Industrial control system data feature extraction method, fault diagnosis method and related device

Abstract

The invention belongs to the field of industrial control systems, and discloses an industrial control system data feature extraction method, a fault diagnosis method and a related device, and the method comprises the steps: obtaining the operation data of an industrial control system; inputting the operation data of the industrial control system into a preset improved sparse auto-encoder to obtain feature data of the operation data of the industrial control system; wherein the improved sparse auto-encoder is obtained by replacing the process of determining the weight of a sparse penalty term in a loss function of the sparse auto-encoder with the process of optimizing by adopting a particle swarm optimization algorithm. Compared with an existing manual determination mode, the particle swarm optimization algorithm is adopted for optimization, the sparse penalty term weight can be further optimized, the feature extraction effect can be directly influenced by selection of the sparse penalty term weight, and therefore high-precision effective extraction of the data features of the industrial control system can be effectively improved by optimizing the sparse penalty term weight, and the extraction efficiency of the industrial control system is improved. Therefore, intelligent algorithms such as fault diagnosis and mode recognition can be called, and the fault diagnosis precision of the intelligent fault diagnosis method can be improved.

Description

Technical Field

[0001] The invention belongs to the field of industrial control systems, and relates to an industrial control system data feature extraction method, a fault diagnosis method and related devices. Background Art

[0002] The method for detecting abnormal operating status of industrial control systems (ICS) during operation mainly relies on the experience of operation and maintenance personnel. By setting thresholds for key operating parameters and detection parameters, an alarm signal is triggered once the parameter changes exceed the threshold range. The operation and maintenance personnel then trace the cause of the fault based on the alarm information.

[0003] However, this method has the problem that the diagnosis speed and accuracy are restricted by the professional knowledge of the operation and maintenance personnel. Especially when a large number of alarm signals appear at the same time, it is impossible to quickly diagnose the cause of the fault from the massive alarm information in a short period of time relying on manpower, which may lead to more serious accident consequences.

[0004] In view of this, the data-driven intelligent fault diagnosis method does not rely on knowledge and experience, and can start from the operating data features, with the advantages of fast diagnosis speed and high diagnostic accuracy. However, the implementation effect of the intelligent fault diagnosis method is closely related to the accuracy of the extracted data features. The existing data feature extraction methods are often not accurate in extracting features from high-dimensional and massive data, which leads to low fault diagnosis accuracy of the intelligent fault diagnosis method. Summary of the invention

[0005] The purpose of the present invention is to overcome the above-mentioned shortcomings of the prior art and to provide an industrial control system data feature extraction method, a fault diagnosis method and related devices.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] According to a first aspect of the present invention, a method for extracting features from industrial control system data is provided, comprising: obtaining operating data of the industrial control system; inputting the operating data of the industrial control system into a preset improved sparse autoencoder to obtain feature data of the operating data of the industrial control system; wherein the improved sparse autoencoder is obtained by replacing the process of determining the weights of sparse penalty terms in the loss function of the sparse autoencoder with a process of optimizing and determining the weights using a particle swarm optimization algorithm.

[0008] Optionally, before inputting the industrial control system operation data into a preset improved sparse autoencoder, the method further includes: performing data correction processing and normalization processing on the industrial control system operation data.

[0009] Optionally, the normalization process specifically includes: performing normalization process using a Z-Score method.

[0010] Optionally, the step of inputting the industrial control system operation data into a preset improved sparse autoencoder to obtain characteristic data of the industrial control system operation data includes: S11: inputting the industrial control system operation data into a preset improved sparse autoencoder to obtain the loss function value of the improved sparse autoencoder; S12: when the loss function value of the improved sparse autoencoder is greater than the preset loss function setting value, adjusting the weight and bias coefficient of the improved sparse autoencoder, and returning to S11; S13: when the loss function value of the improved sparse autoencoder is not greater than the preset loss function setting value, recording the output of the hidden layer of the improved sparse autoencoder; and using a particle swarm optimization algorithm to optimize the weight of the sparse penalty term in the loss function of the improved sparse autoencoder; S14: iterating S11 to S13 to a preset number of iterations, and selecting the output of the hidden layer of the improved sparse autoencoder with the smallest loss function value among all recorded outputs of the hidden layer of the improved sparse autoencoder as the characteristic data of the industrial control system operation data.

[0011] Optionally, it also includes: dividing the industrial control system operation data and the characteristic data of the industrial control system operation data into normal operation data and abnormal operation data and storing them in partitions; wherein the normal operation data is divided into normal steady-state operation data and normal transient operation data and stored in partitions; the abnormal operation data is classified according to the fault type label and stored in partitions.

[0012] A second aspect of the present invention provides an industrial control system fault diagnosis method, comprising: using the above-mentioned industrial control system data feature extraction method to obtain feature data of the industrial control system operation data to be diagnosed; inputting the feature data of the industrial control system operation data to be diagnosed into a pre-trained industrial control system fault diagnosis model to obtain the industrial control system fault diagnosis result.

[0013] According to a third aspect of the present invention, a data feature extraction system for an industrial control system is provided, comprising: a data acquisition module for acquiring operating data of the industrial control system; a feature extraction module for inputting the operating data of the industrial control system into a preset improved sparse autoencoder to obtain feature data of the operating data of the industrial control system; wherein the improved sparse autoencoder is obtained by replacing the process of determining the weight of the sparse penalty term in the loss function of the sparse autoencoder with that determined by optimizing using a particle swarm optimization algorithm.

[0014] In a fourth aspect, the present invention provides an industrial control system fault diagnosis system, comprising: the above-mentioned industrial control system data feature extraction system, used to obtain feature data of the industrial control system operation data to be diagnosed; a fault diagnosis module, used to input the feature data of the industrial control system operation data to be diagnosed into a pre-trained industrial control system fault diagnosis model to obtain the industrial control system fault diagnosis result.

[0015] According to a fifth aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned industrial control system data feature extraction method or the above-mentioned industrial control system fault diagnosis method when executing the computer program.

[0016] According to a sixth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program implements the steps of the above-mentioned industrial control system data feature extraction method, or implements the steps of the above-mentioned industrial control system fault diagnosis method.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The industrial control system data feature extraction method of the present invention, after obtaining the industrial control system operation data, inputs it into a preset improved sparse autoencoder to obtain the feature data of the industrial control system operation data, and realizes the industrial control system data feature extraction based on the improved sparse autoencoder. Among them, the improved sparse autoencoder is obtained by replacing the determination process of the sparse penalty item weight in the loss function of the sparse autoencoder with the particle swarm optimization algorithm for optimization. Compared with the current manual method of determining the sparse penalty item weight, the particle swarm optimization algorithm can be used for optimization to further optimize the sparse penalty item weight, and the selection of the sparse penalty item weight will directly affect the feature extraction effect. Therefore, by optimizing the sparse penalty item weight, the high-precision and effective extraction of the industrial control system data features can be effectively improved, so as to be called by intelligent algorithms such as fault diagnosis and pattern recognition, thereby improving the fault diagnosis accuracy of the intelligent fault diagnosis method. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The figure is a flow chart of a method for extracting data features from an industrial control system according to an embodiment of the present invention.

[0020] Figure 2 This is a detailed flow chart of data feature extraction of an industrial control system according to an embodiment of the present invention.

[0021] Figure 3 The present invention is a flowchart of a method for diagnosing faults in an industrial control system according to an embodiment of the present invention.

[0022] Figure 4 It is a structural block diagram of the industrial control system data feature extraction system according to an embodiment of the present invention.

[0023] Figure 5 The figure is a structural block diagram of the fault diagnosis of the industrial control system according to the embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0027] See also Figure 1 In one embodiment of the present invention, a method for extracting data features of an industrial control system is provided, specifically, a method for extracting data features of an industrial control system based on an improved sparse autoencoder. When the existing data feature extraction methods realize feature extraction of high-dimensional and massive data, the accuracy of feature extraction is often not high, which leads to low fault diagnosis accuracy of the intelligent fault diagnosis method. By extracting feature data of high-precision industrial control system operation data, the fault diagnosis accuracy of the intelligent fault diagnosis method can be improved.

[0028] Specifically, the industrial control system data feature extraction method includes the following steps:

[0029] S1: Obtain the operating data of the industrial control system.

[0030] S2: Input the industrial control system operation data into a preset improved sparse autoencoder to obtain feature data of the industrial control system operation data.

[0031] Among them, the improved sparse autoencoder is obtained by replacing the determination process of the sparse penalty term weight in the loss function of the sparse autoencoder with an optimization determination using a particle swarm optimization algorithm.

[0032] The industrial control system data feature extraction method of the present invention, after obtaining the industrial control system operation data, inputs it into a preset improved sparse autoencoder to obtain the feature data of the industrial control system operation data, and realizes the industrial control system data feature extraction based on the improved sparse autoencoder. Among them, the improved sparse autoencoder is obtained by replacing the determination process of the sparse penalty item weight in the loss function of the sparse autoencoder with the particle swarm optimization algorithm for optimization. Compared with the current manual method of determining the sparse penalty item weight, the particle swarm optimization algorithm can be used for optimization to further optimize the sparse penalty item weight, and the selection of the sparse penalty item weight will directly affect the feature extraction effect. Therefore, by optimizing the sparse penalty item weight, the high-precision and effective extraction of the industrial control system data features can be effectively improved, so as to be called by intelligent algorithms such as fault diagnosis and pattern recognition, thereby improving the fault diagnosis accuracy of the intelligent fault diagnosis method.

[0033] In a possible implementation manner, before inputting the industrial control system operation data into the preset improved sparse autoencoder, the method further includes performing data correction processing and normalization processing on the industrial control system operation data.

[0034] Specifically, the data collection of industrial control system operation data can use IO cards to collect parameters such as switch quantity, analog quantity and pulse quantity, filter the collected industrial control system operation data, and select the industrial control system operation data that is closely related to the operation status of the industrial control system for storage. The storage of industrial control system operation data can be transmitted to the historical database in the centralized control platform through the local area network for data storage.

[0035] Among them, data correction processing includes filling, deleting and correcting data missing, data redundancy and data errors.

[0036] Normalization can be performed using the Z-Score method. The Z-Score method, also known as standard deviation normalization, is a common data preprocessing method that can convert the original data into a distribution with a mean of 0 and a standard deviation of 1, thereby eliminating the dimensional effects between different features and making the data more suitable for subsequent data analysis and modeling. In this way, the original data is converted into a Z-Score form, that is, a distribution with a mean of 0 and a standard deviation of 1. This processing method can retain the distribution information of the original data while reducing the numerical differences between different features, making it easier to compare and analyze. The advantage of the Z-Score method is that it is simple and easy to use, and has a clear mathematical meaning. By performing normalization with the Z-Score method, the dimensional effects between different features can be eliminated, making the data easier to process and understand. At the same time, it can also improve the performance of certain machine learning algorithms, such as logistic regression and support vector machines.

[0037] In a possible implementation manner, the step of inputting the industrial control system operation data into a preset improved sparse autoencoder to obtain feature data of the industrial control system operation data includes:

[0038] S11: Input the operating data of the industrial control system into a preset improved sparse autoencoder to obtain a loss function value of the improved sparse autoencoder.

[0039] S12: When the loss function value of the improved sparse autoencoder is greater than the preset loss function setting value, adjust the weight and bias coefficient of the improved sparse autoencoder, and return to S11.

[0040] S13: When the loss function value of the improved sparse autoencoder is not greater than the preset loss function setting value, the output of the hidden layer of the improved sparse autoencoder is recorded; and the particle swarm optimization algorithm is used to optimize the sparse penalty term weight in the loss function of the improved sparse autoencoder.

[0041] S14: Iterate S11 to S13 to a preset number of iterations, and select the output of the hidden layer of the improved sparse autoencoder with the smallest loss function value among all the recorded outputs of the hidden layer of the improved sparse autoencoder as the feature data of the industrial control system operation data.

[0042] Specifically, the improved sparse autoencoder has been further optimized and improved compared to the traditional sparse autoencoder, and a particle swarm optimization algorithm is used to perform self-optimization operations on some parameters in the sparse autoencoder.

[0043] The implementation process of sparse autoencoder can be divided into the following two parts:

[0044] Coding process: h1 = σ e (W1x+b1)

[0045] Decoding process: y = σ d (W2h1+b2)

[0046] Where W1 and b1 are the weight and bias of encoding; W2 and b2 are the weight and bias of decoding; σ e and σ d is a nonlinear transformation, x is the input data, h1 is the output data of the hidden layer, and y is the reconstructed data.

[0047] The loss function of the traditional autoencoder is: the objective function of minimizing the error between y and x expressed as mean square error, which can be expressed as J(W,b):

[0048]

[0049] Where L(x,y) is the mean square error of x and y.

[0050] In order to achieve the suppression effect, the sparse autoencoder constrains the average activation value of the hidden layer neuron output, uses the KL divergence to force it to be close to a given sparse value, and adds it to the loss function as a penalty term. Therefore, the loss function of the sparse autoencoder can be expressed as J SAE (W):

[0051]

[0052] Among them, β is the weight of the sparse penalty term, is the sparse penalty term, for KL divergence, h is the number of hidden layer neurons, ρ is the activation value of the input data, is the average activation value of the input data on the hidden layer neuron j.

[0053] Taking β as the variable, the further improvement of the loss function optimization of the sparse autoencoder based on the particle swarm optimization algorithm is:

[0054] J SAE (W) = pso(β)

[0055] Among them, pso is the particle swarm optimization algorithm function.

[0056] See also Figure 2 , the process of improving the sparse autoencoder to extract the features of the industrial control system operation data:

[0057] The operating data of the industrial control system is used as the input parameter to initialize the parameters of the improved sparse autoencoder, and the input layer nodes are mapped to the hidden layer nodes using the activation function for encoding. Then, the hidden layer nodes are mapped to the output layer nodes through the activation function to realize the decoding process.

[0058] Taking the reconstruction results and input parameters of the output layer nodes as input, the loss function value Loss of the improved sparse autoencoder is calculated. If the Loss value is less than the loss function set value SetValue, the calculation result meets the requirements, and the data of the hidden layer nodes is output as the feature data of the industrial control system operation data. Otherwise, the weights and biases of the improved sparse encoder are adjusted for iterative calculation. The weights and biases here include encoding and decoding parts, until the loss function value calculation result meets the requirements.

[0059] Since the sparse penalty weight β in the sparse autoencoder needs to be set manually, the choice of the sparse penalty weight β will directly affect the feature extraction effect.

[0060] Therefore, the present invention further improves the sparse autoencoder based on the particle swarm optimization algorithm, sets the maximum number of iterations MaxValue, and when the number of iterations t is not less than MaxValue, the particle swarm optimization algorithm is used to further optimize the sparse penalty item weight β until the number of iterations t reaches the requirement, that is, greater than MaxValue, and outputs the output of the hidden layer of the improved sparse autoencoder with the smallest Loss value among all iteration results as the feature data of the industrial control system operation data.

[0061] In a possible implementation manner, the industrial control system data feature extraction method further includes:

[0062] The industrial control system operation data and the characteristic data of the industrial control system operation data are divided into normal operation data and abnormal operation data and stored in partitions; among them, the normal operation data is divided into normal steady-state operation data and normal transient operation data and stored in partitions; the abnormal operation data is classified according to the fault type label and stored in partitions.

[0063] Specifically, the data storage process follows the principle of operating condition separation. The steady-state operating data and transient operating data of normal operating conditions are stored separately according to the corresponding operating condition states. For the operating data under abnormal operating conditions, they are stored separately according to the fault type of the abnormal condition.

[0064] Among them, the characteristic data of the industrial control system operation data is also stored according to the same storage settings, so as to facilitate subsequent data calls according to different operating conditions and adapt intelligent algorithms such as fault diagnosis and pattern recognition.

[0065] See also Figure 3 In another embodiment of the present invention, a method for diagnosing faults in an industrial control system is provided, which realizes high-precision fault diagnosis based on feature data extracted by the above-mentioned method for extracting feature data of industrial control system.

[0066] Specifically, the industrial control system fault diagnosis method includes the following steps:

[0067] S20: Using the above-mentioned industrial control system data feature extraction method, characteristic data of the operating data of the industrial control system to be diagnosed is obtained.

[0068] S21: Inputting feature data of the operating data of the industrial control system to be diagnosed into a pre-trained industrial control system fault diagnosis model to obtain a fault diagnosis result of the industrial control system.

[0069] Specifically, the industrial control system fault diagnosis model can be built using a neural network model, such as BP neural network, convolutional neural network and recurrent neural network, to achieve the mapping of the characteristic data of the industrial control system operation data to the industrial control system fault type. The pre-training of the industrial control system fault diagnosis model is to use the characteristic data of the labeled historical industrial control system data and the corresponding fault type label as training data for training.

[0070] The following are device embodiments of the present invention, which can be used to implement the method embodiments of the present invention. For details not disclosed in the device embodiments, please refer to the method embodiments of the present invention.

[0071] See also Figure 4 In another embodiment of the present invention, a data feature extraction system for an industrial control system is provided, which can be used to implement the above-mentioned data feature extraction method for an industrial control system. Specifically, the data feature extraction system for an industrial control system includes a data acquisition module and a feature extraction module.

[0072] Among them, the data acquisition module is used to obtain the operating data of the industrial control system; the feature extraction module is used to input the operating data of the industrial control system into a preset improved sparse autoencoder to obtain the feature data of the operating data of the industrial control system; wherein, the improved sparse autoencoder is obtained by replacing the process of determining the weight of the sparse penalty term in the loss function of the sparse autoencoder with the particle swarm optimization algorithm for optimization.

[0073] In a possible implementation manner, before inputting the industrial control system operation data into a preset improved sparse autoencoder, the method further includes: performing data correction processing and normalization processing on the industrial control system operation data.

[0074] In a possible implementation manner, the normalization process specifically includes: performing normalization process using a Z-Score method.

[0075] In a possible implementation manner, the step of inputting the industrial control system operation data into a preset improved sparse autoencoder to obtain feature data of the industrial control system operation data includes:

[0076] S11: Input the operating data of the industrial control system into a preset improved sparse autoencoder to obtain a loss function value of the improved sparse autoencoder.

[0077] S12: When the loss function value of the improved sparse autoencoder is greater than the preset loss function setting value, adjust the weight and bias coefficient of the improved sparse autoencoder, and return to S11.

[0078] S13: When the loss function value of the improved sparse autoencoder is not greater than the preset loss function setting value, the output of the hidden layer of the improved sparse autoencoder is recorded; and the particle swarm optimization algorithm is used to optimize the sparse penalty term weight in the loss function of the improved sparse autoencoder.

[0079] S14: Iterate S11 to S13 to a preset number of iterations, and select the output of the hidden layer of the improved sparse autoencoder with the smallest loss function value among all the recorded outputs of the hidden layer of the improved sparse autoencoder as the feature data of the industrial control system operation data.

[0080] In a possible implementation, it also includes a storage module for dividing the industrial control system operation data and the characteristic data of the industrial control system operation data into normal operation data and abnormal operation data and storing them in partitions; wherein the normal operation data is divided into normal steady-state operation data and normal transient operation data and stored in partitions; and the abnormal operation data is classified according to the fault type label and stored in partitions.

[0081] All relevant contents of each step involved in the above-mentioned embodiment of the industrial control system data feature extraction method can be referred to the functional description of the functional module corresponding to the industrial control system data feature extraction system in the embodiment of the present invention, and will not be repeated here.

[0082] See also Figure 5 In another embodiment of the present invention, an industrial control system fault diagnosis system is provided, which can be used to implement the above-mentioned industrial control system fault diagnosis method. Specifically, the industrial control system fault diagnosis system includes an industrial control system data feature extraction system and a fault diagnosis module.

[0083] Among them, the industrial control system data feature extraction system is used to obtain the feature data of the industrial control system operation data to be diagnosed; the fault diagnosis module is used to input the feature data of the industrial control system operation data to be diagnosed into the pre-trained industrial control system fault diagnosis model to obtain the industrial control system fault diagnosis result.

[0084] All relevant contents of each step involved in the above-mentioned embodiment of the industrial control system fault diagnosis method can be referred to the functional description of the functional modules corresponding to the industrial control system fault diagnosis system in the embodiment of the present invention, and will not be repeated here.

[0085] The division of modules in the embodiments of the present invention is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present invention may be integrated into one processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0086] In another embodiment of the present invention, a computer device is provided, the computer device including a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of the industrial control system data feature extraction method or the industrial control system fault diagnosis method.

[0087] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the industrial control system data feature extraction method or the industrial control system fault diagnosis method in the above-mentioned embodiment.

[0088] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0089] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0092] 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 above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for extracting data features from an industrial control system, characterized in that: include: Obtaining the operation data of industrial control system; Inputting the operation data of the industrial control system into a preset improved sparse autoencoder to obtain feature data of the operation data of the industrial control system; Among them, the improved sparse autoencoder is obtained by replacing the determination process of the sparse penalty term weight in the loss function of the sparse autoencoder with an optimization determination using a particle swarm optimization algorithm.

2. The method for extracting features from industrial control system data according to claim 1, characterized in that: Before inputting the industrial control system operation data into the preset improved sparse autoencoder, the method further includes: The industrial control system operation data is corrected and normalized.

3. The method for extracting features from industrial control system data according to claim 2, characterized in that: The normalization process specifically includes: performing normalization process using a Z-Score method.

4. The method for extracting features from industrial control system data according to claim 1, characterized in that: The step of inputting the industrial control system operation data into a preset improved sparse autoencoder to obtain the characteristic data of the industrial control system operation data includes: S11: inputting the operation data of the industrial control system into a preset improved sparse autoencoder to obtain a loss function value of the improved sparse autoencoder; S12: When the loss function value of the improved sparse autoencoder is greater than the preset loss function setting value, adjust the weight and bias coefficient of the improved sparse autoencoder, and return to S11; S13: when the loss function value of the improved sparse autoencoder is not greater than the preset loss function setting value, the output of the hidden layer of the improved sparse autoencoder is recorded; and the sparse penalty term weight in the loss function of the improved sparse autoencoder is optimized by using a particle swarm optimization algorithm; S14: Iterate S11 to S13 to a preset number of iterations, and select the output of the hidden layer of the improved sparse autoencoder with the smallest loss function value among all the recorded outputs of the hidden layer of the improved sparse autoencoder as the feature data of the industrial control system operation data.

5. The method for extracting industrial control system data features according to claim 1, characterized in that: Also includes: The industrial control system operation data and the characteristic data of the industrial control system operation data are divided into normal operation data and abnormal operation data and stored in partitions; among them, the normal operation data is divided into normal steady-state operation data and normal transient operation data and stored in partitions; the abnormal operation data is classified according to the fault type label and stored in partitions.

6. A method for diagnosing faults in an industrial control system, characterized in that: include: Using the industrial control system data feature extraction method according to any one of claims 1 to 5, characteristic data of the operating data of the industrial control system to be diagnosed is obtained; The characteristic data of the operating data of the industrial control system to be diagnosed is input into the pre-trained industrial control system fault diagnosis model to obtain the industrial control system fault diagnosis result.

7. An industrial control system data feature extraction system, characterized in that: include: Data acquisition module, used to obtain the operation data of the industrial control system; A feature extraction module is used to input the operation data of the industrial control system into a preset improved sparse autoencoder to obtain feature data of the operation data of the industrial control system; Among them, the improved sparse autoencoder is obtained by replacing the determination process of the sparse penalty term weight in the loss function of the sparse autoencoder with an optimization determination using a particle swarm optimization algorithm.

8. An industrial control system fault diagnosis system, characterized in that: include: The industrial control system data feature extraction system of claim 7 is used to obtain feature data of the operating data of the industrial control system to be diagnosed; The fault diagnosis module is used to input the characteristic data of the operating data of the industrial control system to be diagnosed into the pre-trained industrial control system fault diagnosis model to obtain the industrial control system fault diagnosis result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the industrial control system data feature extraction method according to any one of claims 1 to 5 are implemented, or the steps of the industrial control system fault diagnosis method according to claim 6 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the industrial control system data feature extraction method according to any one of claims 1 to 5 are implemented, or the steps of the industrial control system fault diagnosis method according to claim 6 are implemented.

Images