases where the abnormal pattern set of operating signals, the deformation feature set of connection structures, and the abnormal feature set of environmental conditions occur simultaneously under the same fault type. Extract frequently occurring feature combinations and generate a fault feature combination library. The abnormal pattern set of operating signals, the deformation feature set of connection structures, the abnormal feature set of environmental status and the fault feature combination library are summarized as the basis for dynamic diagnosis.

5. The connecting line fault diagnosis method combined with intelligent detection according to claim 4 is characterized in that: The image analysis of the connection structure image data in the historical diagnosis records is performed to extract the deformation features of the connection points by comparing the appearance differences between normal connection points and faulty connection points, thereby generating a connection structure deformation feature set, including: Selecting an appearance image of a normal connection point from historical diagnosis records as a reference image, and selecting an appearance image of a faulty connection point as a comparison image; Performing grayscale processing on the reference image and the comparison image to generate a grayscale reference image and a grayscale comparison image; Perform edge detection on the grayscale reference image and the grayscale contrast image, extract the contour edges of the connection points, and generate the reference contour and the contrast contour; Calculate the overlap ratio between the edge pixels of the comparison contour and the edge pixels of the reference contour to generate an edge blur parameter. The lower the overlap ratio, the higher the edge blur. Count the number of pixels in the abnormal color area of ​​the connection point surface in the comparison image, calculate the ratio of the number of pixels in the abnormal area to the total number of pixels at the connection point, and generate the surface damage area parameter; Calculate the position offset distance between the geometric center of the comparison contour and the geometric center of the reference contour to generate shape offset parameters; The edge blur degree parameter, surface damage area parameter and shape offset parameter are combined into the deformation feature of the connection point. The deformation features of all historical failure cases are collected to generate the connection structure deformation feature set.

6. The connecting line fault diagnosis method combined with intelligent detection according to claim 1, characterized in that: The intelligent diagnosis model is called to jointly analyze multiple types of data in the current integrated data pool. The intelligent diagnosis model identifies abnormal fluctuations in the operating signal data, appearance deformations in the connection structure image data, and abnormal changes in the environmental status data based on the combination of associated features in the dynamic diagnosis basis, and generates a preliminary diagnosis result containing candidate fault types, including: Extract the operating signal data from the current integrated data pool, analyze the changing trend of voltage fluctuation, fluctuation frequency and the occurrence of abnormal amplitude, and obtain the current operating signal characteristics; Extract the connection structure image data from the current integrated data pool, calculate the edge blur, surface damage area and shape offset of the connection point through edge detection and contour comparison, and obtain the current connection structure characteristics; Extract environmental status data from the current integrated data pool, analyze the continuous range, change rate and occurrence of abnormal values ​​of temperature and humidity, and obtain the current environmental status characteristics; From the fault feature combination library based on dynamic diagnosis, traverse the feature combination corresponding to each fault type to check whether the current operating signal feature meets the operating signal abnormality mode of the fault type, whether the current connection structure feature meets the connection structure deformation feature of the fault type, and whether the current environmental state feature meets the environmental state abnormality feature of the fault type; Count the fault types that meet the feature combination conditions, record the number of features that meet the conditions, and generate feature matching degrees; The fault type with the highest feature matching degree is taken as the candidate fault type, and a preliminary diagnosis result including the candidate fault type and feature matching degree is generated.

7. The connecting line fault diagnosis method combined with intelligent detection according to claim 6, characterized in that: The counting of fault types that meet the feature combination conditions, recording the number of features that meet the conditions, and generating a feature matching degree includes: For each fault type feature combination in the dynamic diagnosis basis, the number of abnormal operation signal patterns, connection structure deformation features, and abnormal environmental state features included in the combination is determined as three feature conditions; Check whether the current operating signal characteristics are consistent with the operating signal abnormality pattern of the fault type. If they are consistent, a characteristic condition is met; Check whether the current connection structure characteristics are consistent with the connection structure deformation characteristics of the fault type. If they are consistent, a characteristic condition is met; Check whether the current environmental state characteristics are consistent with the abnormal environmental state characteristics of the fault type. If they are consistent, a characteristic condition is met; Count the number of characteristic conditions that the fault type meets, calculate the ratio of the number of characteristic conditions met to the total number of characteristic conditions, and generate the characteristic matching degree of the fault type; Traverse the feature combinations of all fault types and repeat the above steps to generate the feature matching degree of each fault type.

8. The connecting line fault diagnosis method combined with intelligent detection according to claim 1, characterized in that: The preliminary diagnosis results are cross-validated using multiple types of data in the composite detection data set. By verifying the correspondence between the abnormal operation signal and the deformation of the connection structure, and the matching between the abnormal environmental state and the historical fault environment, the candidate fault type is corrected and a final diagnosis report containing the fault location and type is generated, including: Extract candidate fault types and their feature matching degrees from the preliminary diagnosis results; Extracting the abnormal operation signal pattern, connection structure deformation characteristics and environmental state abnormal characteristics corresponding to the candidate fault type from the dynamic diagnosis basis; Analyze whether the abnormal fluctuations in the current running signal data correspond to the appearance deformation in the connection structure image data in time and space. That is, whether the time when the abnormal fluctuation occurs is consistent with the shooting time of the appearance deformation image, and whether the position of the connection point corresponding to the abnormal fluctuation is consistent with the shooting position of the appearance deformation image, and generate a signal-structure correspondence score; Analyze whether the abnormal changes in the current environmental status data are consistent with the abnormal environmental status characteristics of the fault type in the historical diagnosis records, that is, whether the continuous range, change rate and frequency of abnormal values ​​of temperature and humidity match, and generate an environment and history matching score; The feature matching degree, signal and structure correspondence score, and environment and history matching score are weighted and calculated to generate a comprehensive confidence score; If the comprehensive confidence exceeds the preset threshold, the candidate fault type is confirmed as the final fault type, and the position with the most obvious appearance deformation in the current connection structure image data is extracted as the fault location; Integrate the final fault type and fault location into the final diagnosis report; The step of analyzing whether abnormal fluctuations in the current running signal data correspond to appearance deformations in the connection structure image data in time and space, and generating a signal-structure correspondence score, includes: Extract the time point list of abnormal fluctuations in the current running signal data; Extracting a list of image shooting time points where appearance deformation occurs in the current connection structure image data; Calculate the number of overlaps between the abnormal fluctuation time point and the appearance deformation shooting time point. The greater the number of overlaps, the better the time correspondence. Extract the connection point location list corresponding to abnormal fluctuations in the current running signal data; Extracting a list of shooting positions corresponding to the appearance deformation image in the current connection structure image data; Calculate the number of overlaps between the abnormal fluctuation position and the appearance deformation position. The greater the number of overlaps, the better the spatial correspondence. The number of temporal correspondence coincidences is summed with the number of spatial correspondence coincidences to generate a signal-to-structure correspondence score.

9. The connecting line fault diagnosis method combined with intelligent detection according to claim 1, characterized in that: The sending of the final diagnostic report to the maintenance device includes: Extract the fault type and fault location information from the final diagnostic report; Converting the fault type into a preset fault type description text, wherein the fault type description text includes specific types such as poor contact, insulation damage, and line aging; Converting the fault location into a coordinate identifier in the physical layout diagram of the connection line, wherein the coordinate identifier corresponds to the location of the appearance deformation in the connection structure image data; Associate the fault type description text with the coordinate identifier to generate a diagnostic report containing text description and location annotation; The diagnostic report content is sent to the designated data receiving interface of the maintenance equipment, and the receiving status of the maintenance equipment is monitored. If the reception confirmation information is not received within the specified time, the diagnostic report content is resent until the reception confirmation information is received.

10. A connection line fault diagnosis system combined with intelligent detection, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to run the programs, instructions or codes in the memory to implement the connection line fault diagnosis method combined with intelligent detection as described in any one of claims 1 to 9.

Images