A device abnormality diagnosis method, system and program product based on artificial intelligence

Abstract

The present invention belongs to the technical field of fault diagnosis, and specifically discloses an equipment abnormality diagnosis method, system and program product based on artificial intelligence, which obtains the inspection data group to be analyzed by merging the current inspection record data with the corresponding historical normal inspection data group according to the inspection data type number, and then extracts the array feature set of the inspection data group to be analyzed, and inputs the extracted array feature set into the KNN model for binary classification prediction to obtain the data abnormality classification result, and when there is an abnormal label in each classification result, the inspection data type number and the data abnormality classification result are used to form a feature matrix, and finally the feature matrix is ​​input into the fault classification diagnosis model for fault diagnosis to obtain the fault classification result of the target device. The present invention can realize intelligent equipment inspection data analysis, quickly and accurately diagnose the fault type of the equipment, effectively improve the abnormal diagnosis efficiency of the equipment, and ensure the reliability of the abnormal diagnosis of the equipment.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault diagnosis, and specifically relates to an equipment abnormality diagnosis method, system and program product based on artificial intelligence. Background Art

[0002] In modern enterprises and organizations, the normal operation of equipment is the key to the sustainable development of business, especially in the energy and chemical industry. As its development trend continues to be mechanized, automated, large-scale, high-speed and complex, the requirements for equipment inspection are getting higher and higher. Equipment inspection is to check and diagnose the equipment, grasp the changes in the equipment operation status and the surrounding environment, discover the hidden dangers and defects of the equipment as early as possible, reduce the chance of sudden failure, and keep the equipment in good operating condition. The purpose of equipment inspection is to ensure the normal operation of the equipment, prevent potential failures, improve equipment availability and extend equipment life.

[0003] At present, in some traditional enterprises and organizations, the inspection data recorded after equipment inspections are mostly analyzed and processed manually, that is, professional operation and maintenance personnel determine whether the equipment has abnormalities based on the recorded inspection data, and when the equipment inspection data is judged to be abnormal, the type of fault that caused the equipment abnormality is manually determined based on the abnormal data. This manual equipment abnormality diagnosis method is highly dependent on the professional experience of the operation and maintenance personnel, and the diagnostic efficiency is not high, and the reliability needs to be improved. Summary of the invention

[0004] The purpose of the present invention is to provide an artificial intelligence-based equipment abnormality diagnosis method, system and program product to solve the above-mentioned problems existing in the prior art.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, a method for diagnosing abnormalities of equipment based on artificial intelligence is provided, comprising:

[0007] Acquire a current inspection record data set of a target device, wherein the current inspection record data set includes a plurality of current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data;

[0008] Retrieving from the database a historical normal inspection data group corresponding to each inspection data type number, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with a corresponding inspection data type number;

[0009] According to the inspection data type number, each current inspection record data is merged into the corresponding historical normal inspection data group to obtain the inspection data group to be analyzed corresponding to each inspection data type number, and the current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number;

[0010] Determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the array feature set to be analyzed corresponding to the inspection data type number based on each difference;

[0011] Input the array feature sets to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data anomaly classification results corresponding to each inspection data type number, wherein the data anomaly classification results include abnormal label data or normal label data;

[0012] When there is abnormal label data in the data abnormality classification results corresponding to each inspection data type number, a feature matrix is ​​formed using each inspection data type number and the data abnormality classification results corresponding to each inspection data type number;

[0013] The characteristic matrix is ​​input into a preset fault classification diagnosis model for fault classification diagnosis, and the fault classification result of the target device is obtained and output. The fault classification diagnosis model adopts a trained BP neural network model.

[0014] In a possible design, obtaining a current inspection record data set of a target device includes:

[0015] An initial inspection record data set of the target device is collected, and data cleaning and data deduplication processing are performed on the initial inspection record data set to obtain a current inspection record data set of the target device.

[0016] In a possible design, the current inspection record data is merged into the corresponding historical normal inspection data group according to the inspection data type number to obtain the inspection data group to be analyzed corresponding to each inspection data type number, including:

[0017] Each current inspection record data is added to the end of the corresponding historical normal inspection data group according to the inspection data type number, so as to obtain the inspection data group to be analyzed corresponding to each inspection data type number.

[0018] In a possible design, determining the to-be-analyzed array feature set corresponding to the inspection data type number based on each difference value includes:

[0019] Arrange the differences in sequence to obtain a difference group;

[0020] The difference mean and the difference standard deviation of each difference are calculated, and the difference mean and the difference standard deviation are added to the end of the difference group in sequence to obtain the array feature set to be analyzed corresponding to the inspection data type number.

[0021] In a possible design, before inputting the to-be-analyzed array feature sets corresponding to the inspection data type numbers into the trained KNN model for classification prediction, the method further includes:

[0022] Constructing a KNN model and obtaining a first training set of the KNN model, wherein the first training set includes a number of positive samples of the array feature set to be analyzed that are marked with abnormal label data and a number of negative samples of the array feature set to be analyzed that are marked with normal label data;

[0023] The KNN model is trained using the first training set to obtain a trained KNN model.

[0024] In a possible design, the feature matrix is ​​formed by using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number, including:

[0025] The inspection data type number and its corresponding data anomaly classification result are used to form corresponding data pairs;

[0026] Based on the set arrangement order of the inspection data type numbers, each data pair is used to form a feature matrix, each row of the feature matrix corresponds to a group of data pairs, the first column of the feature matrix contains each inspection data type number, and each inspection data type number is arranged according to the set arrangement order, and the second column of the feature matrix contains the data anomaly classification result corresponding to each inspection data type number.

[0027] In a possible design, before inputting the characteristic matrix into a preset fault classification diagnosis model for fault classification diagnosis, the method further includes:

[0028] Constructing a BP neural network model and obtaining a second training set of the BP neural network model, wherein the second training set includes a number of feature matrix samples marked with corresponding fault classification labels;

[0029] The BP neural network model is trained using the second training set to obtain a fault classification diagnosis model.

[0030] In a second aspect, an artificial intelligence-based equipment abnormality diagnosis system is provided, including a data acquisition unit, a data retrieval unit, a data merging unit, a feature extraction unit, an abnormality classification unit, a feature assembly unit and a fault diagnosis unit, wherein:

[0031] A data acquisition unit, used to acquire a current inspection record data set of a target device, wherein the current inspection record data set includes a plurality of current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data;

[0032] A data retrieval unit, used to retrieve from a database a historical normal inspection data group corresponding to each inspection data type number, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with a corresponding inspection data type number;

[0033] A data merging unit is used to merge each current inspection record data into a corresponding historical normal inspection data group according to the inspection data type number, to obtain an inspection data group to be analyzed corresponding to each inspection data type number, wherein the current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number;

[0034] A feature extraction unit, used to determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the feature set of the array to be analyzed corresponding to the inspection data type number based on each difference;

[0035] An abnormal classification unit is used to input the array feature set to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data abnormality classification result corresponding to each inspection data type number, wherein the data abnormality classification result includes abnormal label data or normal label data;

[0036] A feature forming unit, used to form a feature matrix using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number when there is abnormal label data in the data anomaly classification result corresponding to each inspection data type number;

[0037] The fault diagnosis unit is used to input the characteristic matrix into a preset fault classification diagnosis model for fault classification diagnosis, obtain the fault classification result of the target device, and output the fault classification result of the target device. The fault classification diagnosis model adopts a trained BP neural network model.

[0038] In a third aspect, a device abnormality diagnosis system based on artificial intelligence is provided, comprising:

[0039] A memory for storing instructions;

[0040] A processor is used to read the instructions stored in the memory and execute any one of the methods described in the first aspect according to the instructions.

[0041] In a fourth aspect, a computer-readable storage medium is provided, wherein instructions are stored on the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes any one of the methods described in the first aspect. In addition, a computer program product is provided, and when the computer program product is executed on a computer, the computer executes any one of the methods described in the first aspect.

[0042] Beneficial effect: The present invention obtains the inspection data group to be analyzed by merging the current inspection record data with the corresponding historical normal inspection data group according to the inspection data type number, and then extracts the array feature set of the inspection data group to be analyzed, and inputs the extracted array feature set into the KNN model for binary classification prediction to obtain the data abnormality classification result, and when there is an abnormal label in each classification result, the inspection data type number and the data abnormality classification result are used to form a feature matrix, and finally the feature matrix is ​​input into the fault classification diagnosis model for fault diagnosis to obtain the fault classification result of the target device, which can realize artificial intelligence equipment abnormality diagnosis. The present invention can replace the traditional manual inspection data analysis method, realize intelligent equipment inspection data analysis, and analyze the endogenous correlation between each abnormal data based on the neural network model, quickly and accurately diagnose the fault type of the equipment, effectively improve the abnormal diagnosis efficiency of the equipment, and ensure the reliability of the abnormal diagnosis of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only 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.

[0044] Figure 1 This is a schematic diagram of the steps of the method in Example 1 of the present invention;

[0045] Figure 2 This is a schematic diagram of the structure of the system in Example 2 of the present invention;

[0046] Figure 3 Schematic diagram of the system structure in Example 3 of the present invention. DETAILED DESCRIPTION

[0047] It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. The specific structures and functional details disclosed herein are only used to describe the exemplary embodiments of the present invention. However, the present invention can be embodied in many alternative forms, and it should not be understood that the present invention is limited to the embodiments set forth herein.

[0048] It should be understood that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments can be understood according to specific circumstances.

[0049] In the following description, certain details are provided to facilitate a complete understanding of the example embodiments. However, it will be appreciated by those of ordinary skill in the art that the example embodiments may be implemented without these certain details. For example, devices may be shown in block diagrams to avoid obscuring the examples with unnecessary details. In other embodiments, well-known processes, structures, and techniques may not be shown in unnecessary detail to avoid obscuring the embodiments.

[0050] Embodiment 1:

[0051] This embodiment provides an artificial intelligence-based device abnormality diagnosis method, which can be applied to corresponding device monitoring servers, such as Figure 1 As shown, the method comprises the following steps:

[0052] S1. Acquire a current inspection record data set of a target device, wherein the current inspection record data set includes several current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data.

[0053] In specific implementation, after completing the inspection task of the target device, the initial inspection record data set of the target device can be collected through the corresponding data recording terminal, and the initial inspection record data set can be cleaned and deduplicated to obtain the current inspection record data set of the target device. The device monitoring server obtains the current inspection record data set from the data recording terminal for subsequent analysis and processing. The current inspection record data set includes several current inspection record data groups, and the current inspection record data group includes interrelated inspection data type numbers and current inspection record data.

[0054] S2. Retrieve the historical normal inspection data group corresponding to each inspection data type number from the database, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with the corresponding inspection data type number.

[0055] In specific implementation, the equipment monitoring server can retrieve the historical normal inspection data group corresponding to each inspection data type number from the database, and the historical normal inspection data group contains a number of historical normal inspection data arranged in sequence. The database has a number of historical normal inspection data groups pre-stored, and each historical normal inspection data group is associated with a corresponding inspection data type number.

[0056] S3. According to the inspection data type number, each current inspection record data is merged into the corresponding historical normal inspection data group to obtain the inspection data group to be analyzed corresponding to each inspection data type number. The current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number.

[0057] In specific implementation, the equipment monitoring server can add each current inspection record data to the end of the corresponding historical normal inspection data group according to the inspection data type number to obtain the inspection data group to be analyzed corresponding to each inspection data type number. The current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number.

[0058] S4. Determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the array feature set to be analyzed corresponding to the inspection data type number based on each difference.

[0059] In specific implementation, the equipment monitoring server first calculates the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, such as subtracting the previous inspection data to be analyzed from the latter inspection data to be analyzed, to obtain the difference between two adjacent inspection data to be analyzed. Then, the differences are arranged in sequence to obtain a difference group, and then the difference mean and difference standard deviation of each difference are calculated, and the difference mean and difference standard deviation are added to the end of the difference group in sequence to obtain the feature set of the array to be analyzed corresponding to the inspection data type number. Based on the changes in the differences between adjacent inspection data to be analyzed, combined with the mean and standard deviation characteristics of each difference, it can effectively reflect the data change difference after the current inspection record data is added to the historical normal inspection data group, so that when the data anomaly is subsequently classified, it is based on the characteristics reflected by the feature set of the array to be analyzed to accurately detect whether the current inspection record data is abnormal.

[0060] S5. Input the array feature sets to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data anomaly classification results corresponding to each inspection data type number, wherein the data anomaly classification results include abnormal label data or normal label data.

[0061] In specific implementation, the device monitoring server can pre-build a KNN model and obtain a first training set of the KNN model, wherein the first training set includes a number of positive samples of the array feature set to be analyzed that are marked with abnormal label data and a number of negative samples of the array feature set to be analyzed that are marked with normal label data, and then the KNN model is trained using the first training set to obtain a trained KNN model. The core idea of ​​the KNN (K-NearestNeighbor, K nearest neighbor classification algorithm) model is that if most of the K nearest neighboring samples of a sample in the feature space belong to a certain category, then the sample also belongs to this category and has the characteristics of the samples in this category. Through this classifier, the category to which the input feature belongs can be quickly and efficiently determined, and the corresponding data anomaly classification result, that is, abnormal label data or normal label data, can be obtained.

[0062] S6. When there is abnormal label data in the data anomaly classification results corresponding to each inspection data type number, a feature matrix is ​​formed using each inspection data type number and the data anomaly classification results corresponding to each inspection data type number.

[0063] In specific implementation, when the device monitoring server detects that there is abnormal label data in the data anomaly classification results corresponding to each inspection data type number, it means that at least one of the current inspection record data corresponding to the inspection data type number is abnormal, that is, the target device may have a certain fault that causes the inspection record data to be abnormal, but in some cases, a device fault may cause the inspection record data corresponding to multiple inspection data type numbers to be abnormal, that is, there may be some endogenous relationship between multiple abnormal inspection record data. Based on this, in order to accurately detect the fault type corresponding to different abnormal inspection record data, this embodiment also needs to use each inspection data type number and the data anomaly classification results corresponding to each inspection data type number to form a feature matrix, so that the feature matrix can be used to analyze the endogenous relationship between abnormal inspection record data and diagnose the fault type. The process of forming a feature matrix using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number includes: forming corresponding data pairs using the inspection data type number and the data anomaly classification result corresponding to each inspection data type number; based on the set inspection data type number arrangement order, forming a feature matrix using each data pair, each row of the feature matrix corresponds to a group of data pairs, the first column of the feature matrix contains each inspection data type number, and the inspection data type numbers are arranged in a set arrangement order, and the second column of the feature matrix contains the data anomaly classification result corresponding to each inspection data type number.

[0064] S7. Input the characteristic matrix into a preset fault classification diagnosis model to perform fault classification diagnosis, obtain the fault classification result of the target device, and output the fault classification result of the target device. The fault classification diagnosis model adopts a trained BP neural network model.

[0065] In specific implementation, the equipment monitoring server can pre-build a BP (back propagation) neural network model and obtain a second training set of the BP neural network model, wherein the second training set contains a number of feature matrix samples marked with corresponding fault classification labels, and then the BP neural network model is trained using the second training set to obtain a fault classification diagnosis model. The BP neural network is a multi-layer feedforward neural network, which mainly includes an input layer, a hidden layer and an output layer. There can be multiple hidden layers, each of which contains a number of neurons. The input layer can receive multiple feature matrices as input, and the output layer outputs a calculation result through a corresponding activation function. When performing fault diagnosis, the equipment monitoring server inputs the feature matrix into the fault classification diagnosis model for fault classification diagnosis, analyzes the endogenous relationship between abnormal inspection record data through the fault classification diagnosis model, determines the possible fault type that causes the abnormal inspection record data of the target device, and outputs the fault classification result of the target device, i.e., the corresponding fault classification label data, so that the operation and maintenance personnel can intuitively obtain the type of the target device according to the fault classification result and perform corresponding processing.

[0066] The method of this embodiment can replace the traditional manual inspection data analysis method, realize intelligent equipment inspection data analysis, and analyze the intrinsic correlation between various abnormal data based on the neural network model, quickly and accurately diagnose the fault type of the equipment, effectively improve the abnormal diagnosis efficiency of the equipment, and ensure the reliability of abnormal diagnosis of the equipment.

[0067] Embodiment 2:

[0068] This embodiment provides an artificial intelligence-based device abnormality diagnosis system. Figure 2 As shown, it includes a data acquisition unit, a data retrieval unit, a data merging unit, a feature extraction unit, an abnormal classification unit, a feature building unit and a fault diagnosis unit, wherein:

[0069] A data acquisition unit, used to acquire a current inspection record data set of a target device, wherein the current inspection record data set includes a plurality of current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data;

[0070] A data retrieval unit, used to retrieve from a database a historical normal inspection data group corresponding to each inspection data type number, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with a corresponding inspection data type number;

[0071] A data merging unit is used to merge each current inspection record data into a corresponding historical normal inspection data group according to the inspection data type number, to obtain an inspection data group to be analyzed corresponding to each inspection data type number, wherein the current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number;

[0072] A feature extraction unit, used to determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the feature set of the array to be analyzed corresponding to the inspection data type number based on each difference;

[0073] An abnormal classification unit is used to input the array feature set to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data abnormality classification result corresponding to each inspection data type number, wherein the data abnormality classification result includes abnormal label data or normal label data;

[0074] A feature forming unit, used to form a feature matrix using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number when there is abnormal label data in the data anomaly classification result corresponding to each inspection data type number;

[0075] The fault diagnosis unit is used to input the characteristic matrix into a preset fault classification diagnosis model for fault classification diagnosis, obtain the fault classification result of the target device, and output the fault classification result of the target device. The fault classification diagnosis model adopts a trained BP neural network model.

[0076] Embodiment 3:

[0077] This embodiment provides an artificial intelligence-based device abnormality diagnosis system. Figure 3 As shown, at the hardware level, it includes:

[0078] Data interface, used to establish data connection between the processor and the data recording end;

[0079] A memory for storing instructions;

[0080] The processor is used to read the instructions stored in the memory and execute the artificial intelligence-based device abnormality diagnosis method in Example 1 according to the instructions.

[0081] Optionally, the system further includes an internal bus, through which the processor, the memory and the data interface can be interconnected, and the internal bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0082] The memory may include, but is not limited to, random access memory (RAM), read only memory (ROM), flash memory, first input first output (FIFO) and / or first in last out (FILO) memory, etc. The processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0083] Embodiment 4:

[0084] This embodiment provides a computer-readable storage medium, on which instructions are stored, and when the instructions are executed on a computer, the computer executes the artificial intelligence-based device abnormality diagnosis method in Embodiment 1. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive, and / or a memory stick, etc., and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0085] This embodiment also provides a computer program product, which, when executed on a computer, executes the artificial intelligence-based device abnormality diagnosis method in Embodiment 1. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0086] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A device abnormality diagnosis method based on artificial intelligence, characterized in that: include: Acquire a current inspection record data set of a target device, wherein the current inspection record data set includes a plurality of current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data; Retrieving from the database a historical normal inspection data group corresponding to each inspection data type number, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with a corresponding inspection data type number; According to the inspection data type number, each current inspection record data is merged into the corresponding historical normal inspection data group to obtain the inspection data group to be analyzed corresponding to each inspection data type number, and the current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number; Determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the array feature set to be analyzed corresponding to the inspection data type number based on each difference, wherein the determination of the array feature set to be analyzed corresponding to the inspection data type number based on each difference includes: arranging the differences in sequence to obtain a difference group; calculating the difference mean and the difference standard deviation of each difference, and adding the difference mean and the difference standard deviation to the end of the difference group in sequence to obtain the array feature set to be analyzed corresponding to the inspection data type number; Input the array feature sets to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data anomaly classification results corresponding to each inspection data type number, wherein the data anomaly classification results include abnormal label data or normal label data; When there is abnormal label data in the data abnormality classification results corresponding to each inspection data type number, a feature matrix is ​​formed using each inspection data type number and the data abnormality classification results corresponding to each inspection data type number; The characteristic matrix is ​​input into a preset fault classification diagnosis model for fault classification diagnosis, and the fault classification result of the target device is obtained and output. The fault classification diagnosis model adopts a trained BP neural network model.

2. The method for diagnosing abnormality of equipment based on artificial intelligence according to claim 1, characterized in that: The step of obtaining a current inspection record data set of the target device includes: An initial inspection record data set of the target device is collected, and data cleaning and data deduplication processing are performed on the initial inspection record data set to obtain a current inspection record data set of the target device.

3. The method for diagnosing abnormality of equipment based on artificial intelligence according to claim 1, characterized in that: The method of merging each current inspection record data into the corresponding historical normal inspection data group according to the inspection data type number to obtain the inspection data group to be analyzed corresponding to each inspection data type number includes: Each current inspection record data is added to the end of the corresponding historical normal inspection data group according to the inspection data type number, so as to obtain the inspection data group to be analyzed corresponding to each inspection data type number.

4. The device abnormality diagnosis method based on artificial intelligence according to claim 1 is characterized in that: Before inputting the to-be-analyzed array feature sets corresponding to the inspection data type numbers into the trained KNN model for classification prediction, the method further includes: Constructing a KNN model and obtaining a first training set of the KNN model, wherein the first training set includes a number of positive samples of the array feature set to be analyzed that are marked with abnormal label data and a number of negative samples of the array feature set to be analyzed that are marked with normal label data; The KNN model is trained using the first training set to obtain a trained KNN model.

5. The method for diagnosing abnormality of equipment based on artificial intelligence according to claim 1, characterized in that: The feature matrix is ​​formed by using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number, including: The inspection data type number and its corresponding data anomaly classification result are used to form corresponding data pairs; Based on the set arrangement order of the inspection data type numbers, each data pair is used to form a feature matrix, each row of the feature matrix corresponds to a group of data pairs, the first column of the feature matrix contains each inspection data type number, and each inspection data type number is arranged according to the set arrangement order, and the second column of the feature matrix contains the data anomaly classification result corresponding to each inspection data type number.

6. The method for diagnosing abnormality of equipment based on artificial intelligence according to claim 1, characterized in that: Before inputting the characteristic matrix into a preset fault classification diagnosis model for fault classification diagnosis, the method further includes: Constructing a BP neural network model and obtaining a second training set of the BP neural network model, wherein the second training set includes a number of feature matrix samples marked with corresponding fault classification labels; The BP neural network model is trained using the second training set to obtain a fault classification diagnosis model.

7. An artificial intelligence-based equipment abnormality diagnosis system, characterized in that: It includes a data acquisition unit, a data retrieval unit, a data merging unit, a feature extraction unit, an abnormal classification unit, a feature building unit and a fault diagnosis unit, wherein: A data acquisition unit, used to acquire a current inspection record data set of a target device, wherein the current inspection record data set includes a plurality of current inspection record data groups, and the current inspection record data groups include mutually related inspection data type numbers and current inspection record data; A data retrieval unit, used to retrieve from a database a historical normal inspection data group corresponding to each inspection data type number, wherein the historical normal inspection data group includes a number of historical normal inspection data arranged in sequence, and each historical normal inspection data group is associated with a corresponding inspection data type number; A data merging unit is used to merge each current inspection record data into a corresponding historical normal inspection data group according to the inspection data type number, to obtain an inspection data group to be analyzed corresponding to each inspection data type number, wherein the current inspection record data and each historical normal inspection data in the inspection data group to be analyzed are used as the inspection data to be analyzed with the corresponding inspection data type number; The feature extraction unit is used to determine the difference between two adjacent inspection data to be analyzed in the inspection data group to be analyzed, and determine the array feature set to be analyzed corresponding to the inspection data type number based on each difference, wherein the determination of the array feature set to be analyzed corresponding to the inspection data type number based on each difference comprises: arranging each difference in sequence to obtain a difference group; calculating the difference mean and the difference standard deviation of each difference, and sequentially adding the difference mean and the difference standard deviation to the end of the difference group to obtain the array feature set to be analyzed corresponding to the inspection data type number; An abnormal classification unit is used to input the array feature set to be analyzed corresponding to each inspection data type number into the trained KNN model for classification prediction, and obtain the data abnormality classification result corresponding to each inspection data type number, wherein the data abnormality classification result includes abnormal label data or normal label data; A feature forming unit, used to form a feature matrix using each inspection data type number and the data anomaly classification result corresponding to each inspection data type number when there is abnormal label data in the data anomaly classification result corresponding to each inspection data type number; The fault diagnosis unit is used to input the characteristic matrix into a preset fault classification diagnosis model for fault classification diagnosis, obtain the fault classification result of the target device, and output the fault classification result of the target device. The fault classification diagnosis model adopts a trained BP neural network model.

8. An artificial intelligence-based equipment abnormality diagnosis system, characterized in that: include: A memory for storing instructions; A processor is used to read the instructions stored in the memory and execute the device abnormality diagnosis method described in any one of claims 1-6 according to the instructions.

9. A computer program product, characterized in that When the computer program product is run on a computer, the device abnormality diagnosis method according to any one of claims 1 to 6 is executed.

Images