Intelligent fault detection method and device, equipment and medium

Abstract

The invention relates to the technical field of fault detection, and provides an intelligent fault detection method and device, equipment and a medium, and the method comprises the steps: respectively obtaining a positive data set, a negative data set and an observation data set of weak current equipment; aligning and arranging the data sets in sequence to obtain a data matrix; calculating a data difference value of each column in the data matrix, and taking the column with the maximum data difference value as a reference feature column; arranging other residual columns in a descending order according to correlation coefficients, and sequentially filling the rest columns to two sides of the reference feature column so as to construct a feature significance matrix; a first sliding window and a second sliding window are adopted to scan the data set of the feature saliency matrix, and a window difference value is calculated; and after the whole-line scanning is completed, accumulating a difference value according to the weight, and when the accumulated difference value exceeds a threshold value, triggering fault early warning. By adopting the scheme, for the weak current equipment sensitive to signal change, the fault type of the weak current equipment can be accurately detected in real time.

Description

Technical Field

[0001] The present disclosure relates to the technical field of fault detection, and in particular to an intelligent fault detection method, device, equipment and medium. Background Art

[0002] In the weak-current equipment of steel mills, the working voltage of the weak-current equipment is low and the signal amplitude is small, and it is difficult for traditional detection methods to effectively capture weak fault characteristic signals; secondly, the weak-current equipment is easily affected by factors such as electromagnetic interference and environmental noise, resulting in a low signal-to-noise ratio of fault signals. These characteristics make the fault detection of weak-current equipment more difficult.

[0003] The existing technology has a high false alarm rate in the fault detection of weak-current equipment, cannot effectively identify small signal changes, and is prone to missing or false alarms of faults; while simple statistical analysis methods are difficult to capture the characteristics of faults, resulting in insufficient real-time performance and accuracy of fault detection; in addition, traditional fault detection methods often lack the comprehensive analysis ability of multiple fault types, and it is difficult to achieve accurate classification and positioning of fault types.

[0004] Therefore, there is an urgent need to develop an intelligent fault detection method for weak-current equipment to solve the problems of low detection sensitivity, high false alarm rate, and fuzzy fault classification in the existing technology, and to achieve accurate capture of tiny fault characteristics and real-time identification of type faults. Summary of the Invention

[0005] The present disclosure provides an intelligent fault detection method, device, equipment and medium to achieve real-time and accurate detection of the fault types of weak-current equipment.

[0006] In a first aspect, the present disclosure provides an intelligent fault detection method, including: respectively obtaining N positive data sets and N negative data sets of the weak-current equipment; obtaining the latest 1 observation data set of the weak-current equipment; each data set includes M data, and M is an odd number; aligning and arranging the N positive data sets, 1 observation data set and N negative data sets in sequence to obtain a data matrix of (2N + 1) × M; calculating the data difference value of each column in the data matrix, taking the column with the largest data difference value as the reference feature column, and moving the reference feature column to the middle column position of the data matrix; calculating the correlation coefficient between the other remaining columns and the reference feature column, and arranging the other remaining columns in descending order of the correlation coefficient, and filling them on both sides of the reference feature column in turn, so as to construct a feature significance matrix; Use the first sliding window to scan the observation data set in the (N + 1)-th row of the feature significance matrix; use the second sliding window to synchronously scan the positive data set in the K-th row of the feature significance matrix, where the K-th row is the middle row of the positive data set; synchronously move the two sliding windows, calculate the difference between the two sliding windows in real time, and set the first weight according to the real-time distance between the center position of the sliding window and the reference feature column; after completing the scanning of the entire row, accumulate the differences according to the first weight, and when the accumulated difference exceeds the threshold, trigger a fault warning.

[0007] According to the fault intelligent detection method provided by the present disclosure, respectively obtaining N positive data sets and N negative data sets of the weak current device includes: The N positive data sets are the data when the weak current device is in a normal operating state; The N negative data sets are the data when the weak current device is in N typical fault critical states; the N negative data sets respectively represent N typical faults, and the N typical faults include at least one of short circuit, open circuit, poor contact, signal attenuation, electromagnetic interference, component aging, and environmental interference; The data set is row data, and one data set is a row of data including M data.

[0008] According to the fault intelligent detection method provided by the present disclosure, triggering a fault warning further includes: Use N third sliding windows to synchronously scan the negative data sets in the (N + 1)-th row to the (2N + 1)-th row of the feature significance matrix respectively, and each window is responsible for scanning the data of one typical fault mode; Synchronously move the first sliding window and the N third sliding windows, calculate the feature similarity between the first sliding window and the N third sliding windows in real time; and set the second weight according to the real-time distance between the center position of the sliding window and the reference feature column; After completing the scanning of the entire row, accumulate the similarities of each row according to the second weight, select the row with the largest accumulated similarity, and determine the fault type according to the negative data set corresponding to the row with the largest similarity.

[0009] According to the fault intelligent detection method provided by the present disclosure, calculating the data difference value of each column in the data matrix and taking the column with the largest data difference value as the reference feature column includes: Perform wavelet transform on each column of the data matrix, and extract the wavelet coefficients of each column of data; Calculate the energy distribution of the wavelet coefficients of each column of data; Calculate the entropy value of the energy distribution of each column of data; Determine the column with the largest entropy value as the reference feature column.

[0010] According to the fault intelligent detection method provided by the present disclosure, calculating the correlation coefficient between other remaining columns and the reference feature column includes: Obtaining the time-domain correlation coefficient X between other remaining columns and the reference feature column; Obtaining the frequency-domain correlation coefficient Y between other remaining columns and the reference feature column; Obtaining the waveform correlation coefficient Z between other remaining columns and the reference feature column; Performing weighted calculation on X, Y, and Z to obtain the correlation coefficient.

[0011] According to the fault intelligent detection method provided by the present disclosure, arranging other remaining columns in descending order of the correlation coefficient and filling them on both sides of the reference feature column in sequence includes: Sorting other remaining columns in descending order of the correlation coefficient with the reference feature column; Moving the reference feature column to the middle column position of the data matrix; Adopting the left-right alternating filling rule, filling the first remaining column after sorting to the position immediately adjacent to the left of the reference feature column, filling the second remaining column to the position immediately adjacent to the right of the reference feature column, and filling the subsequent remaining columns to the left empty position according to the odd or even number of the sequence number, and filling the even-numbered columns to the right empty position, so as to construct a feature significance matrix with the reference feature column as the center and the column correlation gradually weakening from both sides in sequence.

[0012] According to the fault intelligent detection method provided by the present disclosure, synchronously moving two sliding windows and calculating the difference between the two sliding windows in real time includes: Synchronously moving two sliding windows to keep the starting positions of the two windows synchronously aligned on their respective rows; the size of the sliding window is optimized according to the signal characteristics of the weak-current device and the fault detection requirements.

[0013] In a second aspect, the present disclosure further provides a fault intelligent detection device, including: A data acquisition module that respectively acquires N positive data sets and N negative data sets of the weak-current device; acquires the latest 1 observation data set of the weak-current device; each data set includes M data, and M is an odd number; A data matrix module that aligns and arranges the N positive data sets, 1 observation data set, and N negative data sets in sequence to obtain a (2N + 1) × M data matrix; A feature significance matrix that calculates the data difference value of each column in the data matrix, takes the column with the largest data difference value as the reference feature column, moves the reference feature column to the middle column position of the data matrix; calculates the correlation coefficient between other remaining columns and the reference feature column, arranges other remaining columns in descending order of the correlation coefficient, and fills them on both sides of the reference feature column in sequence, so as to construct a feature significance matrix; The early warning module scans the observation data set in the (N + 1)-th row of the feature significance matrix using the first sliding window; synchronously scans the positive data set in the K-th row of the feature significance matrix using the second sliding window, where the K-th row is the middle row of the positive data set; synchronously moves the two sliding windows, calculates the difference between the two sliding windows in real time, and sets the first weight according to the real-time distance between the center position of the sliding window and the reference feature column; after completing the scan of the entire row, accumulates the differences according to the first weight, and when the accumulated difference exceeds the threshold, triggers a fault warning.

[0014] Compared with the prior art, the present disclosure relates to a fault intelligent detection method, which includes: respectively obtaining the positive data set, negative data set and observation data set of the weak current device; aligning and arranging the data sets in sequence to obtain a data matrix; calculating the data difference value of each column in the data matrix, and taking the column with the largest data difference value as the reference feature column; arranging the other remaining columns in descending order of the correlation coefficient and filling them on both sides of the reference feature column in sequence, so as to construct a feature significance matrix; using the first sliding window and the second sliding window to scan and calculate the difference in the data set of the feature significance matrix; after completing the scan of the entire row, accumulating the differences according to the weight, and when the accumulated difference exceeds the threshold, triggering a fault warning. By adopting the above scheme, for weak current devices sensitive to signal changes, the sensitivity and accuracy of fault detection of weak current devices can be significantly improved, weak fault signals can be effectively captured, the false alarm rate can be reduced, early warning and precise positioning of faults can be realized, and effective guarantee for the stable operation of weak current devices can be provided. Brief Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions in the present disclosure, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are some embodiments of the present disclosure. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0016] Figure 1 Flowchart of a fault intelligent detection method provided by the present disclosure; Figure 2 Feature significance matrix construction diagram provided by the present disclosure; Figure 3 Schematic diagram of a fault intelligent detection device provided by the present disclosure; Figure 4 Frame structure diagram of an electronic device provided by the present disclosure. Detailed Embodiments

[0017] To make the objectives, technical solutions, and advantages of the present disclosure clearer, the technical solutions in the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the present disclosure. Apparently, the described embodiments are some, but not all, of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative efforts shall fall within the protection scope of the present disclosure.

[0018] In modern steel mills, weak-current devices are key components in the production process, including but not limited to automation control systems, sensor networks, communication devices, monitoring systems, and various intelligent meters, etc. These devices usually operate in an environment of low voltage and small current and are vulnerable to interference from strong electromagnetic sources such as electric arc furnaces, frequency converters, and motor groups in the steel mill. Their normal operation is crucial for the production efficiency and safety of the steel mill.

[0019] Figure 1 is a flowchart of a fault intelligent detection method provided by the present disclosure. As Figure 1 shown, the method includes: Step 1: Obtain N positive data sets and N negative data sets of the weak-current device respectively; obtain the latest 1 observation data set of the weak-current device; each data set includes M data, and M is an odd number. Further, the N positive data sets D i,j + include N rows of data, where i is the row number, representing an integer from 1 to N, and j is the column number, representing an integer from 1 to M; Specifically, the first row of the positive data set includes M data, marked as D 1,M + ={D 1,1 + , D 1,2 + , …, D 1,M +}, which are historical data of the weak-current device in a normal operation state, and each row of data corresponds to the data of a normal operation cycle. For example, M parameters (ripple coefficient, voltage fluctuation, temperature, etc.) of the power supply module of a PLC controller in a certain steel mill during normal operation.

[0020] Similarly, the N negative data sets D i,j - include N rows of data, where i is the row number, representing an integer from 1 to N, and j is the column number, representing an integer from 1 to M; Specifically, the first row of the negative data set includes M data, marked as D 1,M - ={D 1,1 - , D 1,2 - , …, D1,M -} is the data of the weak current equipment in the typical fault critical state, reflecting the typical characteristics in the typical fault.

[0021] Furthermore, the data format of the latest 1 observation data set of the weak current equipment is D j ={D1, D2, …, D M}, which is the real-time data of the current operation cycle of the weak current equipment, has the same number of columns M as the positive and negative data sets, and is obtained in real time through a high-speed acquisition module (sampling rate ≥ 10 kHz) to ensure the timeliness of the data.

[0022] Step 2: Align and arrange the N positive data sets, 1 observation data set and N negative data sets in sequence to obtain a (2N + 1)×M data matrix; Furthermore, merge the arranged positive data sets, observation data set and negative data sets in sequence to form a data matrix with a scale of (2N + 1)×M; The first N rows correspond to the positive data set {D i,j +}, representing the normal operation state of the weak current equipment; the (N + 1)-th row corresponds to the observation data set D j , which is the current real-time operation data of the equipment; the last N rows correspond to the negative data set {D i,j −}, recording the data of the typical fault critical state; The form of the (2N + 1)×M data matrix is:

[0023] Step 3: Calculate the data difference value of each column in the data matrix, take the column with the largest data difference value as the reference feature column, and move the reference feature column to the middle column position of the data matrix; calculate the correlation coefficient between the other remaining columns and the reference feature column, and arrange the other remaining columns in descending order of the correlation coefficient and fill them on both sides of the reference feature column in turn, so as to construct a feature significance matrix; Furthermore, for each column in the data matrix, calculate the data difference value of this column, and the data difference value reflects the change degree of the data in this column between different rows; After calculating the data difference value of each column, select the column with the largest difference value as the reference feature column. The reference feature column reflects the most significant feature in the data matrix and is most relevant to the change of the operation state of the weak current equipment.

[0024] Move the reference feature column to the middle column position of the data matrix, that is, the (M + 1) / 2 column. This process ensures the core position of the reference feature column in the subsequent analysis and is convenient for subsequent feature analysis and fault detection.

[0025] Step 4: Use the first sliding window to scan the observation data set in the (N + 1)-th row of the feature significance matrix; use the second sliding window to synchronously scan the positive data set in the K-th row of the feature significance matrix, where the K-th row is the middle row of the positive data set; synchronously move the two sliding windows, calculate the difference between the two sliding windows in real time, and set the first weight according to the real-time distance between the center position of the sliding window and the reference feature column; after completing the scanning of the entire row, accumulate the differences according to the first weight, and when the accumulated difference exceeds the threshold, trigger a fault warning.

[0026] Further, the positive data set contains N rows of data, representing various operating condition data of the weak current device in the normal operating state. The K-th row, as the middle row of the positive data set, is a typical representative and reference of the normal data features. Selecting the middle row for scanning can obtain the most representative normal operating feature pattern on the basis of comprehensively considering various normal state data in the positive data set. Among them, K = [(N + 1) / 2], and [(N + 1) / 2] is the integer function.

[0027] Further, the first weight is optimized and set according to the position of the sliding window in the feature significance matrix. The sliding window farther away from the reference feature column is given a lower weight to reflect the characteristic that the importance of the data weakens with the increase of the distance, improving the reliability of fault detection.

[0028] Compared with the prior art, the intelligent fault detection method of the present disclosure respectively obtains the positive data set, negative data set and observation data set of the weak current device; aligns and arranges the data sets in sequence to obtain a data matrix; calculates the data difference value of each column in the data matrix, and takes the column with the largest data difference value as the reference feature column; arranges the other remaining columns in descending order of the correlation coefficient and fills them on both sides of the reference feature column in turn, thereby constructing a feature significance matrix; uses the first sliding window and the second sliding window to scan and calculate the difference in the data set of the feature significance matrix; after completing the scanning of the entire row, accumulates the differences according to the weight, and when the accumulated difference exceeds the threshold, triggers a fault warning. By adopting the above scheme, for the weak current device sensitive to signal changes, the sensitivity and accuracy of the weak current device fault detection can be significantly improved, weak fault signals can be effectively captured, the false alarm rate can be reduced, early warning and accurate positioning of faults can be realized, and effective guarantee for the stable operation of the weak current device can be provided.

[0029] In one implementation, the respectively obtaining N positive data sets and N negative data sets of the weak current device in step S1 includes: The N positive data sets are the data of the weak current device in the normal operating state; The N negative data sets are data of the weak-current equipment in N typical fault critical states, reflecting the operation characteristics of the equipment under different fault conditions, and are used for fault detection and classification; the N negative data sets respectively represent N typical faults, and the N typical faults include but are not limited to short circuit, open circuit, poor contact, signal attenuation, electromagnetic interference, component aging, and environmental interference; The data set is row data, and one data set is a row of data including M data.

[0030] Through the above steps, N positive data sets and N negative data sets of the weak-current equipment are successfully obtained, and these data sets provide high-quality basic data for subsequent fault detection and classification.

[0031] In one implementation, after triggering the fault warning in step 4, it further includes: Using N third sliding windows to perform synchronous scanning on the negative data sets in the (N + 1)-th row to the (2N + 1)-th row of the feature significance matrix respectively, and each window is responsible for scanning the data of one typical fault mode; Synchronously move the first sliding window and the N third sliding windows, and calculate the feature similarity between the first sliding window and the N third sliding windows in real time; and set the second weight according to the real-time distance between the center position of the sliding window and the reference feature column; After completing the full-row scanning, accumulate the similarities of each row according to the second weight, select the row with the largest accumulated similarity, and the negative data set corresponding to the row with the largest similarity is the current fault type, and determine the fault type according to the negative data set corresponding to the row with the largest similarity. For example, if the accumulated similarity corresponding to the 3rd third sliding window is the largest, and the negative data set scanned by this window represents the "poor contact" fault, it is determined that the current fault type is poor contact.

[0032] In one implementation, calculating the data difference value of each column in the data matrix and taking the column with the largest data difference value as the reference feature column includes: Perform wavelet transform on each column of the data matrix, and extract the wavelet coefficients of each column of data; Calculate the energy distribution of the wavelet coefficients of each column of data; Calculate the entropy value of the energy distribution of each column of data; Determine the column with the largest entropy value as the reference feature column.

[0033] In one implementation, calculating the correlation coefficient between other remaining columns and the reference feature column includes: Obtain the time-domain correlation coefficient X between other remaining columns and the reference feature column; Obtain the frequency-domain correlation coefficient Y between other remaining columns and the reference feature column; Obtain the waveform correlation coefficient Z between other remaining columns and the reference feature column; Perform weighted calculation on X, Y, and Z to obtain the correlation coefficient.

[0034] In one implementation, arranging the other remaining columns in descending order of the correlation coefficient and filling them on both sides of the reference feature column in sequence includes: Sort the other remaining columns in descending order of the correlation coefficient with the reference feature column; Move the reference feature column to the middle column position of the data matrix; Adopt the left - right alternating filling rule. Fill the first remaining column after sorting to the position immediately adjacent to the left of the reference feature column, the second remaining column to the position immediately adjacent to the right of the reference feature column, and the subsequent remaining columns are filled into the left empty position according to the odd - numbered sequence and into the right empty position according to the even - numbered sequence, so as to construct a feature significance matrix centered on the reference feature column, with the column correlation gradually weakening from the center to both sides.

[0035] Specifically, as Figure 2 shown, assume M = 5, N = 3, and the data matrix is as Figure 2 shown on the left; Calculate the data difference value of each column. Assume the difference values are: column 1 = 0.5, column 2 = 0.6, column 3 = 0.7, column 4 = 0.8, column 5 = 0.9; Select column 5 with the largest difference value as the reference feature column; move column 5 to the middle position, i.e., the 3rd column; Calculate the correlation coefficient between other columns and the reference feature column (column 5). Assume the correlation coefficients are: column 1 = 0.8, column 2 = 0.9, column 3 = 0.7, column 4 = 0.6; sort them in descending order of the correlation coefficient: column 2, column 1, column 3, column 4; Construct the feature significance matrix according to the left - right alternating filling rule: Column 2 (with the highest correlation) is filled to the position immediately adjacent to the left of the reference feature column; Column 1 (with the second - highest correlation) is filled to the position immediately adjacent to the right of the reference feature column; Column 3 (with the third - highest correlation) is filled to the second position to the left of the reference feature column; Column 4 (with the fourth - highest correlation) is filled to the second position to the right of the reference feature column; The final feature significance matrix is as Figure 2 shown on the right.

[0036] In one implementation, synchronously move two sliding windows and calculate the difference between the two sliding windows in real - time, including: Synchronously move two sliding windows, keeping the starting positions of the two windows synchronized and aligned on their respective rows; the size of the sliding window is optimized according to the signal characteristics of the low-voltage equipment and the requirements of fault detection.

[0037] Specifically, the first sliding window starts from column 1, and the second sliding window also starts from column 1. In each time step, the two windows move one data point to the right simultaneously until all columns are covered.

[0038] During the actual operation, the size of the sliding window is dynamically adjusted according to the characteristics of the real-time data. For example, when it is detected that the frequency components of the signal change, the size of the sliding window is automatically adjusted to adapt to the new signal characteristics.

[0039] Next, a fault intelligent detection device provided by the present disclosure will be described. The detection system described below can be mutually corresponded and referred to the detection method described above.

[0040] As Figure 3 shown, a fault intelligent detection device includes: A data acquisition module, which respectively acquires N positive data sets and N negative data sets of the low-voltage equipment; acquires the latest 1 observation data set of the low-voltage equipment; each data set includes M data, and M is an odd number; A data matrix module, which aligns and arranges the N positive data sets, 1 observation data set and N negative data sets in sequence to obtain a (2N + 1) × M data matrix; A feature significance matrix, which calculates the data difference value of each column in the data matrix, takes the column with the largest data difference value as the reference feature column, and moves the reference feature column to the middle column position of the data matrix; calculates the correlation coefficient between the other remaining columns and the reference feature column, and arranges the other remaining columns in descending order of the correlation coefficient, and fills them on both sides of the reference feature column in turn, so as to construct a feature significance matrix; An early warning module, which uses the first sliding window to scan the observation data set in the (N + 1)-th row of the feature significance matrix; uses the second sliding window to synchronously scan the positive data set in the K-th row of the feature significance matrix, and the K-th row is the middle row of the positive data set; synchronously moves the two sliding windows, calculates the difference between the two sliding windows in real time, and sets the first weight according to the real-time distance between the center position of the sliding window and the reference feature column; after completing the whole row scan, accumulates the difference according to the first weight, and when the accumulated difference exceeds the threshold, triggers a fault warning.

[0041] An electronic device provided by an embodiment of the present application, as Figure 4As shown, the electronic device includes a processor 401 and a memory 403. A computer program that can run on the processor is stored in the memory. When the processor executes the computer program, the steps of the method provided in the above embodiments are implemented.

[0042] See Figure 4 , the electronic device further includes: a bus 404 and a communication interface 402. The processor 401, the communication interface 402, and the memory 403 are connected through the bus 404; the processor 401 is used to execute an executable module stored in the memory 403, such as a computer program.

[0043] Among them, the memory 403 may include a high-speed random access memory (Random Access Memory, abbreviated as RAM), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 402 (which can be wired or wireless), a communication connection is realized between this system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0044] The bus 404 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 4 only a bidirectional arrow is used in

[0045] to represent, but it does not mean that there is only one bus or one type of bus.

[0046] The processor 401 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 401 or instructions in the form of software. The above-mentioned processor 401 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 403, and the processor 401 reads the information in the memory 403 and combines its hardware to complete the steps of the above method.

[0047] Corresponding to the above-mentioned fault intelligent detection method, an embodiment of the present application also provides a non-transitory computer-readable storage medium storing computer instructions. The computer-readable storage medium stores computer-executable instructions or a computer program. When the computer-executable instructions or the computer program are called and run by a processor, the computer-executable instructions cause the processor to run the steps of the above-mentioned fault intelligent detection method.

[0048] The fault intelligent detection device provided by the embodiments of the present application may be specific hardware on the device or software or firmware installed on the device, etc. For the device provided by the embodiments of the present application, its implementation principle and the technical effects produced are the same as those of the foregoing method embodiments. For the sake of brief description, for the parts not mentioned in the device embodiments, reference may be made to the corresponding content in the foregoing method embodiments. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can all refer to the corresponding processes in the above method embodiments, and will not be repeated here.

[0049] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some communication interfaces. The indirect coupling or communication connection of the devices or units can be in electrical, mechanical or other forms.

[0050] For another example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0051] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0052] In addition, the functional units in the embodiments provided in the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0053] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the fault intelligent detection method described in various embodiments of this application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM for short), random access memories (RAM for short), magnetic disks, or optical discs that can store program codes.

[0054] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. In addition, the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0055] Finally, it should be noted that: the above-mentioned embodiments are only specific implementation manners of this application, used to illustrate the technical solution of this application, rather than limiting it. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the art within the technical scope disclosed by this application can still modify the technical solutions described in the foregoing embodiments, or can easily think of changes, or make equivalent replacements for some of the technical features; and these modifications, changes, or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. An intelligent fault detection method, characterized in that, It includes the following steps: Obtain N positive datasets and N negative datasets of the low-voltage equipment respectively; obtain the latest 1 observation dataset of the low-voltage equipment; each dataset includes M data, and M is an odd number. Align and arrange the N positive datasets, 1 observation dataset and N negative datasets in sequence to obtain a (2N + 1)×M data matrix. Calculate the data difference value of each column in the data matrix, take the column with the largest data difference value as the reference feature column, and move the reference feature column to the middle column position of the data matrix; calculate the correlation coefficient between the other remaining columns and the reference feature column, and arrange the other remaining columns in descending order of the correlation coefficient, and fill them on both sides of the reference feature column in turn, so as to construct a feature significance matrix. Use the first sliding window to scan the observation dataset in the (N + 1)-th row of the feature significance matrix; use the second sliding window to synchronously scan the positive dataset in the K-th row of the feature significance matrix, and the K-th row is the middle row of the positive dataset; synchronously move the two sliding windows, calculate the difference between the two sliding windows in real time, and set the first weight according to the real-time distance between the center position of the sliding window and the reference feature column. After completing the full-row scan, accumulate the differences according to the first weight. When the accumulated difference exceeds the threshold, trigger a fault warning.

2. The fault intelligent detection method according to claim 1, wherein The step of respectively obtaining N positive datasets and N negative datasets of the low-voltage equipment includes: The N positive datasets are the data when the low-voltage equipment is in normal operation. The N negative datasets are the data when the low-voltage equipment is in N typical fault critical states; the N negative datasets respectively represent N types of typical faults, and the N typical faults include at least one of short circuit, open circuit, poor contact, signal attenuation, electromagnetic interference, component aging and environmental interference. The dataset is row data, and 1 dataset is a row of data including M data.

3. The fault intelligent detection method according to claim 2, characterized in that The triggering of the fault warning also includes: Use N third sliding windows to synchronously scan the negative datasets in the (N + 1)-th row to the (2N + 1)-th row of the feature significance matrix respectively, and each window is responsible for scanning the data of one typical fault mode. Synchronously move the first sliding window and the N third sliding windows, calculate the feature similarity between the first sliding window and the N third sliding windows in real time; and set the second weight according to the real-time distance between the center position of the sliding window and the reference feature column. After completing the full-row scan, accumulate the similarities of each row according to the second weight, select the row with the largest accumulated similarity, and determine the fault type according to the negative dataset corresponding to the row with the largest similarity.

4. The intelligent fault detection method according to claim 1, wherein, The calculation of the data difference value of each column in the data matrix and taking the column with the largest data difference value as the reference feature column includes: Perform wavelet transform on each column of the data matrix, and extract the wavelet coefficients of each column of data. Calculate the energy distribution of the wavelet coefficients of each column of data. Calculate the entropy value of the energy distribution of each column of data. Determine the column with the largest entropy value as the reference feature column.

5. The intelligent fault detection method according to claim 1, wherein The calculation of the correlation coefficient between the other remaining columns and the reference feature column includes: Obtain the time-domain correlation coefficient X between the other remaining columns and the reference feature column. Obtain the frequency domain correlation coefficient Y between the other remaining columns and the reference feature column; Obtain the waveform correlation coefficient Z between the other remaining columns and the reference feature column; Perform weighted calculation on X, Y, and Z to obtain the correlation coefficient.

6. The intelligent fault detection method according to claim 5, characterized in that, The arranging the other remaining columns in descending order of the correlation coefficient and filling them on both sides of the reference feature column in sequence includes: Sort the other remaining columns in descending order of the correlation coefficient with the reference feature column; Move the reference feature column to the middle column position of the data matrix; Adopt the left-right alternating filling rule, fill the first remaining column after sorting to the position immediately adjacent to the left of the reference feature column, fill the second remaining column to the position immediately adjacent to the right of the reference feature column, and fill the subsequent remaining columns to the left empty position according to the odd or even number of the serial number, and fill the even-numbered columns to the right empty position, so as to construct a feature significance matrix with the reference feature column as the center and the column correlation gradually weakening from both sides in sequence.

7. The intelligent fault detection method according to claim 1, characterized in that The synchronously moving two sliding windows and calculating the difference between the two sliding windows in real time includes: Synchronously move two sliding windows to keep the starting positions of the two windows synchronously aligned on their respective rows; the size of the sliding window is optimized according to the signal characteristics of the weak current equipment and the requirements of fault detection.

8. An intelligent fault detection device, characterized in that, Including: A data acquisition module that respectively acquires N positive data sets and N negative data sets of the weak current equipment; Obtain the latest 1 observed data set of the weak current equipment; each data set includes M data, and M is an odd number; A data matrix module that aligns and arranges the N positive data sets, 1 observed data set, and N negative data sets in sequence to obtain a (2N + 1)×M data matrix; A feature significance matrix that calculates the data difference value of each column in the data matrix, takes the column with the largest data difference value as the reference feature column, and moves the reference feature column to the middle column position of the data matrix; calculates the correlation coefficient between the other remaining columns and the reference feature column, arranges the other remaining columns in descending order of the correlation coefficient, and fills them on both sides of the reference feature column in sequence, so as to construct a feature significance matrix; An early warning module that scans the observed data set in the (N + 1)th row of the feature significance matrix using the first sliding window; synchronously scans the positive data set in the Kth row of the feature significance matrix using the second sliding window, and the Kth row is the middle row of the positive data set; synchronously move the two sliding windows, calculate the difference between the two sliding windows in real time, and set the first weight according to the real-time distance between the center position of the sliding window and the reference feature column; After completing the full-row scan, accumulate the differences according to the first weight, and trigger a fault warning when the accumulated difference exceeds the threshold.

9. An electronic device, comprising: A processor; A memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to execute the method according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer-readable storage medium stores instructions or a computer program that, when the instructions or the computer program runs on the device, causes the device to execute the method according to any one of claims 1-7.

Images