An industrial fault diagnosis method based on domain generalization

Abstract

The application relates to an industrial process fault diagnosis method based on field generalization and applied to unknown field fault diagnosis in an industrial scene. Known field data are collected, a pre-training discriminator is obtained by training a generative adversarial network, a multi-discriminator generative adversarial network is constructed, the multi-discriminator generative adversarial network is used to generate data with generalization characteristics, a fault diagnosis model is constructed and trained by using the generated generalization data and the known field data, and the fault diagnosis model is used for online diagnosis of a new field industrial process. The industrial process fault diagnosis method based on field generalization is mainly aimed at the situation that prior knowledge and data are lacking in unknown fields in an industrial scene, and realizes field generalization of fault diagnosis by combining a generative adversarial network and a supervised fault diagnosis model. Successful application of the method is expected to improve the accuracy of fault diagnosis, reduce maintenance cost, and bring substantial benefits to the industrial field.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of industrial process fault diagnosis, and specifically relates to an industrial process fault diagnosis method based on domain generalization. BACKGROUND

[0002] Currently, in industrial production processes, fault diagnosis is a key task to ensure the normal operation of equipment and the production of high-quality products. However, due to the complexity and diversity of industrial scenarios, fault diagnosis for unknown domains has always been a challenge. Traditional fault diagnosis methods usually rely on prior knowledge and a large amount of known domain data to train models, which limits their applicability in unknown domains. In the prior art, some research attempts to use generative adversarial networks to solve the problem of domain generalization in fault diagnosis, which is a powerful deep learning model that can generate realistic data samples. However, some limitations in existing methods have limited their practical application in industrial processes. One of the main challenges is how to achieve high-performance fault diagnosis for unknown domains in the absence of prior knowledge and data.

[0003] Therefore, it is necessary to propose an industrial process fault diagnosis method based on domain generalization, which aims to overcome the limitations of lacking prior knowledge and data for unknown domains by training generative adversarial networks using similar domain data and generating data with generalization characteristics. Such a solution is expected to improve the accuracy and efficiency of industrial process fault diagnosis, reduce maintenance costs, and bring substantial technological progress and economic benefits to the industrial field. SUMMARY

[0004] The application provides an industrial process fault diagnosis method based on domain generalization, which aims to solve the challenge of fault diagnosis for unknown domains in industrial scenarios in the absence of prior knowledge and data. The method trains generative adversarial networks using similar domain data to generate data with generalization characteristics, and combines known domain data to build and train a fault diagnosis model.

[0005] The technical solution adopted by the application to achieve the above-mentioned purposes is as follows:

[0006] An industrial fault diagnosis method based on domain generalization trains generative adversarial networks using similar domain data to generate universal data with generalization characteristics, and combines known domain data to build and train a fault diagnosis model, which is used for fault diagnosis for unknown domains in industrial scenarios in the absence of prior knowledge and data, including the following steps:

[0007] Step 1: Collect sample data of multiple similar domains of the unknown industrial process to be diagnosed, and label the fault labels respectively;

[0008] Step 2: Construct a single discriminator using a generative adversarial network, input each known domain sample data for training, and obtain a pre-trained discriminator corresponding to the known domain;

[0009] Step 3: Construct a multi-discriminator generative adversarial network using all known domain corresponding pre-trained discriminators;

[0010] Step 4: Use the multi-discriminator generative adversarial network to generate data with generalization characteristics using all known domain data as input, to expand the training set and enhance the generalization ability of the model;

[0011] Step 5: Construct a fault diagnosis model and use the generated generalization data and known domain data to train the model to have the generalization ability for unknown domain industrial processes;

[0012] Step 6: Use the trained fault diagnosis model to perform online diagnosis on new domain industrial processes to detect and identify possible faults in real time.

[0013] The data collection method comprises the following steps:

[0014] Determine the running data of the unknown industrial process and the adjacent known domain that needs to be collected; the running data includes sensor or monitoring device data, controller data, and mechanical component data;

[0015] Configure the device location, sampling frequency, and data transmission method for data collection;

[0016] Real-time data collection and monitoring of the running state of the industrial process to obtain sample data at multiple time points;

[0017] Preprocess the collected sample data;

[0018] Use the data labeling method to label each sample data with the corresponding fault label to construct a labeled training data set.

[0019] The data labeling method includes analyzing and judging the sensor data, controller data, and mechanical component data to obtain specific fault labels, and using operation logs and fault report information to infer and label fault labels.

[0020] The multi-discriminator generative adversarial network comprises a plurality of pre-trained discriminators DIS i and an untrained discriminator DIS, and an untrained generator G; the pre-trained discriminators DIS iwhether the generated data belongs to the known domain; the discriminator DIS for determining whether the generated data is real data; the generator G for generating data satisfying the following conditions: the pre-trained discriminators DIS1, DIS2 and the generated sample do not belong to the corresponding known domain; the discriminator DIS determines that the generated sample is real data;

[0021] According to the probability that the discriminator DIS maximizes the real sample into true and the false sample into false, the generator G minimizes the probability that the discriminator DIS identifies as false and the pre-trained discriminator identifies as belonging to the known domain, and the optimization objective is defined as

[0022]

[0023] where x is a real sample, z is random noise, denotes the expectation of the real data distribution, denotes the expectation under the random distribution, and log is the logarithmic function.

[0024] The single pre-trained discriminator DIS i is a discriminator after convergence of a general generative adversarial network, the general generative adversarial network includes a single generator and a single discriminator, both of which adopt a deep neural network structure including a recurrent neural network and a multilayer perceptron;

[0025] The input of the generator network G is random noise z, and the output is the generated data sample G(z); the learning goal of the generator network is to make the generated data sample G(z) as close as possible to the real data distribution x, and have similar characteristics and distribution characteristics as the known domain data;

[0026] The input of the discriminator network DIS is the generated data sample G(z) and the real data sample x of the known domain, and the output is the discrimination result of the input data, indicating whether it is real data; the learning goal of the discriminator network is to accurately discriminate the difference between the generated data sample and the real data sample, and provide feedback about the data to help the generator network continuously improve the generation of universal data.

[0027] According to the optimization objective, the following loss function is defined:

[0028] For the discriminator DIS, the real data x and the generated data G(z) are input into the discriminator respectively, and the loss function is calculated, and the discriminator parameters are updated as

[0029]

[0030] wherein, denotes the loss function for discriminator update, x is a real sample, z is random noise, denotes the expectation under the real data distribution, denotes the expectation under the random distribution, and log is the logarithmic function.

[0031] For several training discriminators DIS i , do not participate in the training of the multi-discriminator generative adversarial network;

[0032] For the generator G, input random noise z, input the generated data G(z) into the total discriminator DIS and each pre-trained discriminator DIS i , calculate the joint loss function and update the generator parameters:

[0033]

[0034] wherein, denotes the loss function for generator update, z is random noise, denotes the expectation under the random distribution, and log is the logarithmic function.

[0035] Repeat the above steps until the joint loss function of the generator meets the threshold or converges, or reaches the predetermined iteration training cutoff; after training is completed, the generator G accepts any noise vector z as input and generates unlabelled data x with generalization characteristics g .

[0036] The fault diagnosis model adopts an encoder-decoder network structure, and a stack of recurrent neural networks and convolutional neural networks is used as the encoder, and a fully connected network is used as the decoder.

[0037] The training of the fault diagnosis model includes the following steps:

[0038] The input of the fault diagnosis model is defined as the collected data x, and the output is the fault state y of the industrial process;

[0039] The generated unlabelled data x with generalization characteristics g is used to pre-train the fault diagnosis model;

[0040] The data x of the known field is used to supervise the training of the fault diagnosis model.

[0041] The method for online diagnosis of the new field industrial process includes the following steps:

[0042] The real-time key parameter data collected from the new field industrial process is preprocessed and used as the input of the fault diagnosis model;

[0043] Step 6-3: Feature matching and conversion are performed on the preprocessed real-time data, known field data and generated generalization data to ensure the consistency and comparability of the input data;

[0044] Step 6-4: Diagnose and analyze the real-time data that has been feature-matched and converted using the trained fault diagnosis model to determine whether there is a fault in the current industrial process;

[0045] According to the output result of the fault diagnosis model, the fault of the industrial process is judged and classified;

[0046] The fault diagnosis of the new field industrial process is continuously monitored, new data samples are collected and the fault diagnosis model is updated to continuously improve the diagnosis accuracy.

[0047] The present application has the following beneficial effects and advantages:

[0048] The present application is an industrial process fault diagnosis method based on field generalization, which can significantly improve the accuracy of fault diagnosis, reduce maintenance cost and improve work efficiency. When facing unknown field industrial processes, it can more accurately detect and diagnose possible faults, effectively reducing misdiagnosis and missed diagnosis. The generalization ability of the fault diagnosis model reduces the dependence on prior knowledge and data. Reducing the data collection and analysis workload in the maintenance process improves the efficiency of fault diagnosis and saves manpower and time cost. The present application improves the accuracy of fault diagnosis and reduces maintenance cost, which can avoid production interruption and downtime caused by faults in the industrial field, and improve the reliability and stability of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The present application is a method flowchart.

[0050] Figure 2 The present application is a multi-discriminator generative adversarial network structure diagram. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation method of the present application will be described in detail below. In the following description, a lot of specific details are set forth in order to fully understand the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the spirit of the present application, so the present application is not limited to the specific implementation disclosed below.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application.

[0053] As Figure 1As shown, it is a method flowchart of the application. An industrial process fault diagnosis method based on domain generalization collects monitoring data of multiple similar known industrial processes in a new domain to be diagnosed, uses a generative adversarial network to obtain a pre-trained discriminator for each known domain, constructs a multi-discriminator generative adversarial network, generates data with domain generalization properties, trains a fault diagnosis model using the generated generalization data and known domain data, and applies it to online fault diagnosis in the new domain. The programming language used in the program execution steps of the application is not limited to MATLAB, Python, etc.

[0054] The specific steps of the application are as follows:

[0055] Step 1: Collect multiple similar known domain data of the unknown industrial process using an efficient data collection method, wherein each known domain data contains a series of samples and corresponding fault labels. Taking the tobacco primary processing process as an example, the specific steps include the following steps:

[0056] Step 1-1: Determine the characteristics and parameters of the unknown industrial process to determine the adjacent known domains that need to be collected. In the tobacco primary processing process, characteristics and parameters can include temperature, humidity, speed, etc. Assuming there are two similar known domains, representing different cigarette brands, then the relevant data of these two brands needs to be collected.

[0057] Step 1-2: Select appropriate sensors or monitoring equipment and deploy them in the industrial process to collect real-time operation data of the industrial process. In order to collect data of the tobacco primary processing process, temperature sensors, humidity sensors, speed sensors, etc. can be used. These sensors will be installed on the primary processing equipment and record relevant operation data in real time.

[0058] Step 1-3: Configure the data collection system, including the location of the sensors, sampling frequency, data transmission method, etc., to ensure accurate and efficient data collection. When configuring the data collection system of the primary processing process, the location of the sensors needs to be determined to ensure that the key parameters in the industrial process can be accurately monitored. In addition, appropriate sampling frequency needs to be set to obtain enough data points, and appropriate data transmission method needs to be selected to ensure the real-time and reliability of the data.

[0059] Step 1-4: Start data collection and continuously monitor the running state of the industrial process to obtain sample data at multiple time points. After the data collection system of the primary processing process is configured, data collection can begin. The sensors will continuously monitor the running state of the industrial process and record sample data at multiple time points.

[0060] Step 1-5: Preprocess the collected sample data, including denoising, filtering, normalization, etc., to improve the quality and usability of the data. The tobacco cut process data collected by the sensor has interference, in order to improve the quality and usability of the data, the collected sample data needs to be preprocessed. This includes removing possible noise, filtering to smooth the data, and normalization to make the data have consistent scale and range.

[0061] Step 1-6: Use data labeling methods to analyze and judge the sensor data, controller data, mechanical component data, etc. to obtain specific fault labels, or use operation logs, fault reports, etc. to infer and label fault labels, and label each sample data with the corresponding fault label to build a labeled training data set.

[0062] Step 2: Train a generative adversarial network for each known field data to obtain a pre-trained discriminator corresponding to the known field. For example, two pre-trained discriminators are generated for the cut process data of two different brands of tobacco.

[0063] Step 2-1: consists of two parts: generator and discriminator, adopts deep neural network structure, including convolutional neural network, recurrent neural network, multilayer perceptron, etc. The two parts compete and cooperate with each other to achieve the goal of generating high-quality, realistic data.

[0064] Step 2-2: The input of the generator network G is random noise z, and the output is the generated data sample G(z); the learning goal of the generator network is to make the generated data sample G(z) as close to the real data distribution x as possible, and have similar characteristics and distribution characteristics as the known field data;

[0065] Step 2-3: The input of the discriminator network DIS is the generated data sample G(z) and the real data sample x of the known field, and the output is the discrimination result of the input data, indicating whether it is real data; the learning goal of the discriminator network is to accurately distinguish the difference between the generated data sample and the real data sample, and provide feedback about the data to help the generator network continuously improve the quality of the generated data.

[0066] Step 3: Use all known field corresponding pre-trained discriminators to build a multi-discriminator generative adversarial network. As shown in Figure 2 The structure diagram of the multi-discriminator generative adversarial network of the present application is shown. It includes:

[0067] Step 3-1: The data of different known fields uses the generative adversarial network to obtain the corresponding pre-trained discriminators;

[0068] Step 3-2: The task of the pre-trained discriminator DIS1 corresponding to the different known fields is to judge whether the generated data belongs to the known field;

[0069] Step 3-3: The pre-trained discriminators DIS1, DIS2 and an untrained discriminator DIS, and an untrained generator G construct a multi-discriminator generative adversarial network;

[0070] Step 3-4: The task of the discriminator DIS is to judge whether the generated data is real data;

[0071] Step 3-5: The goal of the generator G is to generate data that satisfies the following conditions: the pre-trained discriminators DIS1, DIS2 and the generated sample do not belong to the corresponding known field; the discriminator DIS judges that the generated sample is real data;

[0072] Step 3-6: The discriminator DIS wants to maximize the probability of dividing real samples into true and dividing false samples into false, and the generator G wants to minimize the probability of being identified as false by the discriminator DIS and being identified as belonging to the known field by the pre-trained discriminator. The optimization goal is

[0073]

[0074] where x is a real sample, z is random noise, denotes the expectation under the real data distribution, denotes the expectation under the random distribution, and log is the logarithmic function.

[0075] Step 4: Use the multi-discriminator generative adversarial network and all known field data to generate data with generalization characteristics to expand the training set and enhance the generalization ability of the model. Specifically, the following steps are included.

[0076] Step 4-1: For the discriminator DIS, input the real data x and the generated data G(z) into the discriminator respectively, calculate the loss function and update the discriminator parameters, and the loss function is

[0077]

[0078] where, denotes the loss function for discriminator update, x is a real sample, z is random noise, denotes the expectation under the real data distribution, denotes the expectation under the random distribution, and log is the logarithmic function.

[0079] Step 4-2: For the trained discriminators DIS1, DIS2, do not participate in the training of the multi-discriminator generative adversarial network.

[0080] Step 4-3: For the generator G, input random noise z, input the generated data G(z) into the discriminator DIS and the pre-trained discriminators DIS1, DIS2, calculate the loss function and update the generator parameters, the loss function is

[0081]

[0082] wherein, denotes the loss function for generator update, z is random noise, denotes the expectation under the random distribution, log is the logarithmic function.

[0083] Step 4-4: Repeat the above steps until the loss function of the generator is small enough or converges, or reaches a predetermined number of iterations.

[0084] Step 4-5: After training, the generator G will be able to accept any noise vector z as input and generate unlabelled data x with generalization characteristics g .

[0085] Step 5: Build a fault diagnosis model and use the generated generalization data and known field data to train the model to have the ability to generalize to unknown field industrial processes. Specifically, it includes the following steps:

[0086] Step 5-1: Design and build a fault diagnosis model according to the collected data x and fault state y, the input is the collected data x, and the output is the fault state y of the industrial process.

[0087] Step 5-2: Use the generated unlabelled data x g to pre-train the fault diagnosis model.

[0088] Step 5-3: Use the data x of the known field to supervise the training of the fault diagnosis model.

[0089] Step 6: Use the trained fault diagnosis model to perform online diagnosis on the new field industrial process, and real-time detect and identify possible faults. Taking the tobacco primary processing process as an example, it specifically includes the following steps:

[0090] Step 6-1: Collect real-time sensor data, controller data, mechanical component data and other key parameters from the new field industrial process, and use them as input for the operation of the fault diagnosis model. In the tobacco primary processing process, real-time data such as temperature, humidity, speed of the current cigarette brand can be collected for fault diagnosis.

[0091] Step 6-2: Preprocess the collected real-time data, including data cleaning, denoising, feature extraction, and other processing to obtain data that meets the input requirements of the fault diagnosis model. Real-time cleaning of newly collected process data for new brands can be performed to remove outliers and noise; denoising operations can be performed to eliminate interference factors in the data; feature extraction can be performed to extract features related to faults.

[0092] Step 6-3: Feature matching and conversion of preprocessed real-time data with known domain data and generated generalized data to ensure consistency and comparability of input data. Feature matching and conversion of preprocessed real-time data of new brands with known brand data. This ensures that different tobacco brand data can be input into the same fault diagnosis model.

[0093] Step 6-4: Use the trained fault diagnosis model to diagnose and analyze the real-time data after feature matching and conversion to determine whether the current industrial process has a fault.

[0094] Step 6-5: Based on the output of the fault diagnosis model, make fault judgments and classifications for the industrial process, and trigger the alarm mechanism if necessary to remind relevant personnel to take appropriate measures;

[0095] Step 6-6: Continuously monitor the fault diagnosis of new domain industrial processes, collect new data samples and update the fault diagnosis model to continuously improve diagnostic accuracy.

[0096] Taking the tobacco primary processing process as an example, data collection or real-time monitoring is performed on variables such as water addition set value, water addition controller set value, water addition controller internal control set value, outlet temperature set value, hot air temperature set value, outlet moisture set value, actual water addition value, outlet temperature actual value, and outlet moisture actual value. Through data collection of known brands, a pre-trained discriminator corresponding to the brand is generated, a multi-discriminator generative adversarial network is constructed, and data with general properties of the tobacco primary processing process of all known or unknown brands is generated. The fault diagnosis model is trained, and online diagnosis of the tobacco primary processing process of new brands is performed.

[0097] To sum up, the present application proposes an industrial process fault diagnosis method based on domain generalization. Through the combination of efficient data collection, data labeling and generative adversarial network, online diagnosis and fault identification of new domain industrial process are realized. The present application effectively solves the challenges and limitations of traditional industrial process fault diagnosis methods in new domain applications. By training with multiple similar known domain data, the present application can realize fault diagnosis of unknown industrial process, avoiding the problem of requiring a large number of labeled sample data in traditional methods. The present application uses the mechanism of generative adversarial network to generate synthetic data with domain generalization ability, thereby increasing the diversity and coverage of training data and improving the generalization ability and adaptability of the model. The present application can improve the reliability and stability of industrial process in industrial process fault diagnosis, reduce production loss caused by fault, improve production efficiency and product quality, and also provides a general fault diagnosis solution for industrial processes in different fields, which has wide application prospect and economic benefit.

[0098] The above description describes the embodiments that will help those skilled in the art to further understand the present application, but does not limit the present application in any form. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present application. These all belong to the protection scope of the present application.

Claims

1. An industrial fault diagnosis method based on domain generalization, characterized in that: By training a generative adversarial network using data from similar domains, we generate universal data with generalization characteristics. This data is then combined with known domain data to build and train a fault diagnosis model. This model is used for fault diagnosis in industrial scenarios where there is a lack of prior knowledge and data in unknown domains. The process includes the following steps: Step 1: Collect sample data from multiple similar fields of the unknown industrial process to be diagnosed and label them with fault labels; Step 2: Use the generative adversarial network to build a single discriminator, input sample data from each known domain for training, and obtain a pre-trained discriminator corresponding to the known domain; Step 3: Use the pre-trained discriminators corresponding to all known domains to build a multi-discriminator generative adversarial network; Step 4: Use a multi-discriminator generative adversarial network to generate data with generalization properties using all known domain data as input to expand the training set and enhance the generalization ability of the model; Step 5: Build a fault diagnosis model and train it using the generated generalized data and known domain data to enable it to generalize to unknown domain industrial processes. Step 6: Use the trained fault diagnosis model to perform online diagnosis of the new domain industrial process to detect and identify possible faults in real time.

2. The industrial fault diagnosis method based on domain generalization according to claim 1 is characterized in that: The data collection method comprises the following steps: Determine the operating data of the unknown industrial process and the adjacent known areas that need to be collected; the operating data includes: sensor or monitoring equipment data, controller data, and mechanical component data; Configure the device location, sampling frequency, and data transmission method for data collection; Real-time data collection and monitoring of the operating status of industrial processes to obtain sample data at multiple time points; Preprocess the collected sample data; Using data labeling methods, each sample data is labeled with a corresponding fault label to construct a labeled training dataset.

3. The industrial fault diagnosis method based on domain generalization according to claim 2 is characterized in that: The data labeling method includes: analyzing and judging sensor data, controller data, and mechanical component data to obtain specific fault labels; and using operation logs and fault report information to infer and label the fault labels.

4. The industrial fault diagnosis method based on domain generalization according to claim 1 is characterized in that: The multi-discriminator generative adversarial network includes several pre-trained discriminators DIS i and an untrained discriminator DIS, and an untrained generator G; Pre-trained discriminators DIS corresponding to different known domains i Used to determine whether the generated data belongs to the known domain; the discriminator DIS is used to determine whether the generated data is real data; The generator G is used to generate data that meets the following conditions: the pre-trained discriminators DIS1 and DIS2 and the samples generated by the judgment do not belong to the corresponding known domain; The discriminator DIS determines whether the generated sample is real data; According to the discriminator DIS, we need to maximize the probability of classifying real samples as true and false samples as false. The generator G needs to minimize the sum of the probabilities of being identified as false by the discriminator DIS and being identified as belonging to the known domain by the pre-trained discriminator. The optimization goal is defined as Among them, x is the real sample, z is the random noise, refers to the expectation of the real data distribution, It refers to the expectation under random distribution, and log is the logarithmic function.

5. The industrial fault diagnosis method based on domain generalization according to claim 4 is characterized in that: The pre-trained discriminator DIS i It is the discriminator after the convergence of a common generative adversarial network. Common generative adversarial networks include a single generator and a single discriminator, both of which use deep neural network structures, including recurrent neural networks and multi-layer perceptrons. The input of the generator network G is random noise z, and the output is the generated data sample G(z); the learning goal of the generator network is to make the generated data sample G(z) as close as possible to the real data distribution x, and have similar characteristics and distribution characteristics to the known domain data; The input of the discriminator network DIS is the generated data sample G(z) and the real data sample x in the known field, and the output is the discrimination result of the input data, indicating whether it is real data; The learning goal of the discriminator network is to accurately distinguish the differences between generated data samples and real data samples, and provide feedback about the data to help the generator network continuously improve and generate universal data.

6. The industrial fault diagnosis method based on domain generalization according to claim 4 is characterized in that: The following loss function is defined according to the optimization objective: For the discriminator DIS, the real data x and the generated data G(z) are input into the discriminator respectively, and the loss function is calculated And update the discriminator parameters: in, Refers to the loss function used to update the discriminator, x is the real sample, z is random noise, refers to the expectation of the real data distribution, Refers to the expectation under random distribution, log is the logarithmic function; For several training discriminators DIS i , do not participate in the training of multi-discriminator generative adversarial networks; For the generator G, input random noise z, and input the generated data G(z) into the total discriminator DIS and each pre-trained discriminator DIS i , calculate the joint loss function and update the generator parameters: in, Refers to the loss function used for generator update, z is random noise, Refers to the expectation under random distribution, log is the logarithmic function; Repeat the above steps until the joint loss function of the generator meets the threshold or converges, or reaches the predetermined number of iterations training cutoff; after training is completed, the generator G accepts any noise vector z as input and generates unlabeled data x with generalization characteristics g .

7. The industrial fault diagnosis method based on domain generalization according to claim 1 is characterized in that: The fault diagnosis model adopts an encoder-decoder network structure, with a stack of a recurrent neural network and a convolutional neural network as the encoder and a fully connected network as the decoder.

8. The industrial fault diagnosis method based on domain generalization according to claim 7 is characterized in that: Training the fault diagnosis model involves the following steps: The fault diagnosis model is defined as the input of the collected data x and the output is the fault state y of the industrial process; Utilize the generated unlabeled data x with generalization properties g , pre-train the fault diagnosis model; The fault diagnosis model is trained in a supervised manner using data x from known domains.

9. The industrial fault diagnosis method based on domain generalization according to claim 1, characterized in that: The method for online diagnosis of a novel industrial process comprises the following steps: Collect real-time key parameter data from new industrial processes, pre-process them, and use them as input for fault diagnosis models; Step 6-3: Perform feature matching and conversion on the preprocessed real-time data, known domain data, and generated generalized data to ensure consistency and comparability of the input data. Step 6-4: Use the trained fault diagnosis model to diagnose and analyze the real-time data that has undergone feature matching and conversion to determine whether there is a fault in the current industrial process; According to the output results of the fault diagnosis model, fault diagnosis and classification of industrial processes are carried out; Continuously monitor fault diagnosis of industrial processes in new areas, collect new data samples and update fault diagnosis models to continuously improve diagnostic accuracy.

Images