Intelligent monitoring and analyzing method and system for faults of heat exchanger

Abstract

The invention discloses an intelligent monitoring and analyzing method and system for faults of a heat exchanger, and belongs to the field of intelligent monitoring, and the method comprises the following steps: obtaining real-time data of the heat exchanger in a working state, and preprocessing the real-time data to obtain preprocessed data; constructing a fault diagnosis model and a fault prediction model based on a LightGBM framework, and inputting the preprocessed data into the fault diagnosis model to obtain the working state and confidence of the heat exchanger; constructing a time series data set based on the working state of the heat exchanger and the preprocessed data; adding features into the time sequence data set to obtain key fault features; and inputting the confidence coefficient and the key fault features into the fault prediction model to obtain a periodic fault occurrence probability. The running state of the heat exchanger can be monitored in real time, faults can be accurately diagnosed, and reliable fault prediction can be provided.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent monitoring technology, and in particular relates to a heat exchanger fault intelligent monitoring and analysis method and system. Background Art

[0002] When the heat exchanger is working, the cold and hot fluids use the heat transfer elements of the heat exchanger to exchange heat to achieve the effect of heating or cooling to meet the requirements of chemical production. Most of the materials passing through the heat exchanger are fluids and are often mixed with suspended matter, which can easily cause the heat exchanger to scale in long-term use, thereby increasing the thermal resistance of the heat exchanger and greatly reducing the heat exchange capacity of the heat exchanger. In order to ensure the normal operation of the heat exchanger and improve the heat exchange efficiency, it is necessary to monitor various parameters of the heat exchanger in real time as required.

[0003] Existing heat exchanger fault monitoring technologies mainly rely on traditional sensors and regular manual inspections, which can provide certain fault diagnosis capabilities. As technology develops, some advanced monitoring systems begin to integrate machine learning algorithms to improve the accuracy and real-time performance of fault detection. By analyzing sensor data, these systems can identify abnormal patterns in the operation of the heat exchanger, thereby providing early warning of potential failures.

[0004] Although existing heat exchanger fault monitoring technologies have made some progress, they still have some significant defects. First, traditional monitoring methods rely on manual inspection, which is not only inefficient but also easily interfered by human factors, resulting in the accuracy and timeliness of fault detection being affected. Second, existing machine learning models often require a large amount of labeled data for training, which is difficult to obtain in actual industrial environments. In addition, many models have insufficient generalization capabilities when processing complex industrial data and are difficult to adapt to changing working conditions. These problems limit the effectiveness and reliability of existing technologies in practical applications.

[0005] In view of the limitations of existing technologies, it is urgent to develop a new intelligent monitoring and analysis method for heat exchanger faults that can overcome the shortcomings of existing technologies. The new technology should be able to achieve real-time monitoring of the operating status of the heat exchanger, accurately diagnose faults, and provide reliable fault prediction. Summary of the invention

[0006] In order to solve the above technical problems, the present invention proposes a heat exchanger fault intelligent monitoring and analysis method and system to solve the above problems existing in the prior art.

[0007] To achieve the above object, the present invention provides a heat exchanger fault intelligent monitoring and analysis method, comprising the following steps:

[0008] Acquiring real-time data of the heat exchanger in a working state and preprocessing the real-time data to obtain preprocessed data;

[0009] Building a fault diagnosis model and a fault prediction model based on the LightGBM framework, and inputting the preprocessed data into the fault diagnosis model to obtain the working status and confidence of the heat exchanger;

[0010] constructing a time series data set based on the working state of the heat exchanger and the preprocessed data;

[0011] Adding features to the time series data set to obtain key fault features;

[0012] The confidence level and the key fault characteristics are input into the fault prediction model to obtain the probability of periodic fault occurrence.

[0013] Optionally, a fault diagnosis model is constructed based on the LightGBM architecture with the goal of minimizing the multi-classification logarithmic loss;

[0014] Among them, the objective function of the fault diagnosis model is:

[0015]

[0016] In the formula, n is the number of samples, M is the number of categories, and y i,c is an indicator variable, which is 1 when sample i belongs to category c, and 0 otherwise. is the probability that the fault diagnosis model predicts that sample i belongs to c.

[0017] Optionally, the process of obtaining the working status and confidence includes:

[0018] Divide the preprocessed dataset into training set, validation set and test set;

[0019] The fault diagnosis model is trained using a training set, and the parameters of the fault diagnosis model are adjusted with the goal of minimizing the multi-classification logarithmic loss;

[0020] The trained fault diagnosis model is obtained by using the validation set to validate the model after parameter adjustment;

[0021] Using the trained fault diagnosis model to predict the test set, obtain the working state classification result of the heat exchanger and the prediction probability of each classification;

[0022] Based on the predicted probabilities, the confidence is calculated using the difference between the maximum probability and the second largest probability.

[0023] Optionally, the process of adding features to the time series data set to obtain key fault features includes:

[0024] Acquire candidate features based on the preprocessed data;

[0025] Calculate the ratio between each pair of candidate features based on the candidate features to obtain a ratio feature;

[0026] The ratio features are screened based on linear discriminant analysis to obtain key fault features.

[0027] Optionally, the expression for calculating the ratio between each pair of candidate features based on the candidate features is:

[0028]

[0029] In the formula, represents the average ratio feature within the time window [t-Δt, t], Δt represents the length of the time window, V A,T represents the measured value of sensor A at time T, V B,T Represents the measurement value of sensor B at time T.

[0030] Optionally, the key fault features include: the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, the ratio of the operating temperature to the set temperature, the ratio of temperature to pressure, the ratio of the maximum value to the minimum value within the sliding window, and the ratio of maintenance and fault history.

[0031] Optionally, the fault prediction model is expressed as:

[0032] P = LightGBM(β0+β1R temp-flow +β2R work-set +β3R temp-push +β4R max-min +β5R repair-history +β6C)

[0033] Where P represents the fault prediction probability, β0, β1, β2, β3, β4, β5, β6 represent the fault prediction model parameters, and R temp-flow It represents the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, R work-set Indicates the ratio of operating temperature to set temperature, R temp-push Represents the ratio of temperature to pressure, R max-min Represents the ratio of the maximum value to the minimum value in the sliding window, R repair-history represents the ratio of protection and failure history, and C represents the confidence level.

[0034] The present invention also discloses a heat exchanger fault intelligent monitoring and analysis system, comprising: an integrated sensor, a wireless communication module, a fault diagnosis module, a fault prediction module and a user interaction interface;

[0035] Wherein, the integrated sensor is used to collect real-time data of the heat exchanger in the working state;

[0036] The wireless communication module is used to transmit the real-time data to the analysis system;

[0037] The fault diagnosis module is used to determine the working state of the heat exchanger and the corresponding confidence level based on the real-time data;

[0038] The fault prediction module is used to determine the probability of periodic fault occurrence based on the confidence of the working state of the heat exchanger and the key fault characteristics;

[0039] The user interaction interface is used to reflect the working status of the heat exchanger and the probability of failure in the future cycle to the user, and propose corresponding solution strategies.

[0040] The present invention also discloses a computer terminal device, comprising:

[0041] one or more processors;

[0042] A memory, coupled to the processor, for storing one or more programs;

[0043] When the one or more programs are executed by the one or more processors, the one or more processors implement a heat exchanger fault intelligent monitoring and analysis method.

[0044] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, a heat exchanger fault intelligent monitoring and analysis method is implemented.

[0045] Compared with the prior art, the present invention has the following advantages and technical effects:

[0046] The present invention ensures the quality and availability of data through the acquisition and preprocessing of real-time data, laying a solid foundation for subsequent model training and prediction. The fault diagnosis model constructed using the LightGBM framework can not only accurately identify the current working status of the heat exchanger, but also provide confidence assessment, making the fault diagnosis results more credible. Furthermore, combined with the construction of time series data sets and the extraction of key fault features, this technology can deeply analyze the operating trends and potential risks of the heat exchanger. By inputting confidence and key fault features into the fault prediction model, this technology can predict the probability of failure of the heat exchanger in the future, thereby providing a scientific basis for maintenance decisions. Through advanced data processing and machine learning algorithms, this technology realizes real-time monitoring, accurate diagnosis and effective prediction of heat exchanger faults, greatly improving the operating safety and maintenance efficiency of the heat exchanger, and providing strong technical support for the stable operation of industrial systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0048] Figure 1 This is a flow chart of a heat exchanger fault intelligent monitoring and analysis method according to an embodiment of the present invention;

[0049] Figure 2 This is a structural diagram of a heat exchanger fault intelligent monitoring and analysis system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0050] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0051] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0052] Embodiment 1

[0053] like Figure 1 As shown, this embodiment provides a heat exchanger fault intelligent monitoring and analysis method, including the following steps:

[0054] The real-time data of the heat exchanger in the working state is obtained and the real-time data is preprocessed to obtain preprocessed data. The specific implementation process includes:

[0055] As one of the core equipment in industrial systems, the main function of heat exchangers is to achieve effective heat transfer to meet the temperature control requirements in the process. The performance of heat exchangers is directly related to the energy efficiency and stability of the entire system. In the long-term operation process, problems such as erosion and surface deposition on the heat exchange wall will gradually affect its heat exchange efficiency, resulting in the outlet temperature failing to meet the predetermined standard.

[0056] Temperature signal is one of the important parameters for monitoring the performance of heat exchangers. The temperature change of the inlet and outlet working fluid can intuitively reflect the heat exchange situation inside the heat exchanger. When the heat exchanger works normally, the inlet and outlet temperatures of the fluid will show a certain temperature difference after passing through the heat exchanger, which is a direct reflection of the heat exchange efficiency. However, if wall erosion or surface deposition occurs inside the heat exchanger, these abnormal conditions will hinder the normal transfer of heat, resulting in a decrease in the temperature difference and the outlet temperature cannot reach the expected value.

[0057] Wall erosion is usually caused by the continuous impact and chemical action of solid particles or corrosive substances in the fluid on the inner wall of the heat exchanger. This erosion will gradually thin the heat exchange wall, increase the heat transfer resistance of the heat exchanger, and thus reduce the heat exchange efficiency. Surface deposition may be caused by the accumulation of impurities or reaction products in the fluid on the inner wall of the heat exchanger. These deposits will also increase thermal resistance and affect heat exchange efficiency.

[0058] In order to ensure the normal operation of the heat exchanger, it is necessary to monitor the temperature signals of the inlet and outlet working fluids in real time. Once the temperature is abnormal, that is, the outlet temperature is continuously lower than the set value, and this deviation exceeds a certain threshold, timely measures need to be taken. In this case, the machine can be temporarily shut down and the inside of the heat exchanger can be inspected in detail using an endoscope. Endoscope inspection can intuitively observe the erosion and deposition inside the heat exchanger, providing a basis for subsequent cleaning and maintenance.

[0059] After confirming the erosion and deposition inside the heat exchanger, you can use a high-pressure water gun to clean it. The high-pressure water gun can generate a strong water flow impact force to effectively remove the deposits on the inner wall of the heat exchanger and restore it to a clean state. During the cleaning process, you need to pay attention to the pressure and direction of the water flow to avoid secondary damage to the heat exchanger.

[0060] In addition to physical cleaning, chemical cleaning methods can also be considered, using specific chemical cleaning agents to dissolve and remove deposits. This method is suitable for deposits that are difficult to remove by physical methods, but it is necessary to pay attention to the selection of appropriate cleaning agents to avoid corrosion of heat exchanger materials.

[0061] During actual operation, the heat exchanger may also deposit in certain areas, such as corners or baffles, due to impurities or mixtures of different components in the fluid medium, resulting in partial or complete blockage of the internal channels of the equipment. Monitoring the flow rate and pressure difference between the inlet and outlet is an effective diagnostic method. When blockage occurs inside the heat exchanger, the resistance of the fluid passing through the heat exchanger will increase, which will increase the pressure difference between the inlet and outlet. At the same time, the flow rate will also decrease due to the reduction in the cross-sectional area of ​​the channel. By continuously monitoring the changes in these parameters, potential blockage problems can be discovered in time. In addition, monitoring the vibration frequency is another means of judging the blockage of the heat exchanger. Blockage may cause changes in the fluid dynamics characteristics inside the equipment, which in turn affects the vibration characteristics of the heat exchanger. By analyzing the vibration signal of the heat exchanger during operation, abnormal vibration modes caused by blockage can be identified.

[0062] Get real-time data from the heat exchanger's sensors, including temperature, pressure, flow, vibration and other multi-dimensional physical quantities. Collect sensor data from the equipment and build a real-time data set {(x t ,y t)}, where x t represents the sensor data at time t, y t Indicates the corresponding working status label. Since sensor data is often affected by environmental noise and measurement errors, a series of preprocessing operations are required for the raw data.

[0063] The first step of preprocessing is usually denoising to remove random noise and interference in the data and improve the signal-to-noise ratio of the data. This can be achieved through filters, such as low-pass filters or Kalman filters. Next, the data is normalized to convert data of different dimensions to a unified scale. In addition, for missing data, appropriate strategies need to be adopted to handle it, such as interpolation or using statistical methods to estimate missing values. In the data cleaning stage, outliers need to be identified and processed due to sensor failures or extreme working conditions. After completing these steps, the data will be converted into feature vectors.

[0064] The preprocessed data is input into the fault diagnosis model to obtain the working state and confidence level of the heat exchanger. The specific implementation process of this embodiment includes:

[0065] The collected preprocessed data set is carefully divided into three parts: training set, validation set and test set. This embodiment uses the training set to train the fault diagnosis model. During the training process, the goal of the fault diagnosis model is to minimize the multi-classification logarithmic loss. By adjusting the model parameters, including the learning rate, the number of trees, the depth of the tree, etc., the decision boundary of the model can be optimized to better distinguish different fault categories. After each round of training, the model is evaluated using the validation set, and the model parameters are adjusted according to the evaluation results. This process is called model parameter adjustment, and its purpose is to prevent the model from overfitting. Through the feedback of the validation set, it can be ensured that the model can adapt to new data while maintaining high accuracy. Once the model achieves satisfactory performance on the validation set, the trained model will be used to predict the test set. The prediction results of the model on the test set will directly reflect the performance of the model in the actual working state. At this stage, not only the classification results of the model are concerned, but also the prediction probability of the model for each classification. These probability information are crucial for evaluating the confidence of the model.

[0066] Furthermore, in order to calculate the confidence, this embodiment adopts a method based on prediction probability, that is, calculating the difference between the maximum probability and the second largest probability predicted by the model. This method can intuitively reflect the certainty of the model for its prediction results. If the difference between the maximum probability and the second largest probability is large, it means that the model is very confident in its prediction results; on the contrary, if the difference is small, it means that the model is less certain about its prediction results.

[0067] Furthermore, the goal of the fault diagnosis model is a classification model, which is used to determine whether the heat exchanger is working properly or has a certain fault. This embodiment can define a multi-classification problem, in which the model output is the possible working state category of the heat exchanger. A multi-classification model is trained using LightGBM, with the goal of minimizing the multi-classification logarithmic loss. The objective function of the fault diagnosis model is:

[0068]

[0069] In the formula, n is the number of samples, M is the number of categories, and y i,c is an indicator variable, which is 1 when sample i belongs to category c, and 0 otherwise. is the probability that the fault diagnosis model predicts that sample i belongs to c.

[0070] A time series data set is constructed based on the working state of the heat exchanger and the preprocessed data. As a specific implementation process of this embodiment, the following is included: In order to capture the dynamic changes in the operating state of the heat exchanger, a sliding window technology is used to extract features on the time series. The size of the sliding window is determined according to the operating characteristics of the heat exchanger and the typical time scale of the failure. In each sliding window, statistical features such as mean, variance, and skewness are calculated to reflect the operating state of the heat exchanger in different time periods. At the same time, trend analysis is performed on the time series data to identify periodic patterns. In addition, the time series data is annotated, especially in the time period when the failure is known to occur.

[0071] Adding features to the time series data set to obtain key fault features, as a specific implementation process of this embodiment includes:

[0072] Candidate features are obtained based on the preprocessed data, and then ratio feature calculation is performed, that is, by comparing the measured values ​​of two or more candidate features, their ratio is obtained. The ratio feature in this embodiment is the average value within a certain time window, and then linear discriminant analysis is used to select the most discriminative feature from the candidate features.

[0073] Furthermore, this embodiment combines linear discriminant analysis with time series analysis to extract features of fault signals and obtain key fault features.

[0074] Furthermore, this embodiment uses wavelet transform to convert the time domain signal in the time series data set into a frequency domain signal, thereby extracting the frequency domain features of the signal such as frequency component, amplitude and phase, which can reveal the inherent law of the operating state of the heat exchanger. On this basis, the linear discriminant analysis method is applied to find the best projection direction by maximizing the inter-class divergence and minimizing the intra-class divergence, so that the features of different fault types are maximized in this direction, and the features of the same fault type are clustered as much as possible, thereby realizing feature selection and selecting the most discriminative features. In addition, through further time series feature enhancement analysis, the periodic characteristics of the heat exchanger are extracted, and the understanding of the law of fault signal change over time is deepened.

[0075] The key fault features finally screened out include: the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, the ratio of the operating temperature to the set temperature, the ratio of temperature to pressure, the ratio of the maximum value to the minimum value in the sliding window, and the ratio of maintenance and fault history.

[0076] Furthermore, the expression for calculating the ratio between each pair of candidate features is:

[0077]

[0078] In the formula, represents the average ratio feature within the time window [t-Δt, t], Δt represents the length of the time window, V A,T represents the measured value of sensor A at time T, V B,T Represents the measurement value of sensor B at time T.

[0079] Based on the key fault characteristics, this embodiment builds a fault prediction model based on the regression function of LightGBM. The expression of the model is:

[0080] P = LightGBM(β0+β1R temp-flow +β2R work-set +β3R temp-push +β4R max-min +β5R reqair-history +β6C)

[0081] Where P represents the fault prediction probability, β0, β1, β2, β3, β4, β5, β6 represent the fault prediction model parameters, and R temp-flow It represents the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, R work-set Indicates the ratio of operating temperature to set temperature, R temp-push Represents the ratio of temperature to pressure, R max-min Represents the ratio of the maximum value to the minimum value in the sliding window, R repair-history represents the ratio of protection and failure history, and C represents the confidence level.

[0082] The present invention can realize real-time monitoring of the operating status of the heat exchanger, accurately diagnose faults, and provide reliable fault prediction. In addition, the present invention also has a strong data processing capability to process and analyze a large amount of industrial data, and has good generalization ability and adaptability to adapt to different working conditions and environments.

[0083] The present invention can provide scientific and accurate decision support for the maintenance and overhaul of heat exchangers, thereby significantly improving the operating efficiency and safety of industrial systems.

[0084] Embodiment 2

[0085] like Figure 2 As shown, this embodiment provides a heat exchanger fault intelligent monitoring and analysis system, including: an integrated sensor, a wireless communication module, a fault diagnosis module, a fault prediction module and a user interaction interface;

[0086] Integrated sensors are used to collect real-time data of the heat exchanger in working state. Integrated sensors include but are not limited to: temperature sensor unit, used to measure the inlet and outlet temperatures of the heat exchanger; pressure sensor unit, used to monitor the pressure changes inside the heat exchanger; flow sensor unit, used to record the rate at which the fluid flows through the heat exchanger; vibration sensor unit, used to detect the vibration of the heat exchanger during operation.

[0087] The wireless communication module is used to transmit real-time data to the analysis system. The wireless communication module includes: a data encoding unit, which is used to convert data into a format suitable for wireless transmission; a wireless transmission unit, which is used to send data to the remote analysis system via wireless signals; and a data receiving unit, which is used to receive the wirelessly transmitted data at the analysis system end.

[0088] This embodiment also includes a data preprocessing module, which includes: a signal denoising unit, used to eliminate noise interference in the data; a data normalization unit, used to convert data of different dimensions into a unified scale; and a missing value processing unit, used to fill or estimate missing data points.

[0089] The fault diagnosis module is used to determine the working state of the heat exchanger and the corresponding confidence level based on real-time data. The fault diagnosis module includes: a feature extraction unit, which is used to extract features that are helpful for fault diagnosis from the data; a model training unit, which is used to train a fault diagnosis model, such as the LightGBM model; and a fault judgment unit, which is used to classify and evaluate the confidence level of the heat exchanger state using the trained model.

[0090] The fault prediction module is used to determine the probability of periodic faults based on the confidence of the working state of the heat exchanger and the key fault characteristics. The fault prediction module includes: a feature fusion unit, which is used to integrate the confidence and key fault characteristics; a prediction model unit, which is used to build and train the fault prediction model; and a probability calculation unit, which is used to calculate the probability of heat exchanger failure in the future.

[0091] The user interaction interface is used to reflect the working status of the heat exchanger and the probability of failure in the future cycle to the user, and propose corresponding solution strategies. The user interaction interface includes: a status display unit, which is used to show the real-time working status of the heat exchanger to the user; a fault warning unit, which is used to issue a warning to the user when a potential fault is predicted; and a maintenance suggestion unit, which is used to provide maintenance and repair suggestions based on the fault prediction results.

[0092] Through the collaborative work of the above modules and units, the system can realize intelligent monitoring and analysis of heat exchanger faults, improve the accuracy of fault diagnosis and the reliability of fault prediction, and thus provide strong technical support for the maintenance and operation management of heat exchangers.

[0093] Embodiment 3

[0094] This embodiment also discloses a computer terminal device, including:

[0095] one or more processors;

[0096] A memory, coupled to the processor, for storing one or more programs;

[0097] When one or more programs are executed by one or more processors, the one or more processors implement a heat exchanger fault intelligent monitoring and analysis method.

[0098] Embodiment 4

[0099] This embodiment also discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, a method for intelligent monitoring and analysis of heat exchanger faults is implemented.

[0100] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A heat exchanger fault intelligent monitoring and analysis method, characterized in that: The following steps are involved: Acquiring real-time data of the heat exchanger in a working state and preprocessing the real-time data to obtain preprocessed data; Building a fault diagnosis model and a fault prediction model based on the LightGBM framework, and inputting the preprocessed data into the fault diagnosis model to obtain the working status and confidence of the heat exchanger; constructing a time series data set based on the working state of the heat exchanger and the preprocessed data; Adding features to the time series data set to obtain key fault features; The confidence level and the key fault characteristics are input into the fault prediction model to obtain the probability of periodic fault occurrence.

2. The heat exchanger fault intelligent monitoring and analysis method according to claim 1 is characterized in that: Build a fault diagnosis model based on the LightGBM architecture with the goal of minimizing multi-classification logarithmic loss; Among them, the objective function of the fault diagnosis model is: In the formula, n is the number of samples, M is the number of categories, and y i,c is an indicator variable, which is 1 when sample i belongs to category c, and 0 otherwise. is the probability that the fault diagnosis model predicts that sample i belongs to c.

3. The heat exchanger fault intelligent monitoring and analysis method according to claim 1 is characterized in that: The process of obtaining working status and confidence includes: Divide the preprocessed dataset into training set, validation set and test set; The fault diagnosis model is trained using a training set, and the parameters of the fault diagnosis model are adjusted with the goal of minimizing the multi-classification logarithmic loss; The trained fault diagnosis model is obtained by using the validation set to validate the model after parameter adjustment; Using the trained fault diagnosis model to predict the test set, obtain the working state classification result of the heat exchanger and the prediction probability of each classification; Based on the predicted probabilities, the confidence is calculated using the difference between the maximum probability and the second largest probability.

4. The heat exchanger fault intelligent monitoring and analysis method according to claim 1 is characterized in that: The process of adding features to the time series data set to obtain key fault features includes: Acquire candidate features based on the preprocessed data; Calculate the ratio between each pair of candidate features based on the candidate features to obtain a ratio feature; The ratio features are screened based on linear discriminant analysis to obtain key fault features.

5. The heat exchanger fault intelligent monitoring and analysis method according to claim 4 is characterized in that: The expression for calculating the ratio between each pair of candidate features based on the candidate features is: In the formula, represents the average ratio feature within the time window [t-Δt, t], Δt represents the length of the time window, V A,T represents the measured value of sensor A at time T, V B,T Represents the measurement value of sensor B at time T.

6. The heat exchanger fault intelligent monitoring and analysis method according to claim 4 is characterized in that: The key fault features include: the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, the ratio of the operating temperature to the set temperature, the ratio of temperature to pressure, the ratio of the maximum value to the minimum value in the sliding window, and the ratio of maintenance and fault history.

7. The heat exchanger fault intelligent monitoring and analysis method according to claim 1 is characterized in that: The expression of the fault prediction model is: P=LightGBM(β0+β1R temp-flow +β2R work-set +β3R temp-push +β4R max-min +β5R repair-history +β6C) Where P represents the fault prediction probability, β0, β1, β2, β3, β4, β5, β6 represent the fault prediction model parameters, and R temp-flow It represents the ratio of the inlet and outlet temperature difference of the heat exchanger to the fluid flow rate, R work-set Indicates the ratio of operating temperature to set temperature, R temp-push Represents the ratio of temperature to pressure, R max-min Represents the ratio of the maximum value to the minimum value in the sliding window, R repair-history represents the ratio of protection and failure history, and C represents the confidence level.

8. A heat exchanger fault intelligent monitoring and analysis system, characterized in that: include: Integrate sensors, wireless communication modules, fault diagnosis modules, fault prediction modules and user interaction interfaces; Wherein, the integrated sensor is used to collect real-time data of the heat exchanger in the working state; The wireless communication module is used to transmit the real-time data to the analysis system; The fault diagnosis module is used to determine the working state of the heat exchanger and the corresponding confidence level based on the real-time data; The fault prediction module is used to determine the probability of periodic fault occurrence based on the confidence of the working state of the heat exchanger and the key fault characteristics; The user interaction interface is used to reflect the working status of the heat exchanger and the probability of failure in the future cycle to the user, and propose corresponding solution strategies.

9. A computer terminal device, characterized in that: include: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the heat exchanger fault intelligent monitoring and analysis method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the heat exchanger fault intelligent monitoring and analysis method according to any one of claims 1 to 7 is implemented.

Images