Industrial system fault detection method and electronic equipment
By calculating the local outlier factor and real-time data of the sensor, combined with the multi-layer classifier to identify the fault sensor, the problem of the inability to accurately locate the fault sensor in the existing technology is solved, and the accurate positioning and identification of fault modes is realized in the industrial system is improved, and the fault is improved. Check efficiency.
Patent Information
- Application Number
- CN202111434866.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-29
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-11-29
AI Technical Summary
The prior art is unable to accurately locate fault sensors from industrial systems with numerous sensors and to know the fault mode of industrial systems, resulting in difficulty in troubleshooting and clearing.
By obtaining real-time data of multiple sensors, calculating local outliers, determining fault sensors, and inputting their fault data into multi-layer classifiers to identify fault patterns in industrial systems, detection is performed using support vector machines or support vector data descriptions.
It realizes accurate positioning of fault sensors from industrial systems with numerous sensors and identifying fault patterns in industrial systems, improving the efficiency and accuracy of troubleshooting.
Smart Images

Figure CN114429171B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of electronic technology, and in particular to an industrial system fault detection method and electronic equipment. Background Art
[0002] In various industrial fields, as the scale of production becomes larger and larger, the equipment composition in industrial systems becomes increasingly complex. For example, the equipment in the deep-sea space station system includes a variety of different equipment such as machinery, hydraulics, pipelines, and electrical appliances, which have complex physical structures. In addition, there are many types of sensors in each device. However, there are many types of industrial system failures. Existing methods cannot accurately locate faulty sensors and obtain the failure mode of industrial systems from industrial systems with many sensors, which makes it difficult to troubleshoot and eliminate industrial system failures. Summary of the invention
[0003] The present invention provides an industrial system fault detection method and electronic equipment, which are used to solve the defects in the prior art that a faulty sensor cannot be accurately located in an industrial system having a large number of sensors and the fault mode of the industrial system cannot be known.
[0004] The present invention provides an industrial system fault detection method, wherein the industrial system includes a plurality of sensors, and the method includes:
[0005] Acquiring real-time data from the multiple sensors;
[0006] Determine the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested;
[0007] The fault data of the fault sensor is input into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, and each base classifier is used to detect a different fault mode of the industrial system.
[0008] According to an industrial system fault detection method provided by the present invention, the faulty sensor is determined based on the relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested, including:
[0009] respectively calculating a first local outlier factor of the normal state data of the sensor to be tested and a second local outlier factor of the real-time data of the sensor to be tested;
[0010] Obtaining a local outlier factor threshold range from the first local outlier factor based on a three sigma criterion;
[0011] A faulty sensor is determined based on a relationship between the second local outlier factor and the local outlier factor threshold range.
[0012] According to an industrial system fault detection method provided by the present invention, the determining of the fault sensor based on the relationship between the second local outlier factor and the local outlier factor threshold range includes:
[0013] When the second local outlier factor is within the local outlier factor threshold range, determining that the sensor to be tested is a normal sensor; and
[0014] When the second local outlier factor is outside the local outlier factor threshold range, it is determined that the sensor to be tested is a faulty sensor.
[0015] According to an industrial system fault detection method provided by the present invention, the first base classifier in the multi-layer classifier is used to receive fault data of the fault sensor;
[0016] In a case where the fault data of the fault sensor meets the classification condition of the first base classifier, determining that the fault mode of the industrial system is the fault mode corresponding to the first base classifier; or
[0017] In a case where the fault data of the faulty sensor does not satisfy the classification condition of the first base classifier, the fault data of the faulty sensor is input into a next base classifier.
[0018] According to an industrial system fault detection method provided by the present invention, after acquiring the real-time data of the multiple sensors, the method further includes:
[0019] Performing data dimension reduction processing on the real-time data of the multiple sensors.
[0020] According to an industrial system fault detection method provided by the present invention, after inputting the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system, the method further includes:
[0021] Perform linear fitting on the data of the faulty sensor within a preset time period to predict the parameter change trend of the faulty sensor.
[0022] According to an industrial system fault detection method provided by the present invention, the base classifier is described by a support vector machine or support vector data.
[0023] The present invention also provides an industrial system fault detection device, comprising:
[0024] A data acquisition module, used to acquire real-time data of the multiple sensors;
[0025] A faulty sensor determination module, configured to determine a faulty sensor based on a relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested;
[0026] The fault mode recognition module is used to input the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, and each base classifier is used to detect a different fault mode of the industrial system.
[0027] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above industrial system fault detection methods are implemented.
[0028] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above industrial system fault detection methods are implemented.
[0029] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of any one of the above industrial system fault detection methods are implemented.
[0030] The industrial system fault detection method and electronic device provided by the present invention obtain the real-time data of the multiple sensors; and determine the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested; thereby, the embodiment of the present invention accurately locates the faulty sensor from the industrial system with many sensors. The fault mode of the industrial system is obtained by inputting the fault data of the faulty sensor into a multi-layer classifier; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, each of which is used to detect different fault modes of the industrial system. The embodiment of the present invention detects the fault data of the faulty sensor one by one through a plurality of base classifiers, and then identifies the fault mode of the industrial system. The embodiment of the present invention realizes the accurate location of the faulty sensor and the acquisition of the fault mode of the industrial system from the industrial system with many sensors. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0032] Figure 1 It is one of the flow charts of the industrial system fault detection method provided by the present invention;
[0033] Figure 2 This is the second flow chart of the industrial system fault detection method provided by the present invention;
[0034] Figure 3 is a schematic diagram of a multi-layer classifier provided by the present invention;
[0035] Figure 4 This is the third flow chart of the industrial system fault detection method provided by the present invention;
[0036] Figure 5 This is the fourth flow chart of the industrial system fault detection method provided by the present invention;
[0037] Figure 6 It is a structural schematic diagram of the industrial system fault detection device provided by the present invention;
[0038] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0039] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0040] Combine the following Figure 1-Figure 5 The industrial system fault detection method of the present invention is described.
[0041] Wherein, the industrial system includes multiple sensors, please refer to Figure 1 , the industrial system fault detection method of the present invention comprises:
[0042] Step 100: acquiring real-time data of the plurality of sensors;
[0043] Real-time data of multiple sensors are obtained through electronic devices. Specifically, the industrial system can be an industrial system for various application scenarios. For example, it can be a deep-sea space station system. The deep-sea space station system has multiple sensors. For example, the deep-sea space station system may include more than a dozen or even dozens of sensors such as temperature sensors, hydraulic sensors, and water level sensors. By obtaining the real-time data of the multiple sensors to form a data set of multiple sensors, it is convenient to locate the early fault sensors of the industrial system, identify the fault mode of the industrial system, and predict the parameter change trend of the faulty sensor.
[0044] Step 300, determining a faulty sensor based on a relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested;
[0045] The electronic device determines the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested. The normal state data of the sensor to be tested refers to the operating data of the sensor to be tested in a normal working state. This part of the normal state data can be pre-collected by the staff of the industrial system for each sensor when the industrial system is operating normally.
[0046] The LOF (Local Outlier Factor, abbreviated as LOF, i.e., local outlier factor) method is a typical density-based high-precision outlier detection method. In the LOF method, an outlier factor LOF that depends on the neighborhood density is assigned to each data point to determine whether the data point is an outlier. This embodiment identifies faulty sensors among numerous sensors in an industrial system based on the local outlier factor.
[0047] Specifically, in some feasible embodiments, please refer to Figure 2 Step 300, determining the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested, comprises:
[0048] Step 310, respectively calculating a first local outlier factor of the normal state data of the sensor to be tested and a second local outlier factor of the real-time data of the sensor to be tested;
[0049] Specifically, the local outlier factor is calculated as follows: first calculate the k-distance of the sample, then calculate its k-distance neighbors and reachable distance, and then calculate the local reachable density to obtain the local outlier factor of the sample.
[0050] The following is an explanation using a formula. Assume that D is a dataset q, p, and o are objects in the dataset. The k-distance of object p is defined as the distance between p and o, where object o∈D and satisfies the following conditions:
[0051] (1) There are at least k points o' in the set that do not include p, satisfying
[0052] d(p, o')≤d(p, o); Formula (1)
[0053] (2) There are at most k-1 points o' in the set that do not include p, satisfying
[0054] d(p, o')<d(p, o); Formula (2)
[0055] In addition, the k-distance neighbors of object p, N k (P) is defined as:
[0056] N k (p) = {q∈D\{p}|d(p,q)≤k-distance(p)}; Formula (3)
[0057] The reachable distance is:
[0058] reach-dist(p,o)=max(k-distance(p),d(p,o));
[0059] Its local reachable density is:
[0060]
[0061] The local outlier factor of object p is:
[0062]
[0063] The d(p, o) here uses the Manhattan distance.
[0064] The first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested can be calculated by the above formula (1) to formula (5).
[0065] Step 320: Obtain a local outlier factor threshold range from the first local outlier factor based on the three sigma criterion;
[0066] When a first local outlier factor of the normal state data of the sensor to be tested is obtained, the electronic device obtains a local outlier factor threshold range from the first local outlier factor based on the three sigma criterion, determines a local outlier factor interval when a sensor is in a normal state from the first local outlier factor based on the three sigma criterion, and determines a faulty sensor based on the local outlier factor threshold range.
[0067] Step 330: Determine a faulty sensor based on the relationship between the second local outlier factor and the local outlier factor threshold range.
[0068] The electronic device determines the faulty sensor based on the relationship between the second local outlier factor and the local outlier factor threshold range. Specifically, in some feasible embodiments, step 330, determining the faulty sensor based on the relationship between the second local outlier factor and the local outlier factor threshold range, includes:
[0069] Step 331: When the second local outlier factor is within the local outlier factor threshold range, determine that the sensor to be tested is a normal sensor; and
[0070] When the second local outlier factor is within the local outlier factor threshold range, it indicates that the sensor to be tested is not an outlier, and the sensor to be tested is a normal sensor.
[0071] Step 332: When the second local outlier factor is outside the local outlier factor threshold range, determine that the sensor to be tested is a faulty sensor.
[0072] When the second local outlier factor is outside the local outlier factor threshold range, it indicates that the sensor to be tested is an outlier and the sensor to be tested is a faulty sensor. This embodiment determines the faulty sensor by the relationship between the first local outlier factor based on the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested; thus, the embodiment of the present invention accurately locates the faulty sensor in an industrial system with many sensors.
[0073] Step 400: input the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, each of the base classifiers is used to detect a different fault mode of the industrial system.
[0074] Since the multi-layer classifier includes a plurality of base classifiers connected in sequence, the plurality of base classifiers match the fault data of the faulty sensor step by step, thereby identifying different fault modes of the industrial system.
[0075] Time domain statistical analysis is used to diagnose faults with large short-term changes such as mutations based on the changes in time domain statistics. Changes in the operating state of the sensor will cause changes in the statistics of the signal's time domain description. The time domain statistics are the statistical values of each time domain statistical parameter. Whether the sensor is faulty can be directly determined by the change in the signal waveform or diagnosed based on the change in the signal's time domain statistics. It is often difficult to identify by judging the change in the waveform, but the characteristics of the signal can be quickly and effectively reflected based on the time domain statistics. Therefore, diagnosing faults based on changes in the signal's time domain statistics is a simple and effective method. Fault diagnosis using time domain statistical parameters not only takes into account the sensitivity of early faults, but also their stability.
[0076] Therefore, the fault data of the fault sensor can adopt the time domain statistics of the fault sensor. The time domain statistics can be divided into dimensioned eigenvalues and dimensionless eigenvalues according to whether they have dimensions or not. Dimensional eigenvalues include: maximum value, minimum value, peak-to-peak value, mean value, variance, mean square value, root mean square value, etc. Since the numerical value of the dimensional eigenvalue often changes due to changes in some external physical quantities, it brings certain difficulties to practical applications. The time domain statistics of the fault sensor in this embodiment can adopt a variety of dimensionless indicators, including peak factor, pulse factor, margin factor, kurtosis factor, waveform factor and skewness, etc.
[0077] Therefore, each base classifier in the multi-layer classifier in this embodiment can identify whether the above-mentioned time domain statistics such as peak factor, pulse factor, margin factor, kurtosis factor, waveform factor, etc. are normal. For example, the multi-layer classifier includes three base classifiers, and base classifier 1 of the three base classifiers is used to detect whether the peak factor of the faulty sensor is normal; base classifier 2 is used to detect whether the pulse factor of the faulty sensor is normal; base classifier 3 is used to detect whether the margin factor of the faulty sensor is normal. Therefore, when a base classifier detects an abnormality, it can be obtained that a time domain statistic of the faulty sensor is faulty, and the fault mode of the industrial system can be further determined by tracing the source.
[0078] It should be noted that the base classifier may be a support vector machine (SVM) or a support vector data description (SVDD).
[0079] The first base classifier in the multi-layer classifier is used to receive the fault data of the fault sensor; when the fault data of the fault sensor meets the classification condition of the first base classifier, the fault mode of the industrial system is determined to be the fault mode corresponding to the first base classifier; or when the fault data of the fault sensor does not meet the classification condition of the first base classifier, the fault data of the fault sensor is input into the next base classifier.
[0080] It should be noted that the first base classifier refers to any base classifier in the multi-layer classifier. The time domain statistics of the faulty sensor are matched step by step through multiple base classifiers, so as to identify different fault modes of the industrial system.
[0081] For example, see Figure 3 In a typical embodiment, the base classifier 11 of the three base classifiers of the multi-layer classifier 10 is used to detect whether the peak factor of the faulty sensor is normal; the base classifier 12 is used to detect whether the pulse factor of the faulty sensor is normal; and the base classifier 13 is used to detect whether the margin factor of the faulty sensor is normal. When the base classifiers 11 and 12 detect that the peak factor and pulse factor of the faulty sensor are abnormal, the fault mode of the industrial system is a peak factor fault and a pulse factor fault, and the fault of the industrial system is checked and cleared based on the above two faults.
[0082] For other aspects of this application, please refer to Figure 4 , after step 100, the real-time data of the plurality of sensors are obtained, the method further comprises:
[0083] Step 200: Perform data dimension reduction processing on the real-time data of the multiple sensors.
[0084] Since industrial systems are equipped with sensors of many types and numbers, the amount of real-time data from such a large number of sensors is also very huge, which will also bring difficulties to step 300 to calculate the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested.
[0085] Therefore, this embodiment performs data dimensionality reduction processing on the real-time data of the multiple sensors to reduce the amount of data, thereby reducing the amount of calculation in step 300. Specifically, the real-time data of the multiple sensors can be processed by a deep autoencoder. In other embodiments, the real-time data of the multiple sensors can also be processed by principal component analysis (Principal Component Analysis) and isometric mapping (Isometric Mapping) and other methods to achieve data dimensionality reduction processing.
[0086] For other aspects of this application, please refer to Figure 5 Step 400, after inputting the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system, further includes:
[0087] Step 500: Perform linear fitting on the data of the faulty sensor within a preset time period to predict the parameter change trend of the faulty sensor.
[0088] After obtaining the faulty sensor, the electronic device can also predict the parameter change trend of the faulty sensor to learn about and prevent possible faults in the industrial system in advance.
[0089] Specifically, the electronic device predicts the parameter change trend of the fault sensor by performing linear fitting on the data of the fault sensor within a preset time. The method used for linear fitting in this implementation is the least squares method. Assume that a total of n groups of data are collected within the preset time: (t1, y1), (t2, y2), ... (tn, yn), where tn represents the time of data collection, and yn represents the collected parameter value.
[0090] The sample regression model is constructed as follows:
[0091] Right now
[0092] Where e is the error of sample (ti, yi). The square loss function is:
[0093]
[0094] When Q is the smallest, the data fitting effect is the best. It can reflect the changing trend of the parameter. Consider formula (3) as Q about Function, find the value when Q is minimum. Let the partial derivative of formula (3-3) equal to 0:
[0095]
[0096] get
[0097] Thus, multiple data of the faulty sensor within a preset time period are substituted into formula (6) to formula (10). The value of can predict the parameter change trend of the faulty sensor. Specifically, when When it is greater than the upper trend threshold, it is considered that the parameter has an increasing trend: When it is less than the lower trend threshold, it is considered that the parameter has a decreasing trend; when it is between the upper trend threshold and the lower trend threshold, it is considered that the parameter is stable and the data is a normal fluctuation. By performing linear fitting on the data of the faulty sensor within a preset time, the operation is simple and the sensitivity is high, and possible faults in the industrial system can be known and prevented in advance.
[0098] The industrial system fault detection device provided by the present invention is described below. The industrial system fault detection device described below and the industrial system fault detection method described above can be referred to each other.
[0099] Please refer to Figure 6 The present invention also provides an industrial system fault detection device, comprising:
[0100] A data acquisition module 201 is used to acquire real-time data of the multiple sensors;
[0101] A faulty sensor determination module 202, configured to determine a faulty sensor based on a relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested;
[0102] The fault mode identification module 203 is used to input the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, each of which is used to detect a different fault mode of the industrial system.
[0103] The industrial system fault detection device of this embodiment obtains the real-time data of the multiple sensors; and determines the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested; thus, the embodiment of the present invention accurately locates the faulty sensor from the industrial system with many sensors. The fault mode of the industrial system is obtained by inputting the fault data of the faulty sensor into a multi-layer classifier; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, and each base classifier is used to detect different fault modes of the industrial system. The embodiment of the present invention detects the fault data of the faulty sensor one by one through a plurality of base classifiers, and then identifies the fault mode of the industrial system. The embodiment of the present invention realizes the accurate location of the faulty sensor and the acquisition of the fault mode of the industrial system from the industrial system with many sensors. On the basis of the above embodiments, as an optional embodiment, the base fault sensor determination module includes:
[0104] A first calculation module, used to respectively calculate a first local outlier factor of the normal state data of the sensor to be tested and a second local outlier factor of the real-time data of the sensor to be tested;
[0105] A threshold range module, configured to obtain a local outlier factor threshold range from the first local outlier factor based on a three sigma criterion;
[0106] The final determination module is configured to determine a faulty sensor based on a relationship between the second local outlier factor and the local outlier factor threshold range.
[0107] Based on the above embodiments, as an optional embodiment, the final determination module includes:
[0108] a first sub-determination module, configured to determine that the sensor to be tested is a normal sensor when the second local outlier factor is within the local outlier factor threshold range; and
[0109] The second sub-determination module is configured to determine that the sensor to be tested is a faulty sensor when the second local outlier factor is outside the local outlier factor threshold range.
[0110] Based on the above embodiments, as an optional embodiment, the first base classifier in the multi-layer classifier is used to receive the fault data of the fault sensor;
[0111] In a case where the fault data of the fault sensor meets the classification condition of the first base classifier, determining that the fault mode of the industrial system is the fault mode corresponding to the first base classifier; or
[0112] In a case where the fault data of the faulty sensor does not satisfy the classification condition of the first base classifier, the fault data of the faulty sensor is input into a next base classifier.
[0113] Based on the above embodiments, as an optional embodiment, the industrial system fault detection device further includes:
[0114] The dimension reduction module is used to perform data dimension reduction processing on the real-time data of the multiple sensors.
[0115] Based on the above embodiments, as an optional embodiment, the industrial system fault detection device further includes:
[0116] The prediction module is used to perform linear fitting on the data of the fault sensor within a preset time period to predict the parameter change trend of the fault sensor.
[0117] Based on the above embodiments, as an optional embodiment, the base classifier is described using a support vector machine or support vector data.
[0118] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the industrial system fault detection method, wherein the industrial system includes a plurality of sensors, and the method includes: obtaining real-time data of the plurality of sensors; determining the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested; inputting the fault data of the faulty sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, and each of the base classifiers is used to detect different fault modes of the industrial system.
[0119] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.
[0120] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the industrial system fault detection method provided by the above methods. The industrial system includes multiple sensors. The method includes: acquiring real-time data of the multiple sensors; determining a faulty sensor based on the relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested; inputting the fault data of the faulty sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes multiple base classifiers connected in sequence, and each base classifier is used to detect a different fault mode of the industrial system.
[0121] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the industrial system fault detection method provided by the above-mentioned methods, wherein the industrial system includes multiple sensors, and the method includes: acquiring real-time data of the multiple sensors; determining a faulty sensor based on a relationship between a first local outlier factor of normal state data of the sensor to be tested and a second local outlier factor of real-time data of the sensor to be tested; inputting the fault data of the faulty sensor into a multi-layer classifier to obtain a fault mode of the industrial system; wherein the multi-layer classifier includes multiple base classifiers connected in sequence, and each base classifier is used to detect a different fault mode of the industrial system.
[0122] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0123] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for detecting faults in an industrial system, characterized in that: The industrial system includes a plurality of sensors, and the method includes: Acquiring real-time data from the multiple sensors; Determine the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested; Inputting the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system; wherein the multi-layer classifier includes a plurality of base classifiers connected in sequence, each of which is used to detect a different fault mode of the industrial system; The method of determining the faulty sensor based on the relationship between the first local outlier factor of the normal state data of the sensor to be tested and the second local outlier factor of the real-time data of the sensor to be tested includes: respectively calculating a first local outlier factor of the normal state data of the sensor to be tested and a second local outlier factor of the real-time data of the sensor to be tested; Obtaining a local outlier factor threshold range from the first local outlier factor based on a three sigma criterion; determining a faulty sensor based on a relationship between the second local outlier factor and the local outlier factor threshold range; The determining of the faulty sensor based on the relationship between the second local outlier factor and the local outlier factor threshold range includes: When the second local outlier factor is within the local outlier factor threshold range, determining that the sensor to be tested is a normal sensor; and When the second local outlier factor is outside the local outlier factor threshold range, determining that the sensor to be tested is a faulty sensor; The first base classifier in the multi-layer classifier is used to receive the fault data of the fault sensor; In a case where the fault data of the fault sensor meets the classification condition of the first base classifier, determining that the fault mode of the industrial system is the fault mode corresponding to the first base classifier; or In a case where the fault data of the faulty sensor does not satisfy the classification condition of the first base classifier, the fault data of the faulty sensor is input into a next base classifier.
2. The industrial system fault detection method according to claim 1, characterized in that: After acquiring the real-time data of the multiple sensors, the method further includes: Performing data dimension reduction processing on the real-time data of the multiple sensors.
3. The industrial system fault detection method according to claim 1, characterized in that: After inputting the fault data of the fault sensor into a multi-layer classifier to obtain the fault mode of the industrial system, the method further includes: Perform linear fitting on the data of the faulty sensor within a preset time period to predict the parameter change trend of the faulty sensor.
4. The industrial system fault detection method according to any one of claims 1 to 3, characterized in that: The base classifier is described using a support vector machine or support vector data.
5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the industrial system fault detection method according to any one of claims 1 to 4 are implemented.
6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the industrial system fault detection method according to any one of claims 1 to 4 are implemented.
7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the industrial system fault detection method according to any one of claims 1 to 4 are implemented.
Citation Information
Cited By
Fault detection system of power electronic system and detection method thereof
CN120294476A
