Fault diagnosis system of organic waste gas catalytic decomposition device

Abstract

The invention relates to the technical field of intelligent fault diagnosis, in particular to a fault diagnosis system of an organic waste gas catalytic decomposition device. The system comprises a data acquisition module, a data preprocessing module, a model training module, a fault diagnosis module and a report generation module. Data of a device sensor are received in real time, data cleaning, standardization and normalization processing are carried out, a fault diagnosis model based on deep learning is trained through historical data, and accurate diagnosis of faults is achieved. The system can automatically identify the fault type and position, generate a detailed fault diagnosis report, provide a remote monitoring function, allow a remote user to check the equipment state and the fault diagnosis result in real time, adjust the equipment parameters in time, and ensure stable operation of the equipment. According to the invention, the fault diagnosis efficiency and accuracy are improved, the downtime and maintenance cost are reduced, and powerful guarantee is provided for industrial production.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent fault diagnosis, and particularly relates to a fault diagnosis system for an organic waste gas catalytic decomposition device. Background Art

[0002] Organic waste gas catalytic decomposition devices play an important role in industrial production and are used to treat harmful waste gases generated during the production process. However, these devices may malfunction for various reasons during operation, such as sensor failure, catalyst poisoning, equipment aging, etc., resulting in low waste gas treatment efficiency or even equipment shutdown. Traditional fault diagnosis methods rely on manual inspections and empirical judgments, suffering from long diagnosis cycles and low accuracy. With the development of industrial Internet of Things and artificial intelligence technologies, it is particularly important to develop a system that can diagnose faults in organic waste gas catalytic decomposition devices in real time and accurately. Summary of the Invention

[0003] Aiming at the above-mentioned existing technical deficiencies, the purpose of the present invention is to provide a fault diagnosis system for an organic waste gas catalytic decomposition device, which solves the problems of long diagnosis cycle and low accuracy in the prior art.

[0004] To solve the above technical problems, the present invention adopts the following technical solutions: In a first aspect, a fault diagnosis system for an organic waste gas catalytic decomposition device includes: A data acquisition module for receiving device data from sensors deployed on the organic waste gas catalytic decomposition device in real time; A data preprocessing module for preprocessing the device data collected by the data acquisition module; A model training module for training a deep learning-based fault diagnosis model using historical device data; A fault diagnosis module for inputting the preprocessed real-time device data into the deep learning-based fault diagnosis model to determine whether the device has a fault and locate the fault; A report generation module for generating a fault diagnosis report according to the results output by the fault diagnosis module.

[0005] Preferably, in a possible implementation manner of the first aspect, the device data includes flammable organic waste gas concentration data, waste gas thermal conductivity data, toxic and harmful gas concentration data, hydrocarbon gas concentration data, volatile organic compound gas concentration data, internal temperature data of the organic waste gas catalytic decomposition device, internal pressure data of the organic waste gas catalytic decomposition device, and waste gas flow data.

[0006] Preferably, in a possible implementation manner of the first aspect, the data preprocessing includes data cleaning, data standardization, and data normalization processing; the data cleaning is used to remove invalid and abnormal data; the data standardization converts the data to the same scale range; the data normalization maps the data to a preset interval.

[0007] Preferably, in a possible implementation manner of the first aspect, the historical device data includes historical flammable organic waste gas concentration data, historical waste gas thermal conductivity data, historical toxic and harmful gas concentration data, historical hydrocarbon gas concentration data, historical volatile organic compound gas concentration data, historical internal temperature data of the organic waste gas catalytic decomposition device, historical internal pressure data of the organic waste gas catalytic decomposition device, historical waste gas flow data, and historical fault data.

[0008] Preferably, in a possible implementation manner of the first aspect, the fault diagnosis model based on deep learning adopts a graph convolutional network model, and the structure of the graph convolutional network model includes: An input layer for receiving the preprocessed real-time device data; A graph construction layer for constructing a graph structure according to the characteristics of the device data, where nodes represent different device data and edges represent the correlation between device data; A graph convolutional layer for extracting feature information in the graph structure through graph convolutional operations; An output layer for outputting the probability distribution of the fault type and the fault location.

[0009] Preferably, in a possible implementation manner of the first aspect, the judgment process of the fault diagnosis module specifically includes: inputting the preprocessed real-time device data into the trained fault diagnosis model, the model outputs the probability distribution of the fault type and the fault location, and judges whether there is a fault according to a preset threshold. If the fault probability at a certain position is greater than the preset threshold, it is considered that there is a fault, and the fault type and position with the highest probability are selected as the diagnosis result.

[0010] Preferably, in a possible implementation manner of the first aspect, the fault diagnosis report includes the time of fault occurrence, fault type, fault location, fault cause, solution, and maintenance measures.

[0011] Preferably, in a possible implementation manner of the first aspect, the system further includes a remote monitoring module for receiving and displaying in real time the fault diagnosis result output by the fault diagnosis module and the fault diagnosis report generated by the fault diagnosis report generation module, and at the same time allowing remote users to monitor the operating status of the organic waste gas catalytic decomposition device through the user interface, receive and respond to control instructions from remote users to adjust the operating parameters of the device.

[0012] The beneficial effects of the present invention are as follows: By using deep learning algorithms to train a fault diagnosis model, accurate diagnosis of device faults is achieved. The system can automatically identify the fault type and determine the fault location, generate a detailed fault diagnosis report, and provide accurate fault information and solutions for maintenance personnel. In addition, the system also has a remote monitoring function, allowing remote users to view the device operation status and fault diagnosis results in real time, respond in a timely manner and adjust device parameters to ensure the stable operation of the device. The fault diagnosis system of the present invention not only improves the efficiency and accuracy of fault diagnosis, but also reduces the device downtime and maintenance costs, providing a strong guarantee for the continuity and stability of industrial production. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0014] Figure 1 FIG. provides a schematic structural diagram of a fault diagnosis system for an organic waste gas catalytic decomposition device of the present application.

[0015] Explanation of reference numerals: 1 - data acquisition module, 2 - data preprocessing module, 3 - model training module, 4 - fault diagnosis module, 5 - report generation module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0017] Embodiment 1: As Figure 1 shown, the present invention provides a fault diagnosis system for an organic waste gas catalytic decomposition device, including: A data acquisition module 1 for collecting real-time device data, and the device data comes from sensors deployed on the organic waste gas catalytic decomposition device.

[0018] Specifically, the device data includes flammable organic waste gas concentration data, waste gas thermal conductivity data, toxic and harmful gas concentration data, hydrocarbon gas concentration data, volatile organic compound gas concentration data, internal temperature data of the organic waste gas catalytic decomposition device, internal pressure data of the organic waste gas catalytic decomposition device, and waste gas flow data.

[0019] The sensors deployed on the organic waste gas catalytic decomposition device include catalytic combustion type gas sensors, thermal conductivity sensors, electrochemical gas sensors, infrared absorption type gas sensors, optical fiber gas sensors, temperature sensors, pressure sensors, and flow sensors.

[0020] The catalytic combustion type gas sensor detects based on the heat generated by the combustion of the target gas on the catalyst surface, which causes a temperature change in the sensor element. It is mainly used to detect the concentration changes of combustible organic waste gases, such as methane, hydrogen, etc. By monitoring the fluctuations in the waste gas concentration, it can be judged whether the catalytic decomposition device effectively treats the waste gas.

[0021] The thermal conductivity sensor uses the difference in thermal conductivity between combustible gas and air to measure the gas concentration. By measuring the difference in thermal conductivity between the waste gas and air, the change in the composition of the waste gas can be indirectly reflected. When the combustible components in the waste gas decrease, the thermal conductivity will change, thereby judging the catalytic decomposition effect.

[0022] The electrochemical gas sensor detects the gas concentration by measuring the current change generated by the oxidation-reduction reaction of the gas in the electrolyte. It is suitable for detecting the concentration of toxic and harmful gases such as oxygen and carbon monoxide. In the catalytic decomposition device, the residual amount of these gases in the waste gas can be monitored to judge whether the device completely decomposes the organic waste gas.

[0023] The infrared absorption type gas sensor detects by using the selective absorption characteristics of specific gases for infrared light. In the catalytic decomposition device, the concentration of gases such as carbon dioxide in the waste gas is monitored to evaluate the catalytic efficiency of the device.

[0024] The optical fiber gas sensor realizes gas detection by using principles such as optical fiber interference and fluorescence quenching. It has advantages such as anti-electromagnetic interference and remote monitoring, and is suitable for gas detection in special environments. In the catalytic decomposition device, the concentration of certain specific gases in the waste gas is monitored, such as volatile organic compounds.

[0025] The temperature sensor measures the temperature inside the catalytic decomposition device. By monitoring the temperature change inside the device, it can be judged whether the catalyst is in the optimal working temperature range and whether there is overheating or insufficient temperature.

[0026] The pressure sensor measures the pressure inside the catalytic decomposition device. By monitoring the pressure change inside the device, it can be judged whether there is abnormal pressure during the waste gas treatment process, such as too high or too low pressure, which may be related to device blockage, leakage, or catalyst failure.

[0027] The flow sensor measures the waste gas flow rate entering and leaving the catalytic decomposition device. By monitoring the change in the waste gas flow rate, it can be judged whether there are problems such as blockage or leakage in the device, and whether the catalytic decomposition efficiency is affected.

[0028] The data preprocessing module 2 preprocesses the device data collected by the data acquisition module 1.

[0029] Specifically, data preprocessing includes data cleaning, data standardization, and data normalization; data cleaning is used to remove invalid and abnormal data; data standardization converts the data to the same scale range; data normalization maps the data to a preset interval.

[0030] In this embodiment, first, data cleaning is performed. Data is collected from the sensors of the organic waste gas catalytic decomposition device and stored in a Pandas DataFrame. The DataFrame contains the following fields: concentration of combustible organic waste gas (unit: ppm), thermal conductivity of waste gas (unit: W / (m·K)), concentration of toxic and harmful gases (unit: ppm), concentration of hydrocarbon gases (unit: ppm), concentration of volatile organic compound gases (unit: ppm), internal temperature of the organic waste gas catalytic decomposition device (unit: °C), internal pressure of the organic waste gas catalytic decomposition device (unit: kPa), and waste gas flow rate (unit: m³ / h).

[0031] Next, the data cleaning step is carried out to remove invalid and abnormal data. Invalid data mainly includes null values (NaN) and infinite values (inf). For null values and infinite values, first use the replace() method to replace infinite values with reasonable values (such as NaN or specific filling values), and then use the dropna() method to remove null values.

[0032] Abnormal data refers to values that exceed the reasonable range, which may be caused by sensor failures, data transmission errors, or environmental mutations, etc. First, calculate the quartiles (Q1 and Q3) and the interquartile range (IQR = Q3 - Q1) of each field in the DataFrame. Then, set the threshold to 1.5 times the IQR as the boundary for filtering out abnormal values. For each field, calculate its lower bound (lower_bound = Q1 - threshold) and upper bound (upper_bound = Q3 + threshold), and consider the values outside this range as abnormal values for filtering.

[0033] After completing data cleaning, the data standardization step is carried out next. Data standardization is to convert the data to the same scale range for subsequent model training. In this embodiment, the Z-score standardization method is adopted to convert the data into a distribution with a mean of 0 and a standard deviation of 1.

[0034] To achieve data standardization, the StandardScaler class in the scikit-learn library is used. First, a StandardScaler object is initialized. Then, the fields to be standardized are extracted, namely all the sensor data fields mentioned above. Next, the fit_transform() method is used to perform standardization processing on these fields.

[0035] Finally, the data normalization step is carried out. In this embodiment, the Min-Max normalization method is adopted to map the data to the interval [0, 1].

[0036] To achieve data normalization, the MinMaxScaler class in the scikit-learn library is used. First, a MinMaxScaler object is initialized. Then, the same fields as those for data standardization are used for normalization. Similarly, the fit_transform() method is used to perform normalization processing on these fields.

[0037] The model training module 3 uses historical device data to train a deep learning-based fault diagnosis model.

[0038] Specifically, historical flammable organic waste gas concentration data, historical waste gas thermal conductivity data, historical toxic and harmful gas concentration data, historical hydrocarbon gas concentration data, historical volatile organic compound gas concentration data, historical internal temperature data of the organic waste gas catalytic decomposition device, historical internal pressure data of the organic waste gas catalytic decomposition device, historical waste gas flow data, and historical fault data.

[0039] The deep learning-based fault diagnosis model adopts a graph convolutional network model. The structure of the graph convolutional network model includes: An input layer for receiving preprocessed real-time device data. The preprocessed data includes parameters such as combustible gas concentration, thermal conductivity, toxic gas concentration, hydrocarbon concentration, volatile organic compound concentration, temperature and pressure, flow rate, and fault data.

[0040] A graph construction layer that constructs a graph structure according to the characteristics of the preprocessed device data. The nodes represent different device data, and the edges represent the correlation between the data. The correlation is calculated by methods such as correlation coefficient and mutual information. In this embodiment, the correlation coefficient matrix is used as the basis for the edge weights, and a threshold of 0.5 is set to determine the existence condition of the edges. When the absolute value of the correlation coefficient between two nodes is greater than the threshold, it is considered that there is an edge between them, and the correlation coefficient is used as the edge weight.

[0041] The graph convolutional layer extracts feature information in the graph structure through graph convolutional operations. In this embodiment, the GCNConv class in the PyTorchGeometric library is used to implement the graph convolutional operations. First, the graph data is converted into the PyTorch Geometric format, including the node feature matrix and the edge index matrix. Then, a GCN model is constructed, including an input layer and multiple hidden layers. Each hidden layer uses the GCNConv class for graph convolutional operations, and applies the ReLU activation function and Dropout operation to enhance the generalization ability of the model.

[0042] Specifically, the graph data is converted into the PyTorch Geometric format, including the node feature matrix (features) and the edge index matrix (edge_index). A GCN model is constructed, including an input layer and multiple hidden layers. The dimension of the input layer is equal to the number of columns of the node feature matrix, and the dimension of the output layer is equal to the number of fault types. In each hidden layer, the GCNConv class is used for graph convolutional operations, taking the node feature matrix and the edge index matrix as inputs to obtain a new node feature matrix. Then, the ReLU activation function and Dropout operation are applied to enhance the generalization ability of the model.

[0043] The output layer outputs the probability distribution of the fault type and the fault location. In this embodiment, the GCNConv class is used for the last graph convolutional operation, and the log_softmax function is applied to output the probability distribution.

[0044] During the training process, the cross-entropy loss function is used to measure the difference between the model prediction result and the actual fault data, and the optimizer is used to update the weight parameters of the model. Through multiple iterative trainings, a graph convolutional network model that can accurately predict the fault type and the fault location is obtained.

[0045] The fault diagnosis module 4 inputs the preprocessed real-time device data into the fault diagnosis model based on deep learning to determine whether the device has a fault and determine the fault location.

[0046] Specifically, the preprocessed real-time device data is input into the trained fault diagnosis model. The model outputs the probability distribution of the fault type and the fault location. Whether there is a fault is judged according to a preset threshold. If there is a fault, the fault type and location with the highest probability are selected as the diagnosis result.

[0047] The fault diagnosis module 4 first receives the preprocessed real-time device data. These data have undergone preprocessing steps such as data cleaning, format conversion, and normalization to ensure the accuracy and consistency of the data.

[0048] Subsequently, the preprocessed real-time device data is input into the fault diagnosis model based on the graph convolutional network output by the model training module 3. The model analyzes and processes the data using its internal neural network structure. Finally, a probability distribution regarding the fault type and fault location is output. This probability distribution reflects the judgment confidence of the model for the fault type and location corresponding to the current data, where each fault type and location corresponds to a probability value.

[0049] To determine whether the device has a fault and locate the specific fault position, in this embodiment, a threshold of 0.8 is set. In the probability distribution output by the model, if the probability value of a certain fault type and location exceeds the threshold, it is considered that the device has that type of fault, and the fault type and location with the highest probability value are selected as the final diagnosis result.

[0050] The report generation module 5 is used to generate a fault diagnosis report based on the result output by the fault diagnosis module 4.

[0051] Specifically, in the fault diagnosis report, the time of the fault occurrence, the fault type, and the fault location are first recorded in detail. In terms of fault cause analysis, the report combines the fault type and location and gives multiple possible causes based on historical data and expert experience to provide comprehensive reference for the operators. For these causes, the report further provides specific solutions, such as adjusting device parameters, replacing damaged components, cleaning blockages, etc., aiming to quickly restore the normal operation of the device.

[0052] In addition, to prevent similar faults from occurring again, the report also proposes targeted maintenance measures, including regularly checking the device status, replacing aging components, optimizing the operation process, etc., to improve the reliability and stability of the device.

[0053] In the report generation process, the report generation module 5 first receives the result output by the fault diagnosis module 4, then integrates it with information such as device historical data and expert experience, and uses the report template to generate a complete fault diagnosis report. Finally, the report is output in the form of an electronic document for the operators to consult and process.

[0054] In one embodiment, the system further includes a remote monitoring module, which is used to receive and display in real time the fault diagnosis result output by the fault diagnosis module 4 and the fault diagnosis report generated by the fault diagnosis report generation module 5. At the same time, it allows remote users to monitor the operating status of the organic waste gas catalytic decomposition device through the user interface, receive and respond to control instructions from remote users to adjust the operating parameters of the device or trigger the fault diagnosis process.

[0055] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. A fault diagnosis system for an organic waste gas catalytic decomposition device, characterized in that, Including: A data acquisition module, which is used to receive device data from sensors deployed on the organic waste gas catalytic decomposition device in real time; A data preprocessing module, which preprocesses the device data collected by the data acquisition module; A model training module, which trains a deep learning-based fault diagnosis model using historical device data; A fault diagnosis module, which inputs the preprocessed real-time device data into the deep learning-based fault diagnosis model to determine whether the device has a fault and locate the fault; A report generation module, which is used to generate a fault diagnosis report according to the result output by the fault diagnosis module.

2. The fault diagnosis system according to claim 1, characterized in that The device data includes flammable organic waste gas concentration data, waste gas thermal conductivity data, toxic and harmful gas concentration data, hydrocarbon gas concentration data, volatile organic compound gas concentration data, internal temperature data of the organic waste gas catalytic decomposition device, internal pressure data of the organic waste gas catalytic decomposition device, and waste gas flow data.

3. The fault diagnosis system according to claim 1, characterized in that, The data preprocessing includes data cleaning, data standardization, and data normalization processing; the data cleaning is used to remove invalid and abnormal data; the data standardization converts the data to the same scale range; the data normalization maps the data to a preset interval.

4. The fault diagnosis system according to claim 1, wherein The historical device data includes historical flammable organic waste gas concentration data, historical waste gas thermal conductivity data, historical toxic and harmful gas concentration data, historical hydrocarbon gas concentration data, historical volatile organic compound gas concentration data, historical internal temperature data of the organic waste gas catalytic decomposition device, historical internal pressure data of the organic waste gas catalytic decomposition device, historical waste gas flow data, and historical fault data.

5. The fault diagnosis system according to claim 1, characterized in that, The deep learning-based fault diagnosis model uses a graph convolutional network model, and the structure of the graph convolutional network model includes: An input layer, which is used to receive the preprocessed real-time device data; A graph construction layer, which constructs a graph structure according to the characteristics of the device data, where nodes represent different device data and edges represent the correlation between device data; A graph convolutional layer, which extracts feature information in the graph structure through graph convolutional operations; An output layer, which outputs the probability distribution of the fault type and the fault location.

6. The fault diagnosis system according to claim 1, wherein, The judgment process of the fault diagnosis module specifically includes: inputting the preprocessed real-time device data into the trained fault diagnosis model, the model outputs the probability distribution of the fault type and the fault location, and determines whether there is a fault according to a preset threshold. If the fault probability at a certain position is greater than the preset threshold, it is considered that there is a fault, and the fault type and location with the highest probability are selected as the diagnosis result.

7. The fault diagnosis system according to claim 1, characterized in that, The fault diagnosis report includes the time of the fault occurrence, the fault type, the fault location, the fault cause, the solution, and the maintenance measures.

8. The fault diagnosis system according to claim 1, wherein, The system further includes a remote monitoring module, which is used to receive and display the fault diagnosis result output by the fault diagnosis module and the fault diagnosis report generated by the fault diagnosis report generation module in real time. At the same time, it allows remote users to monitor the operating status of the organic waste gas catalytic decomposition device through the user interface, receive and respond to control instructions from remote users to adjust the operating parameters of the device.

Images