Busway Fault Monitoring Method and System

Abstract

Busbar trunking fault monitoring method and system, relating to the field of data processing. The method includes: obtaining a first measurement value data set; calculating a difference data set and a difference value of the difference data set from the first measurement value data set; calculating an original interference evaluation value of each measurement value in the first measurement value data set based on the interval length of each measurement node, and then calculating a target interference evaluation value; using the target interference evaluation value to correct the first measurement value data set to obtain a second measurement value data set; dynamically segmenting the second measurement value data set to obtain second measurement value data sub-blocks; determining the sliding area length according to the target interference evaluation value to obtain a third measurement value data set; uploading the third measurement value data set to the server. In the process of removing interference from data, the busbar trunking fault monitoring method and system of the present application introduce the factor of the interval length of each measurement node, making the interference removal effect of the measurement data more accurate.

Description

Technical Field

[0001] This application relates to the field of data processing, and particularly to a method and system for monitoring busbar trunking unit faults. Background Art

[0002] In a busbar trunking unit transmission system, multiple busbar trunking units form a power transmission line. An industrial data monitoring module can be set on the power transmission line to collect data on operating parameters such as current and voltage of the power transmission line. Existing busbar trunking unit fault monitoring systems only consider processing the measurement data collected by a single industrial data monitoring module, and only perform simple filtering on the measurement data. This means that the greater the degree of interference in the measurement data collected at this position, the less accurate the subsequent fault warning obtained based on the analysis of this measurement data. Summary of the Invention

[0003] To solve the above technical problems, this application provides a method and system for monitoring busbar trunking unit faults. During the data anti-interference process, the factor of the interval length between each measurement node is introduced, making the anti-interference effect of the measurement data more accurate and the subsequent fault warning more accurate.

[0004] The technical solution of this application is as follows:

[0005] In a first aspect, this application provides a method for monitoring busbar trunking unit faults, which is applied to an industrial data gateway. The method includes:

[0006] Obtain the first measurement value dataset of the industrial data monitoring modules of each measurement node on the same transmission line; there are multiple transmission lines. The same transmission line refers to a section of the transmission line without a convergence point or a divergence point. Each transmission route is composed of multiple busbar trunking units. The industrial data monitoring modules on the same transmission line are defined as the same group;

[0007] Calculate the differential dataset of the preset range area corresponding to each measurement value and the difference value of this differential dataset from the first measurement value dataset; where the length of the preset range area is much smaller than the length of the first measurement value dataset, and the length of the preset range area is a positive integer;

[0008] Based on the interval length between each measurement node, calculate the original interference evaluation value of each measurement value in the first measurement value dataset; according to the difference value corresponding to each measurement value, calculate the regularization term of each measurement value and then multiply it by the corresponding original interference evaluation value to obtain the interference influence value of each measurement value; perform normalization processing on the interference influence values of each measurement value to obtain the target interference evaluation value of each measurement value in the first measurement value dataset;

[0009] The first measurement value dataset is corrected using the target interference evaluation values of each measurement value to obtain a second measurement value dataset; the second measurement value dataset is dynamically segmented to obtain second measurement value sub-blocks of each measurement node;

[0010] The sliding region length is determined according to the corresponding target interference evaluation value, and the second measurement value sub-blocks are subjected to a first measurement value cumulative average. After the second measurement value sub-blocks within the sliding region are subjected to a second measurement value cumulative average, a third measurement value dataset is obtained; wherein, the sliding region length is a positive integer;

[0011] Taking the third measurement value dataset as the final measurement value dataset, the final measurement value dataset is uploaded to the server for the server to monitor the operating state of the busway for faults.

[0012] The busway fault monitoring method of the present application groups the industrial data monitoring modules at different measurement node positions. The industrial data monitoring modules on the same transmission line are in the same group, that is, the measurement nodes with similar measurement values are considered. Using the interval length between the measurement nodes, the algorithm is optimized to enhance the accuracy and rationality of the estimation of the interference evaluation value; in addition, the present application designs a preset range area corresponding to each measurement value, and uses the difference value of the differential data set corresponding to each measurement value to dynamically adjust the interference evaluation value, enhancing the adaptability of the optimization algorithm to the local area and monitoring the minute changes in the measurement values; based on the corrected measurement data for dynamic segmentation, grouping the measurement data with close interference evaluation values together helps to improve the effect of removing interference from the measurement data and make the subsequent fault warning more accurate.

[0013] In a preferred embodiment, calculating the differential data set corresponding to each measurement value and the difference value of the differential data set from the first measurement value data set includes:

[0014] Taking the measurement values in the first measurement value data set as key points, defining a preset number before and after the key points as the preset range area corresponding to each measurement value;

[0015] Calculating the difference between every two adjacent measurement values within the preset range area corresponding to each measurement value to obtain the differential data set of the preset range area corresponding to each measurement value and the difference value of the differential data set, thereby obtaining the differential data set of the preset range area corresponding to each measurement value and the difference value of the differential data set; the differential data set is a first-order data sequence.

[0016] It should be noted that, by means of the local characteristics of the differential data set relative to the entire first measurement value data set, the present application reveals the local change characteristics of the data, captures the subtle changes in the measurement data more acutely, and the outlier characteristics of the first-order difference sequence of the differential data set also help to eliminate the interference in the data and improve the accuracy of monitoring the operating parameters of the busbar trunking under complex environments.

[0017] In a preferred embodiment, based on the interval lengths of the respective measurement nodes, the original interference evaluation values of the respective measurement values in the first measurement value data set are calculated, including:

[0018] Taking an industrial data monitoring module as the first measurement node, and the industrial data monitoring modules other than the first measurement node as the second measurement nodes, obtaining the measurement value of the first measurement node and the measurement values of the second measurement nodes at the same time point, and calculating the difference between the two measurement values;

[0019] Calculating the sub-interval metric value between the first measurement node and the second measurement node and the total interval metric value between the first measurement node and the second measurement node;

[0020] Calculating the ratio of the sub-interval metric value to the total interval metric value, and using the ratio as the interval metric evaluation value of the measurement value;

[0021] Performing a product summation of the measurement value differences of the respective measurement values and the interval metric evaluation values to obtain the original interference evaluation value of each measurement value.

[0022] It should be noted that, by calculating the interference evaluation value for the differences between the measurement value data of the respective measurement nodes at the same time point, the greater the difference between the measurement value of the current measurement node and the measurement values of other measurement nodes, the greater its interference evaluation value. For example, the measurement values of two measurement nodes with closer intervals should be more similar. Therefore, the interval length between the two measurement nodes can be used as the weight of the difference, and then the interference evaluation value of the measurement value is obtained by weighted averaging to improve the accuracy of calculating the interference evaluation value.

[0023] In a preferred embodiment, performing a dynamic segmentation on the second measurement value data set to obtain the second measurement value data sub-blocks of the respective measurement nodes, including:

[0024] Performing a pre-segmentation process on the second measurement value data set to obtain the original second measurement value data set sub-blocks; the pre-segmentation process is performed based on a preset segmentation number, and the preset segmentation number is a positive integer;

[0025] Calculating the error pre-estimation values of the original second measurement value data set sub-blocks, and summing up the error pre-estimation values of all the original second measurement value data set sub-blocks to obtain the error result of the second measurement value data set under this preset segmentation number;

[0026] Taking the preset number of segments as the horizontal axis and the error result as the vertical axis, create an error change graph for different preset numbers of segments, determine the position of the inflection point based on the shape of the error change graph, and determine the target sub-block length according to the position of the inflection point; where the target sub-block length is a positive integer;

[0027] Based on the target sub-block length, perform a secondary segmentation on the second measurement value data set to obtain the final second measurement value data set sub-blocks, and use the final second measurement value data set sub-blocks as the second measurement value data sub-blocks corresponding to the measurement nodes, so as to obtain the second measurement value data sub-blocks of each measurement node.

[0028] It should be noted that in this application, by combining the optimal solution method and the elbow method, the loss function is minimized by continuously iterating the number of data segments, so that the measurement values in each second measurement value data sub-block are close in size, and the interference level is also stabilized at a certain level, which is more adaptable to the complex changes in the measurement values in different time periods in the measurement data being different.

[0029] In a preferred embodiment, calculating the error prediction value of the original second measurement value data set sub-block includes:

[0030] Calculate the mean value of the measurement values in the original second measurement value data set sub-block to obtain the predicted value of the original second measurement value data set sub-block;

[0031] Calculate the difference between the measurement value and the predicted value in the original second measurement value data set sub-block and perform a sum of squares to obtain the sum of squared residuals of the original second measurement value data set sub-block.

[0032] In a preferred embodiment, calculating the error prediction value of the original second measurement value data set sub-block includes:

[0033] Calculate the mean absolute error value of the original second measurement value data set sub-block, and use this mean absolute error value as the error prediction value of the original second measurement value data set sub-block.

[0034] In a preferred embodiment, calculating the error prediction value of the original second measurement value data set sub-block includes:

[0035] Calculate the root mean square error value of the original second measurement value data set sub-block, and use this root mean square error value as the error prediction value of the original second measurement value data set sub-block.

[0036] In a preferred embodiment, determining the sliding region length according to the corresponding target interference evaluation value includes:

[0037] Based on the target interference evaluation value of the second measurement value data set sub-block, calculate the second interference evaluation mean value of the second measurement value data set sub-block;

[0038] After calculating the product of the initial sliding region length and the second interference evaluation mean value and adding it to the initial sliding region length, the expected sliding region length is obtained.

[0039] Take the maximum positive integer value of the expected sliding region length to obtain the target sliding region length, and use this target sliding region length as the final sliding region length.

[0040] It should be noted that in this application, the size of the window is dynamically adjusted through the expectation of the interference evaluation value in the data sub-block, and the second measurement value data set is filtered based on the adjusted target sliding window (i.e., the sliding region length), completing the data preprocessing of the first measurement value data set of the industrial data monitoring module of each measurement node.

[0041] In a second aspect, this application also provides a busbar trunking fault monitoring system, including a server, an industrial data gateway, and an industrial data monitoring module. The industrial data monitoring module is used to send measurement data to the industrial data gateway, the industrial data gateway executes the busbar trunking fault monitoring method described in any one of the above, and the server is used to perform fault monitoring on the operating state of the busbar trunking according to the final measurement value data set.

[0042] This busbar trunking fault monitoring system groups the industrial data monitoring modules at different measurement node positions. The industrial data monitoring modules on the same transmission line are in the same group, that is, the measurement nodes with similar measurement values are considered. Using the interval length between the measurement nodes, the algorithm is optimized to enhance the accuracy and rationality of the estimation of the interference evaluation value. In addition, this application designs a preset range area corresponding to each measurement value, and uses the difference value of the differential data set corresponding to each measurement value to dynamically adjust the interference evaluation value, enhancing the adaptability of the optimization algorithm to the local area and monitoring the small changes in the measurement values. Based on the corrected measurement data for dynamic segmentation, the measurement data with similar interference evaluation values are segmented together, which helps to improve the effect of removing interference from the measurement data and make the subsequent fault warning more accurate. Description of the Drawings

[0043] Figure 1 It is a schematic structural diagram of the busbar trunking fault monitoring system according to an embodiment of this application.

[0044] Figure 2 It is a flowchart of the steps of the busbar trunking fault monitoring method according to an embodiment of this application. Detailed Embodiments

[0045] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0046] In the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other. For the convenience of description, concepts such as "first", "second", etc. mentioned in the present application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions executed by these devices, modules or units. The modifications of "one" and "multiple" mentioned in the present application are illustrative rather than restrictive. Those skilled in the art should understand that unless clearly specified otherwise in the context, it should be understood as "one or more".

[0047] The following will describe in detail the specific implementation manners of the present application with reference to the accompanying drawings.

[0048] The power transmission line includes multiple busbars. As Figure 1 shown, in the embodiment of the present application, an industrial data monitoring module 1 is set at the connection of adjacent busbars 3 in this busbar fault monitoring system. Since the lengths and specifications of the selected busbars 3 in the same section of the power transmission line are basically the same, the industrial data monitoring modules can be evenly distributed throughout the power transmission line, that is, the measurement nodes of the same section of the power transmission line are evenly distributed. The industrial data monitoring module 1 is connected to the industrial data gateway 2 to transmit the collected measurement values to the industrial data gateway 2 in real time. After preprocessing the collected measurement data, the industrial data gateway 2 uploads the final measurement data after preprocessing to a server (not shown in the figure), and the server performs fault analysis and early warning on the operating status of the busbar 3 according to the final measurement data. The industrial data monitoring modules can be numbered [1, 2, 3..., m] in sequence from the main line to the branch line according to the busbar 3 wiring plan. The same section of the transmission line refers to a section of the transmission line without a convergence point or a shunt point, and the industrial data monitoring modules of the same section of the transmission line can be identified as the same group.

[0049] Referring to Figure 2 , the busbar fault monitoring method includes the following steps S100 to step 700:

[0050] Step S100: Obtain the first measurement value dataset of the industrial data monitoring modules at each measurement node of the same transmission line; there are multiple transmission lines, and the same transmission line refers to a section of the transmission line without a convergence point or a shunt point. Each transmission line is composed of multiple busbars, and the industrial data monitoring modules of the same transmission line are defined as the same group.

[0051] In the embodiment of the present application, when the industrial data monitoring module collects current data, the measurement data collected by each industrial data monitoring module can be described as , indicating the m th industrial data monitoring module at Measurement values at time points. The industrial data monitoring module can also monitor voltage data. The number and interval of the industrial data monitoring modules can also be arranged according to the actual needs of the system. For example, an industrial data monitoring module is set for every five busbar troughs, etc. The embodiment of the present application does not limit the number of industrial data monitoring modules.

[0052] Step S200: Calculate the difference data set of the preset range area corresponding to each measurement value and the difference value of the difference data set from the first measurement value data set; wherein, the length of the preset range area is much smaller than the length of the first measurement value data set, and the length of the preset range area is a positive integer.

[0053] In the above step S100, the collected measurement data (that is, the first measurement value data set) is time-series data. Usually, the interference points in the time-series data often show as outliers, which have nothing to do with the changes of the surrounding data. The measurement value data in the preset range area and the entire measurement value data set are in a local and overall relationship. In the embodiment of the present application, the difference data set is a first-order data sequence, which can be used to characterize the difference between two adjacent measurement values. Specifically, taking the measurement value in the first measurement value data set as the key point, define a preset number before and after the key point as the preset range area corresponding to each measurement value; if the length of the first measurement value data set is M, the preset number can be set to N, and N is much smaller than M. It can also be understood as the relationship between the whole and the part. Taking any measurement value in the first measurement value data set as the key point, obtain N measurement values before and after for subsequent calculation, then the length of the preset range area is 2N + 1, and calculate the difference between every two adjacent measurement values within the preset range area , obtain 2N differences within the preset range area, and the set of these 2N differences is the first-order difference data of the preset range area corresponding to the measurement value. Taking the set of these 2N differences as the difference data set W, it can be used to represent the difference data set of the measurement value of the m-th industrial data monitoring module within the preset range area at the time point. Simply put, it is the first-order difference sequence of the local area of the measurement value. The difference data sets of other measurement values are calculated according to the above steps in the same way, so as to obtain the difference data sets of the preset range areas corresponding to each measurement value. Through the local characteristics of the difference data set relative to the entire first measurement value data set, the local change characteristics of the data can be revealed, and the subtle changes of the measurement data can be captured more sensitively. The outlier characteristics of the first-order difference sequence of the difference data set are also helpful to eliminate the interference in the data and improve the accuracy of monitoring the operating parameters of the busbar trough in a complex environment. The outlier property of the first-order difference data can be described by the difference value between adjacent measurement values within its local range for each measurement value. The difference value can be variance. By describing the consistency of each change with variance, the outlier property can be characterized, that is, the larger the variance of the first-order difference, the higher the outlier property. That is, the variance of the first-order difference data can be calculated as the difference value of the difference data set, and it can be usedD( ) indicates 。

[0054] In short, step S200 in the embodiments of the present application includes:

[0055] S201: Using the measured values in the first measured value dataset as key points, define a preset number N before and after the key points to serve as the preset range area 2N + 1 corresponding to each measured value.

[0056] S202: Calculate the difference between every two adjacent measured values within the preset range area (2N + 1) corresponding to each measured value , to obtain the differential dataset W of the preset range area corresponding to each measured value, thereby obtaining the differential datasets of the preset range areas corresponding to each measured value and the difference value of this differential dataset of D( ) 。

[0057] Step S300: Based on the interval lengths of each measurement node, calculate the original interference evaluation values of each measured value in the first measured value dataset; according to the difference values corresponding to each measured value, calculate the regularization terms of each measured value and then multiply them with the corresponding original interference evaluation values to obtain the interference influence values of each measured value; perform normalization processing on the interference influence values of each measured value to obtain the target interference evaluation values of each measured value in the first measured value dataset.

[0058] In the embodiments of the present application, the interference evaluation value can also be referred to as the noise degree index. In a theoretical environment, the measurement data collected at each position of the same section of the power transmission line at the same time point should be basically the same. However, in reality, affected by environmental changes and the problems of the device's own structure, there will be differences at different positions. If the difference between the measured value of a measurement node and the measured values of other measurement nodes at the same time point is greater, the possible noise interference of this measurement node is greater. Therefore, it is necessary to calculate the interference evaluation value for the differences between the measured value data of each measurement node at the same time point. The greater the difference between the measured value of the current measurement node and the measured values of other measurement nodes, the greater its interference evaluation value. For example, the measured values of two measurement nodes with closer intervals should be more similar. The measurement nodes in the embodiments of the present application consider the measured value data of the same section of the transmission line. Therefore, the interval length between two measurement nodes can be used as the weight of the difference, and then the interference evaluation value of the measured value is obtained by weighted average to improve the accuracy of the calculation of the interference evaluation value.

[0059] Specifically, use an industrial data monitoring module as the first measurement node, and the industrial data monitoring modules other than the first measurement node are the second measurement nodes. Step S300 includes:

[0060] Step S301: Obtain the measurement value of the first measurement node and the measurement value of the second measurement node at the same time point, and calculate the difference between the two measurement values. That is, according to the difference between the measurement values of the first measurement node and other measurement nodes, the consistency of the measurement values of the two measurement nodes is described. The greater the change amount, the lower the consistency, and the greater the interference evaluation value may be. In other words, it is necessary to calculate the difference between the measurement values of the m th industrial data monitoring module and the k th industrial data monitoring module at time point t .

[0061] Step S302: Calculate the sub-interval metric value between the first measurement node and the second measurement node and the total interval metric value between the first measurement node and the second measurement node.

[0062] The sub-interval metric value is obtained based on the interval length between the first measurement node and a single second measurement node. The interval length is used to characterize the busbar length or the number of busbars from one measurement node to another measurement node. For example, it can be represented by D ( m, k ) representing the interval length between the m th industrial data monitoring module and the k th industrial data monitoring module. The reciprocal of this interval length is used as the sub-interval metric value . The total interval metric value is obtained based on the sum of the reciprocals of the interval lengths between the first measurement node and all second measurement nodes. For example, calculate as the total interval metric value , l represents the l th industrial data monitoring module, M represents the number of industrial data monitoring modules, and the total interval metric value characterizes the importance degree of the consistency between the first measurement node and other measurement nodes.

[0063] Step S303: Calculate the ratio of the sub-interval metric value to the total interval metric value, and use the ratio as the interval metric evaluation value of the measurement value;

[0064] In the embodiments of the present application, the interval metric evaluation value can be used as a weight coefficient. According to the situation that the measurement values of two measurement nodes closer in interval are more similar, the interval metric evaluation value is used as a prediction index for the consistency of the measurement values. The closer the interval length is, the higher the consistency is, and the greater the weight of the corresponding measurement value is.

[0065] Step S304: Multiply and add the measurement value differences of each measurement value and the interval metric evaluation value to obtain the original interference evaluation value of each measurement value.

[0066] In the embodiment of the present application, the interval metric evaluation value obtained in step S303 is used as the interval weight coefficient, and the cumulative sum of the product of the measurement value difference and the weight coefficient is calculated, and the calculation result is used as the original interference evaluation value. Combining the above steps S301 - S304, in the embodiment of the present application, the formula of the original interference evaluation value can be expressed as:

[0067] ;

[0068] wherein, is the original interference evaluation value of the m th industrial data monitoring module at the time point, is the measurement value of the m th industrial data monitoring module at the time point, is the measurement value of the k th industrial data monitoring module at the time point, is the interval weight coefficient of the m th industrial data monitoring module, and M represents the number of industrial data monitoring modules in this section of the transmission line.

[0069] In the embodiment of the present application, the difference value corresponding to each measurement value obtained in step 200 ， is the first-order difference sequence within the local area of the measurement value of the th industrial data monitoring module at the time point, and the difference value ， is also the variance of the first-order difference sequence within the local area of this measurement value Calculate the regularization term of each measurement value : , where exp[] is the exponential algorithm, and the purpose is to make the result between 0 and 1. The regularization term is multiplied by the original interference evaluation value obtained in step 304 to obtain the interference influence value, and the interference influence value is subjected to extreme value normalization processing to remove the maximum value and the minimum value to obtain the target interference evaluation value. In the embodiment of the present application, it is combined with the characteristic that the change of the interference point in the time series data often shows an outlier and has no relation with the change of the surrounding data. The interference evaluation value is adjusted through the outlier property of the interference measurement value. In other words, the outlier property of the measurement data in the embodiment of the present application can be represented by the variance of adjacent measurement values within the local area of the measurement value. By representing the consistency of each change through the variance, the outlier property can be characterized, that is, the greater the variance of the first-order difference, the higher the outlier property, the higher the possible degree of interference, the higher the interference evaluation value, and then a smaller adjustment is made to the interference evaluation value, and the adjusted interference evaluation value is used as the final target interference evaluation value. The target interference evaluation value The formula can be expressed as:

[0070] ;

[0071] wherein, represents the interference evaluation value of the measurement value of the m-th industrial data monitoring module at time point, represents the original interference evaluation value.

[0072] Step 400: Correct the first measurement value data set by using the target interference evaluation value of each measurement value to obtain a second measurement value data set; perform dynamic segmentation on the second measurement value data set to obtain second measurement value data sub-blocks of each measurement node.

[0073] In the embodiments of the present application, by correcting the first measurement value data set by using the target interference evaluation value of each measurement value, the error introduced by the interference value can be reduced, the smoothness and availability of the data can be enhanced, and the correction process of the first measurement value data set is a well-known technology and will not be elaborated here. In step 500, performing dynamic segmentation on the second measurement value data set to obtain second measurement value data sub-blocks of each measurement node may include the following steps 501-step 504:

[0074] Step 401: Perform pre-segmentation processing on the second measurement value data set to obtain original second measurement value data set sub-blocks. The pre-segmentation processing is based on a preset segmentation number, and the preset segmentation number is a positive integer; for example, if the preset segmentation number is set to [4, 5, 6, 7, 8], the second measurement value data set is respectively segmented into 4, 5, 6, 7, and 8 data sub-blocks.

[0075] Step 402: Calculate the error pre-estimation value of the original second measurement value data set sub-blocks, and add up the error pre-estimation values of all the original second measurement value data set sub-blocks to obtain the error result of the second measurement value data set under this preset segmentation number.

[0076] The error pre-estimation value can be calculated by using the sum of squared residuals function, the mean absolute error value function, or the root mean square error value. Preferably, in the embodiments of the present application, it is calculated by using the sum of squared residuals function, specifically: calculate the mean value of the measurement values in the original second measurement value data set sub-block to obtain the predicted value of the original second measurement value data set sub-block; calculate the difference between the measurement value and the predicted value of the original second measurement value data set sub-block and perform a sum of squares to obtain the sum of squared residuals of the original second measurement value data set sub-block, and then add up the sum of squared residuals of the original second measurement value data set sub-blocks under this preset segmentation number to obtain the error result. The formula for the error result S can be:

[0077] ,

[0078] Where Z is the preset number of segments, and Y is the total number of measurements of the sub-blocks of the original second measurement value dataset, is the observed value of the sub-block of the original second measurement value dataset, is the predicted value of the sub-block of the original second measurement value dataset, which can be the mean of all measurements of the sub-block of the original second measurement value dataset, is called the residual. For example, when the preset number of segments is [4, 5, 6, 7, 8], through the above formula calculation, the error results for the preset number of segments [4, 5, 6, 7, 8] can be obtained respectively .

[0079] Step 403: Use the preset number of segments as the horizontal axis and the error results as the vertical axis to create an error change graph for different preset numbers of segments. Determine the position of the inflection point according to the shape of the error change graph, and determine the target sub-block length according to the position of the inflection point; where the target sub-block length is a positive integer.

[0080] By analyzing the curve shape, determine the inflection point, which can also be called the elbow point, and determine the best target sub-block length by the preset number of segments value corresponding to the inflection point position. For example, among the preset number of segments [4, 5, 6, 7, 8], when the inflection point appears at the preset number of segments value of 5, then 5 is used as the target sub-block length. This method can avoid overfitting and quickly determine the target sub-block length. It should be emphasized that in the embodiments of the present application, through the combination of the optimal solution method and the elbow method, by continuously iterating the data segmentation number, the loss function is minimized, so that the measurement values in each second measurement value data sub-block are close in size, and combined with the above steps S401 - S403, the interference degree of the measurement values of the finally segmented data sub-blocks is stabilized at a level, that is, the noise degree is stable, and it is more adaptable to the complex changes in the measurement values in different time periods of the measurement data being disturbed differently.

[0081] Step 404: Perform secondary segmentation on the second measurement value dataset based on the target sub-block length to obtain the final second measurement value dataset sub-blocks, and use the final second measurement value dataset sub-blocks as the second measurement value data sub-blocks corresponding to the measurement nodes, so as to obtain the second measurement value data sub-blocks of each measurement node.

[0082] When using the second measurement value dataset for data segmentation, since the second measurement value dataset is corrected by the target interference evaluation value, that is, the data segmentation is based on the target interference evaluation value, the interference intensity of each second measurement value dataset sub-block is the same, and different interference parameters are applied to each data sub-block, thereby reducing the situation of oversmoothing and facilitating the accuracy of data monitoring.

[0083] Step 500: Determine the length of the sliding region according to the corresponding target interference evaluation value, perform a cumulative average of the measured values on the second measured value data sub-block, and perform a secondary cumulative average of the measured values on the second measured value data sub-block within the sliding region to obtain a third measured value data set; wherein, the length of the sliding region is a positive integer.

[0084] The determination of the length of the sliding region according to the corresponding target interference evaluation value includes the following steps 601-603:

[0085] Step 501: Based on the target interference evaluation value of the second measured value data set sub-block, calculate the second interference evaluation mean value of the second measured value data set sub-block;

[0086] Step 502: After calculating the product of the initial sliding region length and the second interference evaluation mean value, add it to the initial sliding region length to obtain the expected sliding region length;

[0087] Step 503: Take the largest positive integer value of the expected sliding region length to obtain the target sliding region length, and use this target sliding region length as the final sliding region length.

[0088] For the above steps 501-step 503, the formula for the final sliding region length can be expressed as:

[0089] ;

[0090] Wherein, is the preset initial sliding region length, is the mean value of the second interference evaluation values in the second measured value data set sub-block.

[0091] In the embodiments of the present application, Gaussian filtering can be used for processing, and the filtering effect is determined by the window size. For the data sub-block, the greater the interference evaluation value, the longer the required sliding region length. Through the expectation of the interference evaluation value in the data sub-block, the size of the window is dynamically adjusted, and the second measured value data set is filtered based on the adjusted target sliding window (i.e., the length of the sliding region), completing the data preprocessing of the first measured value data set of the industrial data monitoring module of each measurement node.

[0092] Step 600: Use the third measured value data set as the final measured value data set, and upload the final measured value data set to the server for the server to monitor the operating state of the busbar trunking for faults.

[0093] The server can apply statistical analysis and machine learning methods to analyze the abnormal status of the measured values, and use historical data to build a detection model for abnormal operating conditions. When the measured values ​​are detected to deviate from the detection model, potential faults can be identified. By real-time monitoring of the voltage and current parameters of each measurement node of the bus duct, combined with the anomaly detection algorithm, abnormal fluctuations in the operation of the bus duct can be effectively captured, thereby achieving early warning and diagnosis of faults.

[0094] like Figure 1 As shown, the embodiment of the present application also provides a bus duct fault monitoring system, including a server, an industrial data gateway and an industrial data monitoring module, the industrial data monitoring module is used to send the measurement data (that is, the first measurement data set) to the industrial data gateway, the industrial data gateway executes the bus duct fault monitoring method as described above, and the server is used to perform fault monitoring on the bus duct operation status according to the final measurement value data set (that is, the third measurement data set). The method executed by the industrial data gateway can refer to the content of the above application embodiment, which will not be repeated here.

[0095] Similarly, the bus duct fault monitoring system groups the industrial data monitoring modules at different measurement node positions, and takes the industrial data monitoring modules of the same transmission line as the same group, that is, considering the measurement nodes with similar measurement values, and utilizing the interval length between the measurement nodes to optimize the algorithm, thereby enhancing the accuracy and rationality of the interference assessment value estimation; in addition, the application designs a preset range area corresponding to each measurement value, and utilizes the difference value of the differential data set corresponding to each measurement value to dynamically adjust the interference assessment value, thereby enhancing the local adaptability of the optimization algorithm and monitoring slight changes in the measurement value; dynamic segmentation is performed based on the corrected measurement data, so that the measurement data with similar interference assessment values ​​are segmented together, which helps to improve the effect of measuring data interference removal and makes subsequent fault warnings more accurate.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

Claims

1. A bus duct fault monitoring method, characterized in that: Applied to an industrial data gateway, the method comprises: Obtain a first measurement value data set of the industrial data monitoring module of each measurement node of the same transmission line; the transmission lines include multiple transmission lines, the same transmission line refers to a section of the transmission line without a convergence point or a diversion point, each transmission line is composed of multiple bus ducts, and the industrial data monitoring modules of the same transmission line are defined as the same group; Calculate a differential data set of a preset range area corresponding to each measurement value and a difference value of the differential data set from the first measurement value data set; wherein the length of the preset range area is much smaller than the length of the first measurement value data set, and the length of the preset range area is a positive integer; Based on the interval length of each measurement node, the original interference evaluation value of each measurement value in the first measurement value data set is calculated; according to the difference value corresponding to each measurement value, the regularization term of each measurement value is calculated and then multiplied by the corresponding original interference evaluation value to obtain the interference impact value of each measurement value; the interference impact value of each measurement value is normalized to obtain the target interference evaluation value of each measurement value in the first measurement value data set; Correcting the first measurement data set using the target interference evaluation value of each measurement value to obtain a second measurement data set; dynamically segmenting the second measurement data set to obtain second measurement data sub-blocks of each measurement node; Determine the length of the sliding area according to the corresponding target interference assessment value, perform a measurement value cumulative average on the second measurement value data sub-block, perform a second measurement value cumulative average on the second measurement value data sub-block in the sliding area, and obtain a third measurement value data set; wherein the length of the sliding area is a positive integer; The third measurement value data set is used as the final measurement value data set, and the final measurement value data set is uploaded to the server, so that the server performs fault monitoring on the operation status of the bus duct.

2. The method according to claim 1, characterized in that Calculating a differential data set of a preset range area corresponding to each measurement value and a difference value of the differential data set from the first measurement value data set includes: Taking the measurement values ​​in the first measurement value data set as key points, defining a preset number before and after the key points as a preset range area corresponding to each measurement value; The difference between every two adjacent measurement values ​​in the preset range area corresponding to each measurement value is calculated to obtain a differential data set in the preset range area corresponding to each measurement value and a difference value of the differential data set, thereby obtaining a differential data set in the preset range area corresponding to each measurement value and a difference value of the differential data set; the differential data set is a first-order data sequence.

3. The method according to claim 1, characterized in that Based on the interval length of each measurement node, the original interference assessment value of each measurement value in the first measurement value data set is calculated, including: An industrial data monitoring module is used as a first measurement node, and industrial data monitoring modules other than the first measurement node are used as second measurement nodes, a measurement value of the first measurement node and a measurement value of the second measurement node at the same time point are obtained, and a difference between the two measurement values ​​is calculated; Calculating a sub-interval metric value between the first measurement node and the second measurement node and a total interval metric value between the first measurement node and the second measurement node; Calculate the ratio of the sub-interval metric value to the total interval metric value, and use the ratio as the interval metric evaluation value of the measurement value; The measured value difference of each measured value is multiplied and added with the interval metric evaluation value to obtain the original interference evaluation value of each measured value.

4. The method according to claim 1, characterized in that Dynamically segmenting the second measurement value data set to obtain second measurement value data sub-blocks of each measurement node includes: Pre-segmenting the second measurement data set to obtain original second measurement data set sub-blocks; the pre-segmenting is performed based on a preset segmentation number, which is a positive integer; Calculating an estimated error value of an original second measurement value data set sub-block, and adding the estimated error values ​​of all original second measurement value data set sub-blocks to obtain an error result of the second measurement value data set under the preset number of segmentations; Taking the preset number of divisions as the horizontal direction and the error result as the vertical direction, an error change graph under different preset numbers of divisions is produced, the position of the inflection point is determined according to the shape of the error change graph, and the target sub-block length is determined according to the position of the inflection point; wherein the target sub-block length is a positive integer; The second measurement data set is divided twice based on the target sub-block length to obtain a final second measurement data set sub-block, and the final second measurement data set sub-block is used as the second measurement data sub-block corresponding to the measurement node, thereby obtaining the second measurement data sub-block of each measurement node.

5. The method according to claim 4, characterized in that Calculating an error estimate of a sub-block of the original second measurement value data set includes: Calculate the mean of the measured values ​​in the sub-block of the original second measured value data set to obtain the predicted value of the sub-block of the original second measured value data set; The difference between the measured value and the predicted value of the original second measurement value data set sub-block is calculated and squared to obtain the square residual sum of the original second measurement value data set sub-block.

6. The method according to claim 4, characterized in that Calculating an error estimate of a sub-block of the original second measurement value data set includes: The mean absolute error value of the original second measurement value data set sub-block is calculated and taken as the error estimate value of the original second measurement value data set sub-block.

7. The method according to claim 4, characterized in that Calculating an error estimate of a sub-block of the original second measurement value data set includes: A root mean square error value of the original second measurement value data set sub-block is calculated and taken as an error estimate value of the original second measurement value data set sub-block.

8. The method according to claim 1, characterized in that The sliding area length is determined according to the corresponding target interference evaluation value, including: Calculating a second interference assessment mean value of the second measurement value data set sub-block based on the target interference assessment value of the second measurement value data set sub-block; After calculating the product of the initial sliding area length and the second interference evaluation mean, the product is added to the initial sliding area length to obtain the expected sliding area length; The maximum positive integer value is taken for the expected sliding area length to obtain the target sliding area length, and the target sliding area length is used as the final sliding area length.

9. The bus duct fault monitoring system is characterized by: It includes a server, an industrial data gateway and an industrial data monitoring module. The industrial data monitoring module is used to send measurement data to the industrial data gateway. The industrial data gateway executes the bus duct fault monitoring method as described in any one of claims 1 to 8. The server is used to perform fault monitoring on the bus duct operation status according to the final measurement value data set.

Images