Electrical fault positioning diagnosis method for engine production line detection equipment

Abstract

The invention provides an electrical fault positioning diagnosis method for engine production line detection equipment, and relates to the technical field of electrical fault positioning diagnosis, and the method comprises the steps: connecting an engine production line, laying multiple types of detection equipment, and constructing a detection inner layer; performing abnormality diagnosis on each device to obtain production abnormality identification information, and storing the production abnormality identification information in the matched detection data block; based on the detection inner layer, performing data linkage by using the detection data blocks, and constructing a linkage detection outer layer; clustering is carried out, linkage identification is carried out according to the clustering relation and the production abnormity identification information, and conflict information is determined; and performing electrical abnormal data identification analysis, positioning an electrical fault, and generating fault diagnosis information according to the electrical abnormal data and the electrical positioning fault. The technical problems that in the prior art, multiple kinds of detection data on an engine production line cannot be effectively integrated, fragmentation and inconsistency exist in the detection data, and consequently the accuracy and efficiency of electrical fault positioning diagnosis are poor are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical fault location diagnosis, and particularly relates to an electrical fault location diagnosis method for an engine production line detection device. Background Art

[0002] As an important power device, the accurate operating state of an engine is crucial for production efficiency and product quality. In traditional engine production line monitoring, data collected by various types of detection devices may have problems such as inconsistent formats, different frequencies, or large differences in data volume. This fragmentation and inconsistency make it difficult to integrate and uniformly process data, limiting the comprehensive analysis and prediction ability of the overall production line health state; moreover, traditional fault diagnosis methods face problems such as coexistence of multiple fault modes and difficulty in detecting hidden faults. Especially in the case of simultaneous operation of multiple devices and production targets, accurately and quickly locating the fault point becomes particularly important but also particularly challenging. Summary of the Invention

[0003] This application provides an electrical fault location diagnosis method for an engine production line detection device, aiming to solve the technical problem that in the prior art, various detection data on the engine production line cannot be effectively integrated, and the detection data is fragmented and inconsistent, resulting in poor accuracy and efficiency of electrical fault location diagnosis.

[0004] An electrical fault location diagnosis method for an engine production line detection device disclosed in this application, the method includes: connecting multiple types of detection devices arranged on the engine production line to construct a detection inner layer, the detection inner layer includes multiple detection data blocks, the detection data block includes detection device information, detection production target, detection synchronization data, the detection synchronization data includes electrical signals and operating parameters of the engine production line, where the electrical signals include current, voltage, power, phase, and the operating parameters include vibration, temperature; based on the detection synchronization data in the detection data block, perform abnormal diagnosis of each device for each detection production target to obtain production abnormal recognition information, and store the production abnormal recognition information into the matching detection data block; based on the detection inner layer, perform data linkage with the detection data block to construct a linkage detection outer layer; through the linkage detection outer layer, respectively perform clustering on the detection data block according to the detection production target and detection device information, and perform linkage recognition according to the production abnormal recognition information according to the clustering relationship to determine conflict information; perform electrical abnormal data identification analysis according to the conflict information to locate electrical faults, and generate fault diagnosis information according to the electrical abnormal data and electrical location faults.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] By deploying multiple types of detection devices and constructing a unified inner detection layer, the data of different devices is integrated into a unified detection data block, including device information, production targets, and synchronization data, which solves the problem of device integration, effectively integrates the data of various types of detection devices, and improves the consistency and integration efficiency of the data; using the unified detection data block and the linked detection outer layer, through data linkage and clustering analysis, fragmented data is integrated and processed, improving the consistency and accuracy of the data; through clustering analysis and linked identification, the production anomaly identification information is associated with the specific detection data block, enabling the rapid positioning and identification of electrical anomaly data, and further analyzing and locating the actual electrical faults, improving the accuracy and efficiency of fault diagnosis; by accurately diagnosing and quickly responding to electrical faults, the downtime and maintenance costs of the production line can be effectively reduced, and the reliability and production efficiency of the production line are improved.

[0007] The above description is only an overview of the technical solution of the present application. In order to be able to more clearly understand the technical means of the present application, it can be implemented in accordance with the content of the specification. And in order to make the above and other purposes, features, and advantages of the present application more obvious and understandable, the following specifically gives the specific implementation manners of the present application. Brief Description of the Drawings

[0008] Figure 1 This is a schematic flow diagram of a method for locating and diagnosing electrical faults in an engine production line detection device provided by an embodiment of the present application;

[0009] Figure 2 This is a schematic flow diagram of clustering the detection data block in a method for locating and diagnosing electrical faults in an engine production line detection device provided by an embodiment of the present application. Detailed Description of the Invention

[0010] By providing a method for locating and diagnosing electrical faults in an engine production line detection device, the embodiment of the present application solves the technical problems in the prior art that various detection data on the engine production line cannot be effectively integrated, the detection data is fragmented and inconsistent, resulting in poor accuracy and efficiency of electrical fault location and diagnosis.

[0011] After introducing the basic principle of the present application, the following will specifically introduce various non-restrictive implementation manners of the present application in conjunction with the drawings of the specification.

[0012] As Figure 1 shown, the embodiment of the present application provides a method for locating and diagnosing electrical faults in an engine production line detection device, and the method includes:

[0013] Connect multiple types of detection devices on the engine production line to construct a detection inner layer. The detection inner layer includes multiple detection data blocks, and each detection data block includes detection device information, detection production targets, and detection synchronization data. The detection synchronization data includes electrical signals and operating parameters of the engine production line. The electrical signals include current, voltage, power, and phase, and the operating parameters include vibration and temperature.

[0014] Install multiple types of detection devices at different positions on the engine production line to achieve real-time monitoring of the production process. Among them, the multiple types of detection devices include electrical signal detection devices, including current sensors, voltage sensors, power meters, phase detectors, etc., and also include operating parameter detection devices, including vibration sensors and temperature sensors.

[0015] According to the technological process of the engine production line, select key positions to deploy detection devices so that they can fully cover the entire production line. The electrical signal detection devices should be arranged at power input and output points, key control nodes, etc., and the operating parameter detection devices should be arranged at key mechanical components, engine test benches, etc.

[0016] Obtain and store real-time data through multiple types of detection devices to construct multiple detection data blocks. Each detection data block contains detection device information, detection production targets, and detection synchronization data. Among them, the detection device information includes device type, location, number, etc., the detection production targets include tasks on the current production line, product types, etc., and the detection synchronization data includes specific electrical signals, including current, voltage, power, phase, and operating parameters, including vibration and temperature. Integrate the multiple detection data blocks to form a detection inner layer, realizing real-time detection of various data on the engine production line and providing accurate data support for subsequent fault diagnosis.

[0017] Based on the detection synchronization data in the detection data block, perform abnormal diagnosis of each detection production target for each device to obtain production anomaly recognition information, and store the production anomaly recognition information in the matching detection data block.

[0018] Utilize the detection synchronization data in the detection data block to perform abnormal diagnosis on each detection production target, identify and store production anomaly information. Specifically, extract the detection synchronization data from the detection data block, including electrical signals and operating parameters, identify abnormal fluctuations in the electrical signals and operating parameters, and establish a time series correlation mapping based on the respective abnormal fluctuation recognition results for abnormal identification to obtain the production anomaly recognition information.

[0019] Store the comprehensively analyzed production anomaly recognition information into the matching detection data block, so that each data block contains complete detection data and anomaly recognition information.

[0020] Furthermore, based on the detection synchronization data in the detection data block, perform abnormal diagnosis of each device for each detection production target to obtain production abnormality recognition information, including:

[0021] Identify abnormal fluctuations based on the electrical signals and operating parameters respectively, and locate abnormal electrical signals and abnormal fluctuation parameters; establish a time-series correlation mapping between the abnormal electrical signals and the abnormal fluctuation parameters; use the abnormal electrical signals as the central data and the abnormal fluctuation parameters as the auxiliary data to perform abnormal recognition and obtain the production abnormality recognition information.

[0022] Analyze the abnormal fluctuations in the electrical signals to locate abnormal electrical signals. Exemplarily, use a threshold-based method to set the normal range threshold of the electrical signals, detect fluctuations exceeding the threshold, and identify abnormal electrical signals; use the same method to analyze the abnormal fluctuations in the operating parameters and locate abnormal fluctuation parameters.

[0023] Use the dynamic time warping algorithm to align the time series of abnormal electrical signals and abnormal operating parameters, analyze the similarity of the time series data, identify the correlation relationship, and through time series similarity analysis, screen out the abnormal electrical signals and abnormal operating parameters with strong correlation. Based on the screening results, establish a time-series correlation mapping between the abnormal electrical signals and the abnormal operating parameters to provide a basis for subsequent abnormal recognition.

[0024] Taking the abnormal electrical signals as the center and combining the abnormal operating parameters as the auxiliary data, perform comprehensive abnormal recognition. Specifically, construct a multivariate abnormal detection model, use the abnormal electrical signals and abnormal operating parameters as input variables, take the diagnosis result of the abnormal electrical signals as the center, perform auxiliary analysis of the abnormal electrical signals through the abnormal fluctuation parameters, identify the causes and impacts of the abnormal fluctuations, determine the abnormal situations in the production process, and based on the comprehensive analysis results, generate production abnormality recognition information, including the types of abnormal electrical signals, abnormal fluctuation parameters, correlation relationships, abnormal causes, etc., to provide a basis for production optimization and fault troubleshooting.

[0025] Furthermore, establishing the time-series correlation mapping between the abnormal electrical signals and the abnormal fluctuation parameters includes:

[0026] Use the dynamic time warping algorithm to perform time-series alignment on the abnormal electrical signals and the abnormal fluctuation parameters, and analyze the similarity of the time series data; according to the similarity of the time series data, screen out the consistent time alignment relationships of abnormal peaks, and establish the time-series correlation mapping between the abnormal electrical signals and the abnormal fluctuation parameters.

[0027] The dynamic time warping algorithm is a method for measuring the similarity of time series. It allows time series to be non-linearly aligned on the time axis to find the best matching path between the two series.

[0028] Obtain the time series data of abnormal electrical signals and abnormal fluctuation parameters, and calculate the distance matrix between the two time series. The distance measurement methods include Euclidean distance, Manhattan distance, etc. The elements of the distance matrix represent the distances between the two series values at different time points.

[0029] Based on the distance matrix, use the dynamic programming method to find the minimum cumulative distance path from the start to the end. This path represents the best alignment of the two time series on the time axis. Dynamic programming finds the minimum cumulative distance at the current point by recursively comparing the minimum distances of the previous step. Starting from the last point of the minimum cumulative distance matrix, backtrack to find the entire alignment path, which represents the best match of the two series on the time axis. According to the found alignment path, analyze the similarity of the two time series. The cumulative distance or average distance can be calculated to quantify the degree of similarity. A smaller distance indicates that the two are more consistent or similar in time.

[0030] In the time series of abnormal electrical signals and abnormal fluctuation parameters, identify and mark the abnormal peak points or fluctuation extreme values. According to the abnormal peak points or fluctuation extreme values, screen out their consistent alignment relationships on the time axis, that is, find the time points when significant changes or fluctuations occur simultaneously in the abnormal electrical signals and abnormal fluctuation parameters. According to the screened consistent time alignment relationships, establish the time series correlation mapping between the abnormal electrical signals and abnormal fluctuation parameters, which means that at these time points, the changes in the abnormal electrical signals can be correlated with the changes in the abnormal fluctuation parameters.

[0031] Furthermore, taking the abnormal electrical signal as the central data and the abnormal fluctuation parameter as the auxiliary data, perform anomaly identification, including:

[0032] Construct a multivariate anomaly detection model. Taking the abnormal electrical signal and the abnormal fluctuation parameter as input variables, with the diagnosis result of the abnormal electrical signal as the center, perform auxiliary analysis of the abnormal electrical signal through the abnormal fluctuation parameter, and output the production anomaly identification information, which is output with the abnormal diagnosis result of the electrical signal.

[0033] The multivariate anomaly detection model aims to identify the abnormal situations occurring in the production process by comprehensively analyzing multiple related variables, including abnormal electrical signals and abnormal fluctuation parameters. This model utilizes the mutual relationships between multiple input variables to improve the accuracy and reliability of anomaly detection.

[0034] Construct a multivariate anomaly detection model. Specifically, select an appropriate multivariate anomaly detection model, such as multiple linear regression, support vector machine, neural network, etc. These models can process multi-dimensional data and identify abnormal patterns. Use historical data or experimental data to train the model. The training data includes normal operation data and known abnormal data so that the model can learn to distinguish normal and abnormal situations. Use a validation data set to validate the model and evaluate its performance. Use a test data set to further test the accuracy and reliability of the model. Finally, obtain a multivariate anomaly detection model that meets the preset requirements.

[0035] The model generates preliminary diagnostic results through the input abnormal electrical signals. These results identify potential abnormal signals. Use abnormal fluctuation parameters for auxiliary analysis. By comparing the abnormal electrical signals and abnormal fluctuation parameters, abnormal situations can be identified more accurately. For example, if the current is abnormal and accompanied by a temperature increase, the possibility of an electrical fault can be further confirmed. Integrate the above analysis results and output production anomaly identification information, which includes the specific diagnostic results of the abnormal electrical signals and the support information obtained from the auxiliary analysis.

[0036] Based on the detection inner layer, perform data linkage with the detection data block to construct a linkage detection outer layer.

[0037] Through the linkage of the detection data block, integrate the data of each detection point to construct a higher-level linkage detection outer layer for comprehensive monitoring and fault diagnosis of the production line. Specifically, establish the association relationship between each detection data block, and define the linkage rules of the data block according to the actual situation of the production line. The linkage rules include data synchronization, abnormal linkage, parameter association, etc., and perform data linkage according to the linkage rules.

[0038] Based on the data linkage of the detection data block, construct a linkage detection outer layer. The linkage detection outer layer includes multiple linkage detection nodes. Each node corresponds to one or more detection data blocks. Each linkage detection node has independent data processing and analysis capabilities and can be linked with other nodes.

[0039] Through the linkage detection outer layer, cluster the detection data blocks according to the detection production target and detection equipment information respectively, and perform linkage identification according to the production anomaly identification information according to the clustering relationship to determine the conflict information.

[0040] The detection data blocks are clustered using a clustering algorithm to form a clustering relationship. Specifically, the detection data blocks are clustered according to the detection production target to form a horizontal clustering relationship, which means that the detection data blocks with the same target will be grouped together; according to the detection device information, such as device type, parameter indicators, etc., the detection data blocks are clustered to form a vertical clustering relationship, which means that the detection data blocks of the same type of device or the same parameter indicator will be grouped together. Based on the horizontal and vertical clustering relationships, a comprehensive clustering relationship of the detection data blocks is established to form a multi-dimensional and multi-level clustering structure for convenient linkage analysis.

[0041] According to the horizontal clustering relationship, the similarity of the production anomaly recognition information is compared to identify the similarity conflict nodes and locate the horizontal conflict electrical data; according to the vertical clustering relationship, the production anomaly recognition information is compared for similarity with the same parameter indicators to identify and locate the vertical conflict electrical data. The horizontal conflict electrical data and the vertical conflict electrical data are overlapped and fused to determine the final conflict information. By fusing these two different-dimensional conflict data, the electrical faults in the production line can be more accurately located and analyzed. For example, if a certain electrical fault appears under a specific production target and a specific device type at the same time, it can be determined that this fault has a high credibility and importance.

[0042] Furthermore, as Figure 2 shown, the clustering of the detection data blocks according to the detection production target and the detection device information respectively includes:

[0043] The detection data blocks are clustered according to the detection production target to determine the horizontal clustering relationship, which is the clustering relationship of the same target; the detection data blocks are clustered according to the device type of the detection device information to determine the vertical clustering relationship, which is the clustering relationship of the same parameter indicators; the clustering relationship is formed based on the horizontal clustering relationship and the vertical clustering relationship. The detection data blocks are collected, and each detection data block contains detection device information, detection production target, and detection synchronization data.

[0044] The detection production targets in all the detection data blocks are classified. The detection production target can be to produce a certain specific model of engine, a certain stage of the production process, or a specific production task. The detection data blocks with the same detection production target are clustered together to form a horizontal clustering relationship, which means that all the data blocks related to the same production target are grouped in one set. For example, assuming the production target is to produce a certain specific model of engine, then all the detection data blocks related to this model will be clustered together.

[0045] Classify the device types in all detected data blocks. The device types can be current sensors, temperature sensors, vibration sensors, etc. Cluster the detected data blocks with the same device type together to form a vertical clustering relationship. The vertical clustering relationship means that all data blocks related to the same parameter index are grouped together. For example, assume the detection device is a sensor for measuring current, then all data blocks related to current measurement will be clustered together.

[0046] Combine the horizontal clustering relationship and the vertical clustering relationship to form a comprehensive clustering relationship network. In this way, each detected data block can consider both the production target and the device type, thus realizing multi-dimensional data analysis and fault diagnosis.

[0047] Furthermore, perform linkage recognition according to the clustering relationship based on the production anomaly recognition information to determine the conflict information, including:

[0048] According to the horizontal clustering relationship, perform similarity comparison on the production anomaly recognition information to determine the similarity conflict nodes and locate the horizontal conflict electrical data; according to the vertical clustering relationship, perform similarity comparison of the same parameter index on the production anomaly recognition information to identify and locate the vertical conflict electrical data; use the horizontal conflict electrical data and the vertical conflict electrical data for overlapping fusion to determine the conflict information.

[0049] Perform similarity comparison on the production anomaly recognition information in the same horizontal clustering relationship. The similarity comparison can adopt various algorithms, such as Euclidean distance, cosine similarity, etc., to calculate the similarity between each abnormal data. For example, if the current abnormal data in two detected data blocks has a high similarity, it can be considered that they have a potential association. According to the similarity comparison result, determine the similarity conflict nodes. The similarity conflict nodes refer to the abnormal data nodes with a similarity higher than a certain threshold in the same horizontal clustering relationship. For example, if the voltage abnormal data in multiple detected data blocks has a high similarity, these data will be marked as similarity conflict nodes.

[0050] For the determined similarity conflict nodes, further analyze their electrical data to locate the horizontal conflict electrical data. The horizontal conflict electrical data refers to the electrical data with similar abnormal characteristics under the same production target. For example, by analyzing the similarity conflict nodes, the abnormal data of specific parameters such as current, voltage, and power can be located.

[0051] Perform similarity comparison of the same parameter indicators for the production anomaly recognition information under the same longitudinal clustering relationship. For example, if the current anomaly data in multiple detection data blocks have highly similar numerical characteristics within the same time period, this may indicate the existence of a common electrical fault point. According to the results of the similarity comparison, identify and locate the longitudinal conflict electrical data. Longitudinal conflict electrical data refers to electrical data with similar anomaly characteristics under the same equipment type. For example, by analyzing the current anomaly data with high similarity, the problems existing in specific current sensors or current monitoring modules can be located.

[0052] Overlap and fuse the horizontal conflict electrical data and the longitudinal conflict electrical data. Specifically, conduct cross-analysis on the horizontal conflict electrical data and the longitudinal conflict electrical data to find the intersection part between the two, analyze the specific electrical characteristics and anomaly situations of the intersection part, and determine the co-existing conflict information, such as possible fault causes or the time and location of anomaly occurrence. For example, if a certain electrical fault appears simultaneously under a specific production target and a specific equipment type, then it can be determined that this fault has a relatively high credibility and importance.

[0053] Furthermore, use the horizontal conflict electrical data and the longitudinal conflict electrical data for overlap fusion to determine the conflict information, including:

[0054] Collect the historical data of different signals of the same equipment, and fit the horizontal anomaly probability function. The expression of the horizontal anomaly probability function is: where X is the observation vector composed of different signals X = [x 1 , x 2, …, x n , μ is the mean vector μ = [μ 1 , μ 2, …, μ n , A is the covariance matrix, the dimension of the covariance matrix is n×n, |A| is the determinant of the covariance matrix, A -1 is the inverse matrix of the covariance matrix; collect the anomaly data of the same type of signals, and fit the longitudinal anomaly probability function. The expression of the longitudinal anomaly probability function is: μ i is the mean value of the electrical signal for the current operation, σ is the standard deviation of the electrical signal for the current operation, x i is the value of the currently observed electrical signal; use the horizontal anomaly probability and the longitudinal anomaly probability as the fusion coefficients for overlap fusion to obtain the conflict information.

[0055] Specifically, collect the historical data of different signals of the same equipment, and fit the horizontal anomaly probability function. The expression of the horizontal anomaly probability function is:

[0056] where (2π)n / 2 |A| 1 / 2 is part of the normalization of the probability density function to ensure that the sum of probabilities is 1, (2π) n / 2 is the normalization constant of the Gaussian distribution, |A| 1 / 2 , which is the square root of the determinant of the covariance matrix A and is used to consider the covariance structure between different signals.

[0057] is the exponential term of the Gaussian distribution, which describes the square of the Mahalanobis distance between the observation vector X and the mean vector μ, multiplied by the inverse of the covariance matrix. It measures the deviation between the observation vector and the mean vector and takes into account the correlation and variance between signals.

[0058] This probability density function is used for anomaly detection and pattern recognition of multi-dimensional data. Especially when the relationship between signals is not just individual, it can more comprehensively evaluate the overall anomaly degree of multiple signals, rather than considering the anomaly of each signal separately. By adjusting the mean vector μ and the covariance matrix A, it can adapt to different data distributions and anomaly detection requirements. A smaller Mahalanobis distance means that the observation vector X is closer to the mean vector μ, and vice versa, indicating a greater possibility of anomaly.

[0059] Specifically, collect anomaly data of similar signals and fit the longitudinal anomaly probability function. The expression of the longitudinal anomaly probability function is as follows:

[0060] where is used to normalize the probability density function to ensure that the sum of probabilities is 1, is the normalization constant of the Gaussian distribution, and σ is the standard deviation of the current electrical signal being operated on, which is used to measure the range of variation of the signal around the mean μ i surrounding. is the exponential term of the Gaussian distribution, which describes the deviation degree between the observed electrical signal value x i and the mean μ i between.

[0061] This probability density function is used for anomaly detection and pattern recognition of a single electrical signal. By comparing the mean μ i and the standard deviation σ, it can be evaluated whether the current observed signal x i deviates from the normal operating range, thereby identifying anomalies. A smaller probability density value indicates that the difference between the observed signal x i and the mean μ i is small, indicating that the signal is in a normal operating state. On the contrary, a larger probability density value may indicate that the observed signal x i has anomalies or sudden changes.

[0062] For each observed signal vector, the lateral anomaly probability is calculated according to the lateral anomaly probability function. For each individual electrical signal, the longitudinal anomaly probability is calculated according to the longitudinal anomaly probability function. The lateral anomaly probability and the longitudinal anomaly probability are used as weight coefficients for weighted averaging, so that the anomaly situations in different directions can be integrated, and the overall severity and impact of the conflict information can be determined according to their respective weights. According to the result of overlapping fusion, the conflict information is determined. A higher overlapping fusion value indicates a more serious or extensive conflict, and corresponding repair measures need to be taken.

[0063] Based on the conflict information, electrical anomaly data identification and analysis are carried out to locate electrical faults, and fault diagnosis information is generated according to the electrical anomaly data and the located electrical faults.

[0064] Extract electrical signal data related to the conflict information, including current, voltage, power, phase, etc., and construct fault diagnosis channels for each electrical signal. This fault diagnosis channel is a threshold corresponding to the fault type set for different electrical signals or an operation analysis model, which can be a fault identification table trained or fitted through experimental data or empirical data, similar to a fuzzy control table, or an operation model. The conflict information is imported into each electrical signal fault diagnosis channel for matching analysis, and matching is carried out according to the quantity and value of the electrical signals of the conflict information to determine the final fault result, that is, the electrical fault.

[0065] Based on the anomaly data and the conflict information, locate the specific electrical fault, determine the root cause and location of the fault. Exemplarily, a multivariate analysis method is adopted. Using a multivariate analysis method, such as multivariate regression analysis, factor analysis, etc., analyze the relationship between different electrical signals, locate the fault source, and combine the electrical signals and operating parameters for comprehensive analysis to determine the specific location of the fault.

[0066] According to the electrical anomaly data and the located electrical faults, generate fault diagnosis information, including fault type, occurrence time, occurrence location, influence range, fault cause, recommended treatment measures, etc. The fault diagnosis information details the characteristics and impacts of the fault, providing a reference for fault handling. This can not only timely detect and solve electrical faults in production, but also provide strong data support and decision-making basis for the maintenance of the production line, ensuring the stable operation of the production line.

[0067] Furthermore, based on the conflict information, electrical anomaly data identification and analysis are carried out to locate electrical faults, including:

[0068] Construct fault diagnosis channels for each electrical signal, import the conflict information into each electrical signal fault diagnosis channel for matching analysis, and determine the electrical fault.

[0069] Obtain different types of electrical signals, including current, voltage, power, phase, etc. Design independent fault diagnosis channels for each electrical signal. Each channel is designed based on the specific fault types and diagnosis models of the signal. Construct a comprehensive fault diagnosis channel that can comprehensively analyze multiple electrical signals.

[0070] Specifically, for different electrical signals, set corresponding fault types and thresholds. These thresholds can be set through experimental data or empirical data. Select and train appropriate analysis models, such as rule-based models, machine learning-based models, etc., to establish each fault diagnosis channel. Import the electrical signal data in the conflict information into the corresponding fault diagnosis channel. In a single signal diagnosis channel, analyze the imported electrical signals according to the preset thresholds and diagnosis models, identify the data points that exceed the thresholds or are abnormal, and determine the corresponding fault types. Collect the matching analysis results of each fault diagnosis channel, conduct a comprehensive evaluation, and make a comprehensive judgment on the matching analysis results according to the quantity and value of the electrical signals to determine the final fault type and location, and determine the electrical fault.

[0071] In summary, the electrical fault location and diagnosis method of an engine production line detection device provided by the embodiments of the present application has the following technical effects:

[0072] By arranging multiple types of detection devices and constructing a unified detection inner layer, integrating the data of different devices into a unified detection data block, including device information, production targets, and synchronization data, the problem of device integration is solved, effectively integrating the data of various types of detection devices, and improving the consistency and integration efficiency of the data; using the unified detection data block and the linked detection outer layer, through data linkage and clustering analysis, integrating and processing fragmented data, improving the consistency and accuracy of the data; through clustering analysis and linked identification, associating the production anomaly identification information with the specific detection data block, being able to quickly locate and identify electrical anomaly data, and further analyze and locate the actual electrical fault, improving the accuracy and efficiency of fault diagnosis; by accurately diagnosing and quickly responding to electrical faults, the downtime and maintenance costs of the production line can be effectively reduced, and the reliability and production efficiency of the production line are improved.

[0073] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for locating and diagnosing electrical faults in engine production line testing equipment, characterized in that: The electrical fault location diagnosis method of the engine production line detection equipment comprises: Connect the engine production line to deploy multiple types of testing equipment and build a testing inner layer, which includes multiple testing data blocks. The testing data blocks include testing equipment information, testing production targets, and testing synchronization data. The testing synchronization data includes electrical signals and operating parameters of the engine production line, wherein the electrical signals include current, voltage, power, and phase, and the operating parameters include vibration and temperature; Based on the detection synchronization data in the detection data block, each device abnormality diagnosis is performed on each detection production target to obtain production abnormality identification information, and the production abnormality identification information is stored in the matching detection data block; Based on the detection inner layer, data linkage is performed with the detection data block to construct a linkage detection outer layer; Through the linkage detection outer layer, the detection data blocks are clustered according to the detection production target and the detection equipment information, and linkage identification is performed according to the production abnormality identification information according to the clustering relationship to determine the conflict information; The electrical abnormality data is identified and analyzed based on the conflict information to locate the electrical fault, and fault diagnosis information is generated based on the electrical abnormality data and the electrical fault location.

2. The electrical fault location diagnosis method for engine production line detection equipment according to claim 1, characterized in that: The clustering of the detection data blocks according to the detection production target and the detection equipment information respectively includes: Clustering the detection data blocks according to the detection production target to determine a horizontal clustering relationship, where the horizontal clustering relationship is a clustering relationship with the same target; Clustering the detection data blocks according to the device type of the detection device information to determine a vertical clustering relationship, where the vertical clustering relationship is a clustering relationship with the same parameter index; The clustering relationship is established based on the horizontal clustering relationship and the vertical clustering relationship.

3. The electrical fault location diagnosis method for engine production line detection equipment according to claim 2, characterized in that: The step of performing linkage identification according to the production abnormality identification information based on the cluster relationship to determine the conflict information includes: According to the horizontal clustering relationship, the production abnormality identification information is compared for similarity, similarity conflict nodes are determined, and horizontal conflict electrical data are located; According to the vertical clustering relationship, the production abnormality identification information is compared with the parameter index similarity to identify and locate the vertical conflicting electrical data; The conflict information is determined by overlapping and fusing the horizontal conflict electrical data with the vertical conflict electrical data.

4. The electrical fault location diagnosis method for engine production line detection equipment according to claim 1, characterized in that: Based on the detection synchronization data in the detection data block, each device abnormality diagnosis is performed on each detection production target to obtain production abnormality identification information, including: Respectively identify abnormal fluctuations according to the electrical signals and operating parameters, and locate abnormal electrical signals and abnormal fluctuation parameters; Establishing a time-series correlation mapping between the abnormal electrical signal and the abnormal fluctuation parameter; The abnormal electrical signal is used as the central data, and the abnormal fluctuation parameter is used as the auxiliary data to perform abnormality identification to obtain the production abnormality identification information.

5. The electrical fault location diagnosis method for engine production line detection equipment according to claim 4, characterized in that: Establishing a time series correlation mapping between the abnormal electrical signal and the abnormal fluctuation parameter includes: Using a dynamic time planning algorithm to perform time series alignment on the abnormal electrical signal and the abnormal fluctuation parameter, and analyzing the similarity of time series data; According to the similarity of the time series data, the consistent time alignment relationship of the abnormal peak values ​​is screened, and a time series correlation mapping between the abnormal electrical signal and the abnormal fluctuation parameter is established.

6. The electrical fault location diagnosis method for engine production line testing equipment according to claim 4, characterized in that: Using the abnormal electrical signal as the central data and the abnormal fluctuation parameter as the auxiliary data, abnormality identification is performed, including: A multivariate anomaly detection model is constructed, with the abnormal electrical signal and the abnormal fluctuation parameter as input variables, and the diagnosis result of the abnormal electrical signal as the center. The abnormal electrical signal is auxiliary analyzed through the abnormal fluctuation parameter, and the production anomaly identification information is output. The production anomaly identification information is output as the electrical signal abnormality diagnosis result.

7. The electrical fault location diagnosis method for engine production line testing equipment according to claim 3, characterized in that: Overlapping and fusing the horizontal conflict electrical data and the vertical conflict electrical data to determine the conflict information includes: Collect historical data of different signals of the same device and fit the lateral abnormality probability function, the expression of which is: Among them, X is the observation vector composed of different signals X = [x1, x 2, …,x n ], μ is the mean vector μ=[μ1,μ 2, …,μ n ], A is the covariance matrix, the dimension of the covariance matrix is ​​n×n, |A| is the determinant of the covariance matrix, A -1 is the inverse of the covariance matrix; Collect abnormal data of similar signals and fit the longitudinal abnormal probability function, the expression of which is: μ i is the mean value of the electrical signal of the current operation, σ is the standard deviation of the electrical signal of the current operation, x i is the current observed electrical signal value; The conflict information is obtained by performing overlapping fusion using the horizontal anomaly probability and the vertical anomaly probability as fusion coefficients respectively.

8. The electrical fault location diagnosis method for engine production line testing equipment according to claim 7, characterized in that: According to the conflict information, electrical abnormality data identification and analysis are performed to locate electrical faults, including: Each electrical signal fault diagnosis channel is constructed, and the conflict information is imported into each electrical signal fault diagnosis channel for matching analysis to determine the electrical fault.

Images