Fault diagnosis model training method and device, equipment and medium

Abstract

Embodiments of the present application disclose a fault diagnosis model training method and device, equipment and a medium, relating to the technical field of fault diagnosis. The method comprises: obtaining reference sample data generated by a sample device under normal operation; determining a fault type for fault diagnosis of the sample device; generating fault correlation data according to the reference sample data and the fault type; and generating fault diagnosis training data according to the fault correlation data, for training a fault diagnosis model. The above scheme introduces a fault type, determines fault correlation data, and then determines fault diagnosis training data, thereby improving the accuracy of the determined fault diagnosis training data, and further improving the training accuracy of the fault diagnosis model and the accuracy of subsequent fault diagnosis using the trained fault diagnosis model.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of fault diagnosis technology, and in particular to a fault diagnosis model training method, apparatus, equipment and medium. Background Technology

[0002] In recent years, fault diagnosis technology for industrial processes has been frequently used to monitor industrial processes, providing timely judgment on whether abnormal signals have occurred in industrial processes.

[0003] In existing technologies, fault diagnosis is typically performed using models. Therefore, improving the training accuracy of these models, and consequently the accuracy of fault diagnosis based on the trained models, is crucial. Summary of the Invention

[0004] This invention provides a fault diagnosis model training method, apparatus, equipment, and medium to improve the training accuracy of the model, thereby improving the accuracy of fault diagnosis based on the trained model.

[0005] According to one aspect of the present invention, a fault diagnosis model training method is provided, characterized in that it includes:

[0006] Acquire reference sample data generated by the sample equipment under normal operating conditions;

[0007] Determine the type of fault for fault diagnosis of the sample device;

[0008] Based on the reference sample data and the fault type, fault association data is generated;

[0009] Based on the fault association data, fault diagnosis training data is generated to train the fault diagnosis model.

[0010] According to another aspect of the present invention, a fault diagnosis model training apparatus is provided, comprising:

[0011] The reference sample data acquisition module is used to acquire reference sample data generated by the sample device under normal operating conditions.

[0012] The fault type determination module is used to determine the fault type of the sample device for fault diagnosis.

[0013] The fault association data generation module is used to generate fault association data based on the reference sample data and the fault type;

[0014] The fault diagnosis model training module is used to generate fault diagnosis training data based on the fault association data, and to train the fault diagnosis model.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] At least one processor; and

[0017] A memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute the fault diagnosis model training method according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the fault diagnosis model training method according to any embodiment of the present invention.

[0020] This invention provides a fault diagnosis model training scheme. The scheme involves acquiring reference sample data generated by a sample device under normal operating conditions; determining the fault type for fault diagnosis of the sample device; generating fault association data based on the reference sample data and the fault type; and generating fault diagnosis training data based on the fault association data for training the fault diagnosis model. This scheme improves the accuracy of the determined fault diagnosis training data by introducing fault types, determining fault association data, and then determining the fault diagnosis training data. This, in turn, improves the training accuracy of the fault diagnosis model and the accuracy of subsequent fault diagnosis using the trained fault diagnosis model.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a fault diagnosis model training method provided in Embodiment 1 of the present invention;

[0024] Figure 2 This is a flowchart of a fault diagnosis model training method provided in Embodiment 2 of the present invention;

[0025] Figure 3This is a schematic diagram of the structure of a fault diagnosis model training device provided in Embodiment 3 of the present invention;

[0026] Figure 4 This is a schematic diagram of the structure of an electronic device for implementing a fault diagnosis model training method, provided in Embodiment 4 of the present invention. Detailed Implementation

[0027] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0028] Example 1

[0029] Figure 1 This is a flowchart of a fault diagnosis model training method provided in Embodiment 1 of the present invention. This embodiment is applicable to the training of fault diagnosis models. The method can be executed by a fault diagnosis device, which can be implemented in hardware and / or software. The device can be configured in an electronic device that carries the fault diagnosis model training function.

[0030] See Figure 1 The fault diagnosis model training method shown includes:

[0031] S110. Obtain reference sample data generated by the sample device under normal operating conditions.

[0032] In this context, "sample equipment" refers to industrial equipment that can provide reference sample data. Reference sample data refers to data that can be used to train a fault diagnosis model. "Normal operation" refers to the sample equipment operating without faults. This embodiment of the invention does not limit the method of acquiring reference sample data; it can be set by technicians based on experience. For example, sensors can be used to acquire reference sample data.

[0033] It should be noted that when using sensors to acquire reference sample data, since different types of sensors have different sensitivities to different fault types, in order to improve the diversity and comprehensiveness of the reference sample data, at least one type of sensor can be used to acquire the reference sample data.

[0034] Specifically, it involves acquiring reference sample data generated at different times when the sample equipment is operating normally.

[0035] S120. Determine the type of fault for fault diagnosis of the sample equipment.

[0036] Here, fault type refers to the direction or type of fault diagnosis required. This invention does not limit the types and / or number of fault types; these can be set by technicians based on experience. In an optional embodiment, the fault type can be determined based on the type of sensor.

[0037] S130. Generate fault association data based on reference sample data and fault type.

[0038] Fault-related data refers to data related to fault diagnosis generated based on reference sample data and fault types. For example, fault-related data can be a set of feature vectors.

[0039] S140. Generate fault diagnosis training data based on fault correlation data to train the fault diagnosis model.

[0040] In this context, fault diagnosis training data refers to data that can be used to provide a basis for adjusting the parameters of the fault diagnosis model. For example, fault diagnosis training data can be represented in matrix form. Specifically, fault diagnosis training data can be a projection matrix (or projected data).

[0041] This invention provides a fault diagnosis model training scheme. The scheme involves acquiring reference sample data generated by a sample device under normal operating conditions; determining the fault type for fault diagnosis of the sample device; generating fault association data based on the reference sample data and the fault type; and generating fault diagnosis training data based on the fault association data for training the fault diagnosis model. This scheme improves the accuracy of the determined fault diagnosis training data by introducing fault types, determining fault association data, and then determining the fault diagnosis training data. This, in turn, improves the training accuracy of the fault diagnosis model and the accuracy of subsequent fault diagnosis using the trained fault diagnosis model.

[0042] Example 2

[0043] Figure 2 This is a flowchart of a fault diagnosis model training method provided in Embodiment 2 of the present invention. Based on the above embodiments, this embodiment further refines the operation of "generating fault association data based on reference sample data and fault type" into "generating intra-fault class association data based on reference sample data; generating inter-fault class association data for different fault types based on reference sample data and fault type; obtaining inter-fault class association data based on inter-fault class association data; generating fault association data including intra-fault class association data and inter-fault class association data," thereby improving the fault association data determination mechanism. It should be noted that parts not detailed in this embodiment can be found in the descriptions of other embodiments.

[0044] See Figure 2 The fault diagnosis model training method shown includes:

[0045] S210. Obtain reference sample data generated by the sample device under normal operating conditions.

[0046] S220. Determine the type of fault for fault diagnosis of the sample equipment.

[0047] S230. Generate fault class-related data based on reference sample data.

[0048] Among them, the fault class intra-association data is used to characterize the intra-class aggregation of reference sample data collected at any given time. For example, the fault class intra-association data can be represented in matrix form. Specifically, the fault class intra-association data (or fault class intra-association matrix) can be determined using the following formula:

[0049]

[0050] Among them, S W Indicates associated data within the fault class; e n This represents an n-dimensional vector consisting entirely of 1s; T represents torque. Let X represent the tensor Kroneck product; X represents the set of reference sample data, X∈R d×n ; d represents the dimension or the number of sensors; n represents the number of reference sample data in the time series, that is, the number of reference sample data at a certain moment.

[0051] It should be noted that the purpose of generating intra-class correlation data for faults is to enable intra-class aggregation of reference sample data obtained at different times, thereby avoiding the excessive dispersion of reference sample data collected at different times.

[0052] Specifically, based on the reference sample data obtained at different times, fault class correlation data is generated for the corresponding time.

[0053] S240. Based on the reference sample data and fault type, generate correlation data between sub-fault classes of different fault types.

[0054] The inter-class correlation data of sub-faults is used to characterize the inter-class dispersion of different fault types and reference sample data. For example, the inter-class correlation data of sub-faults can be represented in matrix form. Specifically, the inter-class correlation data of sub-faults (or the inter-class correlation matrix) can be determined using the following formula:

[0055] S Fi =(Ξ i ·f i (Ξ) i ·f i ) T ;

[0056] Among them, S Fi This represents the inter-fault class association data formed by the i-th type of fault; Ξ i The direction vector representing the fault type; f i This represents the fault amplitude of the i-th fault type; i represents the i-th dimension or sensor.

[0057] It should be noted that when determining the correlation data between sub-fault classes based on the above formula, an expectation E can be added to make the determined correlation data between sub-fault classes more accurate. Specifically, the expression after adding the expectation E is:

[0058] S Fi =(Ξ i ·E(f i ))(Ξ i ·E(f i )) T ;

[0059] Specifically, based on the determined fault type, the sensors sensitive to that fault type are identified. For example, if there are d sensors, and the sensitive directions for the first s sensors correspond to a given fault type, then the remaining ds sensors are insensitive to the determined fault type. Therefore, the fault-sensitive direction vectors are Ξ1 to Ξ... s The following conditions must be met:

[0060] Ξ1 = [100…0] T

[0061] Ξ2 = [0100…0] T

[0062]

[0063] Furthermore, for the fault direction vector corresponding to a sensor that is insensitive to a given fault type, the following condition must be satisfied:

[0064]

[0065]

[0066]

[0067] Here, ε represents a very small positive number, which can avoid matrix ill-conditioning problems in the process of solving the generalized eigenvalue problem, and also ensure that the dimension of the subspace is not constrained.

[0068] Furthermore, to reduce computational load, it can be assumed that each sensor direction is equal and may experience the same amplitude fault. When more accurate fault information cannot be obtained, the fault amplitudes of each fault type can be set to be equal, i.e., ||f1||=||f2||=…||f i ||, where ||f i || indicates the fault amplitude.

[0069] In one optional embodiment, generating inter-fault class correlation data for different fault types based on reference sample data and fault types includes: normalizing the reference sample data so that the mean of each reference sample data is a default value; and generating inter-fault class correlation data for different fault types based on fault types.

[0070] In this embodiment of the invention, the size of the default value is not limited and can be set by a technician based on experience. For example, the default value can be 0. Similarly, the normalization method is not limited and can be set by a technician based on experience. For instance, if the default value is 0, zero-mean normalization and unit variance normalization can be used.

[0071] Understandably, by introducing normalization, the mean of each reference sample data is set to the default value, which improves the efficiency of determining the correlation data between sub-fault classes and reduces the amount of computation.

[0072] S250. Obtain the inter-fault class association data based on the inter-fault class association data.

[0073] Here, inter-fault class correlation data refers to the collection of correlation data between sub-fault classes. For example, inter-fault class correlation data can be represented in matrix form. Specifically, the inter-fault class correlation data (or inter-fault class correlation matrix) can be determined using the following formula:

[0074]

[0075] Among them, S F This represents the data associated with different fault classes, that is, the collection of data associated with different fault types.

[0076] S260. Generate fault association data that includes fault association data within fault classes and fault association data between fault classes.

[0077] It should be noted that the data related to faults within and between fault classes stores feature vectors.

[0078] S270. Based on the fault correlation data, generate fault diagnosis training data for training the fault diagnosis model.

[0079] In one optional embodiment, generating fault diagnosis training data based on fault association data includes: determining fault association feature data based on fault association data within fault classes and fault association data between fault classes; obtaining fault diagnosis feature data based on the fault association feature data; and generating fault diagnosis training data based on the fault diagnosis feature data.

[0080] In this context, fault-related feature data refers to the eigenvalues ​​of each feature vector in the fault-related data. For example, fault-related feature data can be represented by an eigenvalue diagonal matrix. Fault-diagnostic feature data refers to the eigenvalues ​​of at least some feature vectors in the fault-related data. For example, fault-diagnostic feature data can be represented by a maximum eigenvalue diagonal matrix.

[0081] For example, according to Fault-related feature data are determined using the following formula:

[0082] S F B=λS W B;

[0083] Where tr represents the trace of the matrix; λ represents the eigenvalue; and B represents the fault diagnosis training data.

[0084] It is understandable that by introducing fault-related feature data, fault diagnosis feature data can be obtained, and then fault diagnosis training data can be generated based on the fault diagnosis feature data. This ensures the comprehensiveness of the fault diagnosis training data while reducing its dimensionality and improving the efficiency of subsequent processing.

[0085] In one optional embodiment, obtaining fault diagnosis feature data based on fault association feature data includes: selecting a large preset threshold number of fault association feature data as fault diagnosis feature data; correspondingly, generating fault diagnosis training data based on the fault diagnosis feature data includes: determining fault association data corresponding to the fault diagnosis feature data; and using the fault association data corresponding to the fault diagnosis feature data as fault diagnosis training data.

[0086] In this embodiment of the invention, the size of the preset threshold is not limited in any way and can be set by a technician based on experience. For example, the preset threshold can be the subspace dimension, and the size of the subspace dimension can be determined according to preset rules. This embodiment of the invention also does not limit the preset rules, which can be set by a technician based on experience. For example, the preset rules can be the cumulative percentage variance (CPV) criterion or the Akaike information criterion (AIC).

[0087] Specifically, fault-related feature data with larger feature values ​​in the subspace dimension are selected from the fault-related feature data and used as fault diagnosis feature data; the fault-related data corresponding to the fault diagnosis feature data are determined and used as fault diagnosis training data.

[0088] In an optional embodiment, the fault association data can be viewed as a metaspace. This metaspace is then dimensionality-reduced to obtain a subspace containing at least some of the feature vectors from the metaspace. Fault detection for corresponding fault types is then performed using the determined subspace to improve fault diagnosis performance. The metaspace contains feature vectors representing each fault type, while the subspace contains feature vectors representing more important fault types.

[0089] Understandably, by introducing a preset threshold, the dimensionality of fault-related data is reduced, thereby lowering the dimensionality of fault diagnosis training data and facilitating subsequent fault diagnosis.

[0090] The fault diagnosis model training scheme provided in this invention refines the operation of generating fault association data based on reference sample data and fault type into generating intra-fault class association data based on reference sample data; generating inter-fault class association data for different fault types based on reference sample data and fault type; obtaining inter-fault class association data based on inter-fault class association data; and generating fault association data including both intra-fault class association data and inter-fault class association data, thus improving the fault association data determination mechanism. This scheme, by introducing intra-fault class association data and inter-fault class association data to determine fault association data, achieves the determination of corresponding data under both normal operation and fault occurrence scenarios, improving the comprehensiveness and accuracy of the determined fault association data.

[0091] Based on the above technical solution, the method further includes: determining a fault detection threshold for fault diagnosis according to a preset threshold; and determining the fault status of reference sample data according to the fault detection threshold.

[0092] The fault detection threshold can be used to determine whether the data is faulty. For example, the fault detection threshold could be T. 2 The control limits corresponding to the statistic. Fault states can include both faulty and faultless states.

[0093] For example, when the fault detection threshold is T 2 The control time limit corresponding to the statistic can be determined using the following formula to determine the fault detection threshold:

[0094]

[0095] in, T represents 2The control limits corresponding to the statistic; p represents the subspace dimension; F represents the F-distribution; α represents the significance level; η 2 Indicates control limits.

[0096] Understandably, by introducing a fault detection threshold, the fault state can be accurately determined, thus improving the accuracy of the determined fault state.

[0097] In this embodiment of the invention, when training the fault diagnosis model, the input data may include reference sample data X∈R. d×n The direction vector Ξ of the fault type i Fault amplitude of fault type ||f i ||, preset threshold (or subspace dimension p) and significance level α; the output data may include fault diagnosis training data B, fault diagnosis feature data Λ, and fault detection threshold (or T) 2 (Control limits corresponding to the statistic).

[0098] In an optional embodiment, the method further includes: obtaining a fault diagnosis model; wherein the fault diagnosis model is trained using a fault diagnosis model training method; obtaining the data to be tested of the device under test at the current time; inputting the data to be tested into the fault diagnosis model to determine the fault state of the data to be tested.

[0099] The data to be tested refers to the data that requires fault diagnosis. The equipment to be tested refers to the equipment that requires fault diagnosis.

[0100] In one optional embodiment, when the fault diagnosis model is used, the input data may include the data to be detected, x∈R. d The fault diagnosis training data B and fault diagnosis feature data Λ obtained from the training, as well as the fault detection threshold (or T) 2 (Control limits corresponding to the statistics); the output data is the fault status of the data to be detected.

[0101] Specifically, when using the fault diagnosis model, the T corresponding to the data to be detected can be determined using the following formula. 2 Statistic:

[0102] T 2 =x T BΛ -1 F T x;

[0103] Furthermore, if If the test result is positive, the fault status of the data to be tested is determined to be faulty; otherwise, the fault status of the data to be tested is determined to be fault-free, meaning the data to be tested is normal.

[0104] It is understandable that by using the fault diagnosis model trained by the above method, the accuracy of the determined fault state is improved when performing fault diagnosis on the data to be detected.

[0105] It should be noted that preprocessing of the input data is required when training or using the fault diagnosis model. This embodiment of the invention does not impose any limitations on the preprocessing method; it can be set by technical personnel based on experience.

[0106] Example 3

[0107] Figure 3 This is a schematic diagram of a fault diagnosis model training device provided in Embodiment 3 of the present invention. This embodiment is applicable to the training of fault diagnosis models. The method can be executed by a fault diagnosis device, which can be implemented in hardware and / or software. The device can be configured in an electronic device that carries the fault diagnosis model training function.

[0108] like Figure 3 As shown, the device includes: a reference sample data acquisition module 310, a fault type determination module 320, a fault association data generation module 330, and a fault diagnosis model training module 340. Among them,

[0109] Reference sample data acquisition module 310 is used to acquire reference sample data generated by the sample device under normal operating conditions;

[0110] The fault type determination module 320 is used to determine the fault type of the sample device for fault diagnosis.

[0111] The fault association data generation module 330 is used to generate fault association data based on reference sample data and fault type;

[0112] The fault diagnosis model training module 340 is used to generate fault diagnosis training data based on fault correlation data, which is used to train the fault diagnosis model.

[0113] This invention provides a fault diagnosis model training scheme. The scheme involves: acquiring reference sample data generated by a sample device under normal operating conditions using a reference sample data acquisition module; determining the fault type for fault diagnosis using a fault type determination module; generating fault association data based on the reference sample data and fault type using a fault association data generation module; and generating fault diagnosis training data based on the fault association data using a fault diagnosis model training module. This scheme improves the accuracy of the determined fault diagnosis training data by introducing fault type, determining fault association data, and then determining the fault diagnosis training data. This, in turn, improves the training accuracy of the fault diagnosis model and the accuracy of subsequent fault diagnosis using the trained fault diagnosis model.

[0114] Optionally, the fault correlation data generation module 330 includes:

[0115] The intra-class association data generation unit is used to generate intra-class association data of faults based on reference sample data;

[0116] The sub-class association data generation unit is used to generate inter-fault class association data for different fault types based on reference sample data and fault types;

[0117] The inter-class association data generation unit is used to obtain inter-fault class association data based on the inter-fault class association data of sub-fault classes;

[0118] The associated data generation unit is used to generate fault association data, including intra-fault association data and inter-fault association data.

[0119] Optional, inter-class association data generation unit, specifically used for:

[0120] The reference sample data is normalized so that the mean of each reference sample data is the default value;

[0121] Based on the fault type, generate correlation data between sub-fault classes of different fault types.

[0122] Optional, the fault diagnosis model training module 340 includes:

[0123] The fault association feature data determination unit is used to determine fault association feature data based on the association data within fault classes and the association data between fault classes.

[0124] The fault diagnosis feature data determination unit is used to obtain fault diagnosis feature data based on fault association feature data.

[0125] The fault diagnosis training data determination unit is used to generate fault diagnosis training data based on fault diagnosis feature data.

[0126] Optionally, the fault diagnosis feature data determination unit is specifically used for:

[0127] Select a relatively large number of fault-related feature data with a preset threshold as fault diagnosis feature data;

[0128] Accordingly, based on the fault diagnosis feature data, fault diagnosis training data is generated, including:

[0129] Determine the fault-related data corresponding to the fault diagnosis feature data;

[0130] Fault-related data corresponding to fault diagnosis feature data are used as fault diagnosis training data.

[0131] Optionally, the device may also include:

[0132] The fault detection threshold determination unit is used to determine the fault detection threshold for fault diagnosis based on a preset threshold.

[0133] The fault status determination unit is used to determine the fault status of the reference sample data based on the fault detection threshold.

[0134] Optionally, the device may also include:

[0135] The model acquisition module is used to acquire the fault diagnosis model; the fault diagnosis model is trained using a fault diagnosis model training method.

[0136] The test data acquisition module is used to acquire the test data of the device under test at the current moment.

[0137] The fault status determination module is used to input the data to be detected into the fault diagnosis model and determine the fault status of the data to be detected.

[0138] The fault diagnosis model training device provided in this embodiment of the invention can execute the fault diagnosis model training method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing each fault diagnosis model training method.

[0139] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision and disclosure of reference sample data, fault types, and data to be tested all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0140] Example 4

[0141] Figure 4 This is a schematic diagram of the structure of an electronic device for implementing a fault diagnosis model training method according to Embodiment 4 of the present invention. The electronic device 410 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0142] like Figure 4As shown, the electronic device 410 includes at least one processor 411 and a memory, such as a read-only memory (ROM) 412 or a random access memory (RAM) 413, communicatively connected to the at least one processor 411. The memory stores computer programs executable by the at least one processor. The processor 411 can perform various appropriate actions and processes based on the computer program stored in the ROM 412 or loaded from storage unit 418 into the RAM 413. The RAM 413 may also store various programs and data required for the operation of the electronic device 410. The processor 411, ROM 412, and RAM 413 are interconnected via a bus 414. An input / output (I / O) interface 415 is also connected to the bus 414.

[0143] Multiple components in electronic device 410 are connected to I / O interface 415, including: input unit 416, such as keyboard, mouse, etc.; output unit 417, such as various types of displays, speakers, etc.; storage unit 418, such as disk, optical disk, etc.; and communication unit 419, such as network card, modem, wireless transceiver, etc. Communication unit 419 allows electronic device 410 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0144] Processor 411 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 411 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 411 performs the various methods and processes described above, such as fault diagnosis model training methods.

[0145] In some embodiments, the fault diagnosis model training method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 418. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 410 via ROM 412 and / or communication unit 419. When the computer program is loaded into RAM 413 and executed by processor 411, one or more steps of the fault diagnosis model training method described above may be performed. Alternatively, in other embodiments, processor 411 may be configured to execute the fault diagnosis model training method by any other suitable means (e.g., by means of firmware).

[0146] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0147] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0148] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0149] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0150] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0151] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0152] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0153] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for training a fault diagnosis model, characterized in that, include: Acquire reference sample data generated by the sample equipment under normal operating conditions; Determine the type of fault for fault diagnosis of the sample device; Based on the reference sample data and the fault type, fault association data is generated; Based on the fault association data, fault diagnosis training data is generated for training the fault diagnosis model. The step of generating fault association data based on the reference sample data and the fault type includes: Based on the reference sample data, fault class intra-association data is generated; wherein, the fault class intra-association data is used to characterize the intra-class aggregation of reference sample data collected at any collection time; Based on the reference sample data and the fault type, inter-class association data for different fault types is generated; wherein, the inter-class association data for different fault types is used to characterize the inter-class dispersion of different fault types and the reference sample data; Based on the inter-sub-fault class correlation data, inter-fault class correlation data is obtained; wherein, the inter-fault class correlation data refers to the set of inter-sub-fault class correlation data. Generate fault association data that includes intra-fault association data and inter-fault association data.

2. The method according to claim 1, characterized in that, The step of generating inter-fault class association data for different fault types based on the reference sample data and the fault type includes: The reference sample data is normalized so that the mean of each reference sample data is the default value. Based on the fault type, generate inter-fault class association data for different fault types.

3. The method according to claim 1, characterized in that, The step of generating fault diagnosis training data based on the fault association data includes: Based on the intra-fault association data and the inter-fault association data, fault association feature data is determined; Based on the fault association feature data, fault diagnosis feature data is obtained; Based on the fault diagnosis feature data, fault diagnosis training data is generated.

4. The method according to claim 3, characterized in that, The step of obtaining fault diagnosis feature data based on the fault association feature data includes: Select a relatively large number of fault-related feature data with a preset threshold as fault diagnosis feature data; Accordingly, generating fault diagnosis training data based on the fault diagnosis feature data includes: Determine the fault association data corresponding to the fault diagnosis feature data; The fault association data corresponding to the fault diagnosis feature data is used as fault diagnosis training data.

5. The method according to claim 4, characterized in that, The method further includes: Based on the preset threshold, a fault detection threshold for fault diagnosis is determined; The fault status of the reference sample data is determined based on the fault detection threshold.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: Obtain a fault diagnosis model; wherein the fault diagnosis model is trained using the method described in any one of claims 1-5; Acquire the test data of the device under test at the current moment; The data to be detected is input into the fault diagnosis model to determine the fault status of the data to be detected.

7. A fault diagnosis model training device, characterized in that, include: The reference sample data acquisition module is used to acquire reference sample data generated by the sample device under normal operating conditions. The fault type determination module is used to determine the fault type of the sample device for fault diagnosis. The fault association data generation module is used to generate fault association data based on the reference sample data and the fault type; The fault diagnosis model training module is used to generate fault diagnosis training data based on the fault association data, and to train the fault diagnosis model. The fault-related data generation module includes: The intra-class correlation data generation unit is used to generate intra-class correlation data of faults based on reference sample data; wherein, the intra-class correlation data of faults is used to characterize the intra-class aggregation of reference sample data collected at any collection time; The sub-class association data generation unit is used to generate sub-fault class association data for different fault types based on reference sample data and fault types; wherein, the sub-fault class association data is used to characterize the inter-class dispersion of different fault types and reference sample data; The inter-class association data generation unit is used to obtain inter-fault class association data based on the inter-fault class association data of sub-fault classes; wherein, the inter-fault class association data refers to the set of inter-fault class association data of sub-fault classes. The associated data generation unit is used to generate fault association data, including intra-fault association data and inter-fault association data.

8. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a fault diagnosis model training method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a fault diagnosis model training method as described in any one of claims 1-6.

Images