A monitoring system for gas-liquid distributors in packed towers for catalytic reactions

Abstract

The present invention relates to the technical field of monitoring systems, and discloses a monitoring system for a gas-liquid distributor of a packed tower for catalytic reactions, comprising a particle image velocimeter installed on the gas-liquid distributor of the packed tower, at least one pressure sensor installed on the cross section of the distributor for measuring the pressure value of each point; a differential pressure transmitter installed between the inlet and outlet of the distributor; at least one electromagnetic flowmeter installed on the inlet and outlet of the distributor; a vibration sensor installed on each supporting structure of the distributor; a window camera installed on the window of the distributor; and a laser level installed on the distributor, wherein the particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level are respectively connected to the Internet through a wireless data access module. The present invention can issue an early warning when the system performance begins to deviate from the normal range, which helps to prevent potential failures or process deterioration.

Description

Technical Field

[0001] The invention relates to the technical field of distributor monitoring systems, in particular to a gas-liquid distributor monitoring system for a packed tower of a catalytic reaction. Background Art

[0002] Gas-liquid distributors are key components in packed towers. In the chemical production process, many process steps involve material exchange between gas and liquid phases, such as absorption, desorption, and distillation. These processes require that the gas and liquid phases can fully contact each other to achieve efficient mass transfer. As a commonly used gas-liquid contact equipment, packed towers provide a wide contact space for gas and liquid phases by filling them with fillers with a large specific surface area. However, in order to make the gas and liquid phases evenly distributed and effectively contact each other in the packing layer, it is necessary to rely on gas-liquid distributors to achieve this.

[0003] The main function of the gas-liquid distributor is to distribute the gas and liquid entering the packed tower evenly and stably to the packing layer, ensuring that the gas-liquid two-phase forms an effective contact interface on the packing surface. Through reasonable distribution design, it is possible to avoid excessive or low local gas-liquid concentration, reduce the occurrence of undesirable flow phenomena such as channel flow and wall flow, and thus improve mass transfer efficiency and separation effect. The design of gas-liquid distributors is diverse to adapt to different operating conditions and tower diameters. Common types include trough liquid distributors, tube distributors, spray distributors, porous plate distributors, etc. These distributors have their own characteristics. For example, trough distributors are suitable for occasions with large liquid flow, and spray distributors are suitable for occasions that require fine mist spraying. Factors such as the material, structure, pore size, and distribution density of the distributor need to be considered during design.

[0004] Currently, monitoring of the liquid distributor mainly includes flow monitoring: by installing flow meters at the top and bottom of the tower, the flow of gas and liquid is monitored in real time to ensure that the two are evenly distributed in the tower.

[0005] Pressure monitoring: Pressure sensors are installed at different heights of the tower to monitor the pressure changes in each section of the tower to determine whether the gas-liquid flow state is normal.

[0006] Temperature monitoring: Temperature sensors are used to monitor the temperature distribution in the tower to evaluate mass transfer efficiency and heat exchange.

[0007] Image recognition uses cameras and image recognition technology to monitor the gas-liquid distribution in the packing layer and detect abnormalities in a timely manner.

[0008] Data Analysis and Modeling: Combine sensor data with operating parameters and use data analysis software to build an operational model of the packed tower to predict and diagnose potential problems.

[0009] The current monitoring system for liquid distributors has the following defects; although the existing monitoring technology has played an important role in improving the stability and efficiency of packed tower operation, there are still some defects:

[0010] Limited monitoring accuracy: Due to the influence of sensor accuracy and installation location, some monitoring data may have errors and it is difficult to fully reflect the actual situation in the tower.

[0011] Insufficient real-time performance: In some cases, the update speed of monitoring data may not meet the needs of rapid response, resulting in delayed problem discovery and processing.

[0012] Lack of intelligent diagnosis: The current monitoring system mainly focuses on data collection and display, while there is still room for improvement in intelligent fault diagnosis and prediction.

[0013] Poor adaptability to specific working conditions: The operating conditions and working conditions of different packed towers vary greatly. In particular, the current monitoring of enterprise distributors is still based on setting thresholds for monitoring data such as pressure sensors, and a simple triggering warning mechanism when the threshold is exceeded and an alarm is issued. In actual applications, since, for example, there are multiple pressure sensors installed on a liquid distributor, the traditional method is to issue an alarm when one of the pressure sensors exceeds the threshold for a period of time. However, it is often normal for a single sensor to briefly exceed the threshold, and in actual production, there are multiple sensor values ​​that exceed the threshold but the system is operating normally. Therefore, the traditional method of simply exceeding the threshold and issuing an alarm is not intelligent, and will easily increase the burden on staff, be time-consuming and labor-intensive, and greatly increase the maintenance cost and energy consumption of the equipment.

[0014] Therefore, in order to solve the above problems, a monitoring system for the gas-liquid distributor of a packed tower for catalytic reaction is needed. Summary of the invention

[0015] The object of the present invention is to provide a monitoring system for a gas-liquid distributor in a packed tower for a catalytic reaction. The present invention can consider multiple sensor data at the same time, thereby providing a comprehensive system status assessment, avoiding false alarms or missed alarms that may be caused by relying solely on a single sensor threshold. The abnormality detection capability, the present invention identifies the normal range of system behavior, and can detect abnormalities in complex patterns or interactive effects that may not be displayed by a single sensor data. It can consider multiple sensor data at the same time, thereby providing a comprehensive system status assessment, avoiding false alarms or missed alarms that may be caused by relying solely on a single sensor threshold. The present invention can provide early warning of the system. By continuously monitoring statistics (such as T² and Q statistics), the present invention can issue early warnings when system performance begins to deviate from the normal range, helping to prevent potential failures or process deterioration.

[0016] The present invention is achieved in that:

[0017] The present invention provides a monitoring system for a gas-liquid distributor of a packed tower for catalytic reaction, comprising a particle image velocimeter installed on the gas-liquid distributor of the packed tower, and at least one pressure sensor installed on the cross section of the distributor for measuring the pressure value of each point; reflecting the uniformity of fluid distribution;

[0018] The differential pressure transmitter installed between the distributor inlet and outlet is used to measure the pressure difference between the distributor inlet and outlet to determine whether there is a blockage.

[0019] At least one electromagnetic flowmeter installed at the inlet and outlet of the distributor; used to monitor the gas or liquid flow through the distributor;

[0020] Vibration sensors installed on each supporting structure of the distributor; used to monitor the vibration data of the distributor and its supporting structure;

[0021] A window camera installed on the distributor window is used to observe the flow of fluid inside the distributor and whether there is any blockage through visual inspection;

[0022] A laser level installed on the distributor to measure the level of the distributor;

[0023] The particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level are respectively connected to the Internet through a wireless data access module, wherein the wireless data access module includes a Wifi module, a Bluetooth module, a Lora module or a 4G DTU.

[0024] A monitoring terminal is connected via the Internet, and the monitoring terminal includes a server connected to the particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level, and the server is connected to a PC terminal, and the server is wirelessly connected to at least one intelligent terminal.

[0025] Furthermore, the fluid velocity in the distributor is measured by a particle image velocimeter, the pressure difference between the inlet and outlet of the distributor is measured by a differential pressure transmitter, the gas or liquid flow through the distributor is monitored by an electromagnetic flowmeter, the vibration data of the distributor and its supporting structure are monitored by a vibration sensor, the fluid flow inside the distributor is visually inspected by a window camera, and the levelness of the distributor is measured by a laser level.

[0026] The smart terminal is a smart phone or a tablet computer, and there are multiple smart terminals.

[0027] Furthermore, the present invention provides a method for monitoring a gas-liquid distributor of a packed tower for a catalytic reaction, which is specifically performed in the server according to the following steps:

[0028] S1: First, the particle image velocimetry data, pressure data, differential pressure data, electromagnetic flow data, vibration data and laser level data are cleaned through the data acquisition and cleaning module; specifically, outliers are removed, thresholds are set, and data points beyond the normal range are eliminated;

[0029] Fill missing values, use interpolation or regression methods to fill missing data values; remove duplicate data caused by repeated sampling or data redundancy; and remove data noise through filters, as shown in formula (1)-formula (2);

[0030] Formula (1)

[0031] Formula (2)

[0032] in, is the output data after filtering, is the input signal, is the smoothing factor, ranging between 0 and 1, is the window size, the number of samples used for averaging;

[0033] S2: Calculate the average flow rate of gas and liquid in the distributor; as shown in formula (3);

[0034] Formula (3)

[0035] in, yes The flow reading at each time point, is the total sample size;

[0036] S3: Calculate the average pressure difference before and after the distributor, as shown in formula (4);

[0037] Formula (4)

[0038] S4: Use variance or to express the uniformity of gas-liquid distribution in the distributor, as shown in formula (5);

[0039] Formula (5)

[0040] in, is the average value, It is The readings of the measuring points;

[0041] S5: In the server, through the data model training module, the data sorted out in steps S1-S4 is used to build a PCA model using the historical data set, and the main change direction of the data is found by finding the eigenvector corresponding to the maximum eigenvalue of the data covariance matrix, and the principal component analysis of the normal operating status data of the distributor is performed; as shown in formula (6);

[0042] Formula (6)

[0043] in, is the score matrix, is the transpose of the loading matrix, and The projections corresponding to the principal components and the original variables respectively;

[0044] S6: Establish control limits and calculate two key statistics based on the PCA model. Statistics and Statistics; such as formula (7)-formula (8);

[0045] Formula (7)

[0046] Formula (8)

[0047] in, is the PCA score vector, is a scoring matrix estimated based on historical data The covariance matrix of is the reconstruction error vector;

[0048] S7: Determine control limits based on historical data sets and statistic, calculate its corresponding control limits, using Distribution and The distribution is determined at a 95% or 99% confidence level;

[0049] S8: In the server, the abnormal judgment module uses the model to perform real-time monitoring, updates the PCA model with the newly collected sensor data, and calculates the real-time and Statistics, calculated in real time and The statistics are compared with the control limits determined in step S7; when the real-time or When the statistic exceeds its corresponding control limit, it is judged that the current operating state deviates from the normal range and there is an abnormality;

[0050] S9: Anomalies are detected, and the PCA loading matrix is ​​further analyzed to determine the data with the largest abnormal contribution from the variables or sensor data, based on the results of the anomaly location;

[0051] S10: In the server, an abnormal warning information is sent to the intelligent terminal through the warning information sending module for the abnormal result of the positioning.

[0052] Further, the data of the window camera is used to construct a PCA model to convert the image data into a numerical feature vector, which is specifically performed in the following steps:

[0053] S6.1: First, perform image preprocessing to convert the color image into a grayscale image, as shown in formula (9);

[0054] Formula (9)

[0055] in, Represents the pixel intensity of red, green, and blue channels respectively;

[0056] S6.2: Scale the image to a uniform size and perform normalization to normalize the pixel values ​​to the interval [0, 1]; as shown in equations (10)-(11);

[0057] Formula (10)

[0058] in, for The image matrix, is its pixel value, and are the row and column indices respectively;

[0059] Formula (11)

[0060] in, and Row and column indices starting from 0, is the width and number of columns of the image,

[0061] S6.3: Combining the feature vectors of all images into a data matrix , each row represents the feature vector of an image, and the data matrix is ​​centered so that the mean of each column is zero;

[0062] S6.4: Based on the centralized data, calculate the covariance matrix C;

[0063] As shown in formula (12) - formula (13);

[0064] Formula (12)

[0065] Formula (13)

[0066] in, is the first matrix after centralization List, is the original data matrix List, For one A matrix where each row is the same vector , is the mean of each column.

[0067] S6.5: Before selection The eigenvectors corresponding to the largest eigenvalues ​​constitute the principal component matrix ;

[0068] S6.6: Apply PCA using the principal component matrix The data matrix Projected into low-dimensional space, we get the score matrix .

[0069] Furthermore, the server is provided with a warning information issuing module to issue warning information, and the warning information issuing objects include the project leader's smart terminal, the on-duty personnel's smart terminal and the fault corresponding processing personnel's smart terminal. The project leader's smart terminal and the fault corresponding processing personnel's smart terminal are respectively provided with an abnormal processing situation and equipment status module, and an abnormal situation processing summary and suggestion module. The warning terminal directly issued to the responsible person can flexibly allocate the person in charge of the plan, and implement it in time, provide an abnormal situation processing summary after the fault is handled, and have a significant role in promoting subsequent equipment maintenance and reducing the failure rate.

[0070] Furthermore, the present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, the method described above is implemented.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] 1. Ability to consider multiple sensor data simultaneously to provide a comprehensive system status assessment, avoiding false positives or negatives that may result from relying solely on a single sensor threshold.

[0073] Anomaly detection capabilities:

[0074] The present invention uses the statistical model PCA to identify the normal range of system behavior and can detect anomalies in complex patterns or interaction effects that may not be displayed by single sensor data.

[0075] 2. Conduct system early warning:

[0076] By continuously monitoring statistics such as T² and Q statistics, the present invention can provide early warnings when system performance begins to deviate from normal ranges, helping to prevent potential failures or process deterioration.

[0077] 3. The present invention assists in fault diagnosis. When an abnormality is detected, the present invention can help locate the source of the abnormality and identify which sensor data or process variables have the most significant changes by analyzing the load matrix of the model, thereby guiding fault diagnosis.

[0078] 4. The present invention reduces maintenance costs. Through accurate anomaly detection and early warning, it can reduce unnecessary unplanned downtime, reduce maintenance costs, avoid excessive maintenance, and improve equipment availability and efficiency.

[0079] 5. Optimize process control. The real-time monitoring and analysis results provided by the present invention can be used to adjust process parameters and optimize control strategies, thereby improving process performance and product quality.

[0080] 6. The method of the present invention can be adjusted as process characteristics and operating conditions change, and its effectiveness can be maintained by regularly updating and verifying the model. The comprehensive evaluation and anomaly detection information provided can be used as a decision support tool to help engineers and managers make more informed operation and maintenance decisions based on data. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It is understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0082] Figure 1 is a flow chart of the method of the present invention;

[0083] Figure 2 It is a system structure diagram of the present invention;

[0084] Figure 3 It is a server system diagram of the present invention. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but is only for selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0086] See also Figure 1-Figure 3 , the present invention provides a monitoring system for a gas-liquid distributor of a packed tower for a catalytic reaction;

[0087] The present invention provides a monitoring system for a gas-liquid distributor of a packed tower for catalytic reaction, comprising a particle image velocimeter installed on the gas-liquid distributor of the packed tower, and at least one pressure sensor installed on the cross section of the distributor for measuring the pressure value of each point; reflecting the uniformity of fluid distribution;

[0088] The differential pressure transmitter installed between the distributor inlet and outlet is used to measure the pressure difference between the distributor inlet and outlet to determine whether there is a blockage.

[0089] At least one electromagnetic flowmeter installed at the inlet and outlet of the distributor; used to monitor the gas or liquid flow through the distributor;

[0090] Vibration sensors installed on each supporting structure of the distributor; used to monitor the vibration data of the distributor and its supporting structure;

[0091] A window camera installed on the distributor window is used to observe the flow of fluid inside the distributor and whether there is any blockage through visual inspection;

[0092] A laser level installed on the distributor to measure the level of the distributor;

[0093] The particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level are respectively connected to the Internet through a wireless data access module, wherein the wireless data access module includes a Wifi module, a Bluetooth module, a Lora module or a 4G DTU.

[0094] A monitoring terminal is connected via the Internet, and the monitoring terminal includes a server connected to the particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level, and the server is connected to a PC terminal, and the server is wirelessly connected to at least one intelligent terminal. The velocity of the fluid in the distributor is measured by the particle image velocimeter, the pressure difference between the inlet and outlet of the distributor is measured by the differential pressure transmitter, the gas or liquid flow through the distributor is monitored by the electromagnetic flowmeter, the vibration data of the distributor and its supporting structure are monitored by the vibration sensor, the flow of the fluid inside the distributor is visually inspected by the window camera, and the horizontality of the distributor is measured by the laser level.

[0095] The smart terminal is a smart phone or a tablet computer, and there are multiple smart terminals.

[0096] In this embodiment, the present invention provides a method for monitoring a gas-liquid distributor of a packed tower for a catalytic reaction, which is specifically performed in the server according to the following steps:

[0097] S1: First, the particle image velocimetry data, pressure data, differential pressure data, electromagnetic flow data, vibration data and laser level data are cleaned through the data acquisition and cleaning module; specifically, outliers are removed, thresholds are set, and data points beyond the normal range are eliminated;

[0098] Fill missing values, use interpolation or regression methods to fill missing data values; remove duplicate data caused by repeated sampling or data redundancy; and remove data noise through filters, as shown in formula (1)-formula (2);

[0099] Formula (1)

[0100] Formula (2)

[0101] in, is the output data after filtering, is the input signal, is the smoothing factor, ranging between 0 and 1, is the window size, the number of samples used for averaging;

[0102] S2: Calculate the average flow rate of gas and liquid in the distributor; as shown in formula (3);

[0103] Formula (3)

[0104] in, yes The flow reading at each time point, is the total sample size;

[0105] S3: Calculate the average pressure difference before and after the distributor, as shown in formula (4);

[0106] Formula (4)

[0107] S4: Use variance or to express the uniformity of gas-liquid distribution in the distributor, as shown in formula (5);

[0108] Formula (5)

[0109] in, is the average value, It is The readings of the measuring points;

[0110] S5: In the server, through the data model training module, the data sorted out in steps S1-S4 is used to build a PCA model using the historical data set, and the main change direction of the data is found by finding the eigenvector corresponding to the maximum eigenvalue of the data covariance matrix, and the principal component analysis of the normal operating status data of the distributor is performed; as shown in formula (6);

[0111] Formula (6)

[0112] in, is the score matrix, is the transpose of the loading matrix, and The projections corresponding to the principal components and the original variables respectively;

[0113] S6: Establish control limits and calculate two key statistics based on the PCA model. Statistics and Statistics; such as formula (7)-formula (8);

[0114] Formula (7)

[0115] Formula (8)

[0116] in, is the PCA score vector, is a scoring matrix estimated based on historical data The covariance matrix of is the reconstruction error vector;

[0117] S7: Determine control limits based on historical data sets and statistic, calculate its corresponding control limits, using Distribution and The distribution is determined at a 95% or 99% confidence level;

[0118] S8: In the server, the abnormal judgment module uses the model to perform real-time monitoring, updates the PCA model with the newly collected sensor data, and calculates the real-time and Statistics, calculated in real time and The statistics are compared with the control limits determined in step S7; when the real-time or When the statistic exceeds its corresponding control limit, it is judged that the current operating state deviates from the normal range and there is an abnormality;

[0119] S9: Anomalies are detected, and the PCA loading matrix is ​​further analyzed to determine the data with the largest abnormal contribution from the variables or sensor data, based on the results of the anomaly location;

[0120] S10: In the server, an abnormal warning information is sent to the intelligent terminal through the warning information sending module for the abnormal result of the positioning.

[0121] In this embodiment, the data of the window camera is used to construct a PCA model, and the image data is converted into a numerical feature vector, which is specifically performed in the following steps:

[0122] S6.1: First, perform image preprocessing to convert the color image into a grayscale image, as shown in formula (9);

[0123] Formula (9)

[0124] in, Represents the pixel intensity of red, green, and blue channels respectively;

[0125] S6.2: Scale the image to a uniform size and perform normalization to normalize the pixel values ​​to the interval [0, 1]; as shown in equations (10)-(11);

[0126] Formula (10)

[0127] in, for The image matrix, is its pixel value, and are the row and column indices respectively;

[0128] Formula (11)

[0129] in, and Row and column indices starting from 0, is the width and number of columns of the image; for example, a 3×3 image matrix , the element matrix is ​​as follows:

[0130]

[0131] Convert it to a one-dimensional vector , as follows;

[0132] .

[0133] S6.3: Combining the feature vectors of all images into a data matrix , each row represents the feature vector of an image, and the data matrix is ​​centered so that the mean of each column is zero;

[0134] S6.4: Based on centralized data, for example images, each of which has been converted into a The one-dimensional vector of features, the feature vectors of all images are stacked together to form a The data matrix , each row represents the feature vector of an image, and each column represents the performance of a specific feature in all images. The feature vector of an image is , then the data matrix It can be expressed as

[0135]

[0136] Among them, among them, Indicates The image The value of a feature;

[0137] Calculate the covariance matrix C;

[0138] Assumptions No. The mean of the column is , then the centralized data matrix You can use Subtract the mean value of each column from the column to get: ;

[0139] in is a A matrix where each row is the same vector , more specifically, the mean of each column , as shown in formula (12)-formula (13);

[0140] Formula (12)

[0141] Formula (13)

[0142] in, is the first matrix after centralization List, is the original data matrix List, For one A matrix where each row is the same vector , is the mean of each column.

[0143] S6.5: Before selection The eigenvectors corresponding to the largest eigenvalues ​​constitute the principal component matrix ;

[0144] S6.6: Apply PCA using the principal component matrix The data matrix Projected into low-dimensional space, we get the score matrix .

[0145] Furthermore, the server is provided with a warning information issuing module to issue warning information, and the warning information issuing objects include the project leader's smart terminal, the on-duty personnel's smart terminal and the fault corresponding processing personnel's smart terminal. The project leader's smart terminal and the fault corresponding processing personnel's smart terminal are respectively provided with an abnormal processing situation and equipment status module, and an abnormal situation processing summary and suggestion module. The warning terminal directly issued to the responsible person can flexibly allocate the person in charge of the plan, and implement it in time, provide an abnormal situation processing summary after the fault is handled, and have a significant role in promoting subsequent equipment maintenance and reducing the failure rate.

[0146] In this embodiment, the present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, the method described above is implemented.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention has various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A monitoring system for a gas-liquid distributor of a packed tower for a catalytic reaction, characterized in that: It includes a particle image velocimeter installed on the gas-liquid distributor of the packed tower, at least one pressure sensor installed on the cross section of the distributor for measuring the pressure value of each point; and a differential pressure transmitter installed between the inlet and outlet of the distributor; At least one electromagnetic flow meter installed at the inlet and outlet of the distributor; Vibration sensors installed on each support structure of the distributor; window cameras installed on the distributor window; laser level installed on the distributor, The particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level are respectively connected to the Internet through a wireless data access module, and are connected to a monitoring terminal through the Internet. The monitoring terminal includes a server connected to the particle image velocimeter, pressure sensor, differential pressure transmitter, electromagnetic flowmeter, vibration sensor, window camera and laser level, and the server is connected to a PC terminal, and the server is wirelessly connected to at least one intelligent terminal; On the server, follow these steps: S1: First, the particle image velocimetry data, pressure data, differential pressure data, electromagnetic flow data, vibration data and laser level data are cleaned through the data acquisition and cleaning module; specifically, outliers are removed, thresholds are set, and data points beyond the normal range are eliminated; Fill missing values, use interpolation or regression methods to fill missing data values; remove duplicate data caused by repeated sampling or data redundancy; and remove data noise through filters, as shown in formula (1)-formula (2); Formula (1); Formula (2); in, is the output data after filtering, is the input signal, is the smoothing factor, ranging between 0 and 1, is the window size, the number of samples used for averaging; S2: Calculate the average flow rate of gas and liquid in the distributor; as shown in formula (3); Formula (3); in, is the flow reading at time point i, is the total sample size; S3: Calculate the average pressure difference before and after the distributor, as shown in formula (4); Formula (4); S4: Use variance or to express the uniformity of gas-liquid distribution in the distributor, as shown in formula (5); Formula (5); in, is the average value, is the reading of the ith measurement point; S5: In the server, through the data model training module, the data sorted out in steps S1-S4 is used to build a PCA model using the historical data set, and the main change direction of the data is found by finding the eigenvector corresponding to the maximum eigenvalue of the data covariance matrix, and the principal component analysis of the normal operating status data of the distributor is performed; as shown in formula (6); Formula (6); in, is the score matrix, is the transpose of the loading matrix, and The projections corresponding to the principal components and the original variables respectively; S6: Establish control limits and calculate two key statistics based on the PCA model. Statistics and Statistics; such as formula (7)-formula (8); Formula (7); Formula (8); in, is the PCA score vector, is a scoring matrix estimated based on historical data The covariance matrix of is the reconstruction error vector; S7: Determine control limits based on historical data sets and Q statistics, and calculate their corresponding control limits using the F distribution and The distribution is determined at a 95% or 99% confidence level; S8: In the server, the abnormal judgment module uses the model to perform real-time monitoring, updates the PCA model with the newly collected sensor data, and calculates the real-time and Q statistics, which will be calculated in real time and Q statistics are compared with the control limits determined in step S7; when the real-time Or when the Q statistic exceeds its corresponding control limit, it is judged that the current operating state deviates from the normal range and there is an abnormality; S9: Anomalies are detected, and the PCA loading matrix is ​​further analyzed to determine the data with the largest abnormal contribution from the variables or sensor data, based on the results of the anomaly location; S10: In the server, an abnormal warning information is sent to the intelligent terminal through the warning information sending module for the abnormal result of the positioning.

2. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 1, characterized in that: The fluid velocity in the distributor is measured by a particle image velocimeter, the pressure difference between the inlet and outlet of the distributor is measured by a differential pressure transmitter, the gas or liquid flow through the distributor is monitored by an electromagnetic flowmeter, the vibration data of the distributor and its supporting structure are monitored by a vibration sensor, the fluid flow inside the distributor is visually inspected by a window camera, and the levelness of the distributor is measured by a laser level.

3. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 1, characterized in that: in, The wireless data access module includes a Wifi module, a Bluetooth module, a Lora module or a 4G DTU.

4. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 1, characterized in that: The smart terminal is a smart phone or a tablet computer, and there are multiple smart terminals.

5. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 1, characterized in that: The data from the window camera is used to build the PCA model and convert the image data into a numerical feature vector. The specific steps are as follows: S6.1: First, perform image preprocessing to convert the color image into a grayscale image, as shown in formula (9); Formula (9); in, Represents the pixel intensity of red, green, and blue channels respectively; S6.2: Scale the image to a uniform size and perform normalization to normalize the pixel values ​​to the interval [0, 1]; as shown in equations (10)-(11); Formula (10); in, for The image matrix, is its pixel value, and are the row and column indices respectively; Formula (11); in, and Row and column indices starting from 0, is the width and number of columns of the image, S6.3: Combining the feature vectors of all images into a data matrix , each row represents the feature vector of an image, and the data matrix is ​​centered so that the mean of each column is zero; S6.4: Based on the centralized data, calculate the covariance matrix C; S6.5: Before selection k The eigenvectors corresponding to the largest eigenvalues ​​constitute the principal component matrix P ; S6.6: Apply PCA using the principal component matrix The data matrix Projected into low-dimensional space, we get the score matrix .

6. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 5, characterized in that: In step S6.4, as in equation (12)-equation (13); Formula (12); Formula (13); in, is the first matrix after centralization List, is the original data matrix List, For one A matrix where each row is the same vector , is the mean of each column.

7. The gas-liquid distributor monitoring system for a packed tower of a catalytic reaction according to claim 1, characterized in that: The server is provided with a warning information issuing module for issuing warning information. The warning information issuing objects include the project leader's smart terminal, the on-duty personnel's smart terminal and the fault corresponding processing personnel's smart terminal. The project leader's smart terminal and the fault corresponding processing personnel's smart terminal are respectively provided with an abnormal handling situation and equipment status module, and an abnormal situation handling summary and suggestion module.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the main controller, the system as claimed in any one of claims 1 to 7 is implemented.

Images