An intelligent control system for acid mist absorption tower based on big data

Abstract

The present application provides an intelligent control system for an acid mist absorption tower based on big data, which relates to the field of automated control technology, including: an operation data collection unit collects the operation data of the acid mist absorption tower, and constructs a large database including a local monitoring database and a same-family collaborative monitoring database; a relationship fitting unit fits the influence relationship between the absorption evaluation results and the control parameters according to the local monitoring database, and constructs a preliminary adaptive control model; a model optimization unit performs incremental learning and verification on the preliminary adaptive control model according to the same-family collaborative monitoring database to obtain an adaptive control model; a model deployment unit disassembles the input, operation, and output logical relationships of the adaptive control model, and deploys a modular control system. The present application solves the technical problem that the prior art lacks adaptive adjustment means for real-time operating conditions, resulting in unstable control effects and slow control response speeds, thereby improving the control accuracy and control response speed of the acid mist absorption tower.

Description

Technical Field

[0001] The present application relates to the field of automation control technology, and in particular to an intelligent control system for an acid mist absorption tower based on big data. Background Art

[0002] Acid mist absorption tower is an environmental protection equipment specially used for treating acidic gas. It is widely used in chemical, metallurgy, electric power and other industries. The equipment uses the principle of acid-base neutralization to bring acidic gas into contact with alkaline absorption liquid to achieve gas purification.

[0003] Currently, many acid mist absorbers still rely on traditional control methods based on experience or preset rules. These methods typically employ fixed control models and perform manual or automatic adjustments by monitoring parameters such as gas flow, liquid flow, temperature, and pressure. Due to variations in exhaust gas composition, climatic conditions, and equipment aging, traditional control methods struggle to respond to complex environmental changes in real time, resulting in reduced control accuracy and impacting the operational stability of the acid mist absorber. Under these circumstances, the operating efficiency and exhaust gas treatment effectiveness of the acid mist absorber cannot be continuously optimized, and may even lead to energy waste and substandard environmental performance. Summary of the Invention

[0004] The present application provides an intelligent control system for an acid mist absorption tower based on big data, which solves the technical problem that the existing technology lacks adaptive adjustment means for real-time operating conditions and is unable to cope with complex and changeable operating environments, resulting in unstable control effects and slow control response speeds of the acid mist absorption tower, thereby affecting the operational stability of the acid mist absorption tower. The application achieves the technical effect of improving the control accuracy and control response speed of the acid mist absorption tower, thereby enhancing the operational stability of the acid mist absorption tower.

[0005] In view of the above problems, the present application provides an intelligent control system for an acid mist absorption tower based on big data, the system comprising: an operation data collection unit for collecting the operation data of the acid mist absorption tower through sensors and monitoring equipment, including gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and pH value of the absorption liquid, and constructing a large database, including a local monitoring database and a same-family collaborative monitoring database; a relationship fitting unit for extracting key features of absorption control based on the local monitoring database, fitting the influence relationship between absorption evaluation results and control parameters, and constructing a preliminary adaptive control model; a model optimization unit for constructing a test data set, an augmented reality data set, and a similar collaborative monitoring database based on the same-family collaborative monitoring database. An incremental learning set is used to perform incremental learning on the preliminary adaptive control model, and the test data set is used to verify the preliminary adaptive control model until the convergence target is met, so as to obtain an adaptive control model; a model deployment unit is used to disassemble the input, operation, and output logical relationship of the adaptive control model, and deploy a modular control system including an input module, a logic operation module, and an output module. The input module is used to connect the acquisition path, and the logic operation module has the adaptive control model built in to perform control operation analysis on the real-time monitoring data. The output module is used to feed back the control parameters output by the logic operation module to the control center for parameter control.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By combining the local monitoring database with a collaborative monitoring database from the same family, a multidimensional dataset spanning regions and devices was constructed. This big data integration not only covers traditional monitoring data (such as gas flow, liquid flow, and temperature), but also includes key indicators such as tail gas acid mist concentration and absorption liquid pH, providing rich data support for subsequent control decisions and enhancing data comprehensiveness. A relationship fitting unit extracts key control features of the absorber tower and establishes a relationship between absorption effect and control parameters, thus providing a precise control model for the operation of the acid mist absorber tower. This model not only makes adaptive adjustments based on actual operating data but also dynamically optimizes operating parameters, improving control effectiveness and the stability of the acid mist absorber tower. Using an incremental learning method, combined with the collaborative monitoring database to optimize the control model, the control system can be continuously optimized and improved during operation. This allows for rapid adjustment of control strategies to changes in the environment and operating conditions, reducing errors and improving control response speed. The adaptive control model is broken down into an input module, a logic operation module, and an output module, forming a modular control system. This modular design provides the control system with high flexibility and maintainability, enabling it to better adapt to diverse site conditions and equipment requirements. Moreover, the modular control system enables clear separation of data acquisition, calculation and output, facilitating precise control and data management.

[0008] In summary, this application collects and analyzes multi-dimensional operating data in real time, builds and optimizes an adaptive control model, and can automatically adjust control parameters according to real-time operating condition changes, significantly improving the control accuracy and stability of the acid mist absorption tower, and ensuring the optimal operation of the acid mist absorption tower under different operating conditions. Rich data sources and multi-stage model optimization improve the accuracy, adaptability and robustness of the control model. Modular system deployment facilitates maintenance and upgrades, and can quickly adjust control strategies when encountering new operating conditions, enhancing response speed and control flexibility, effectively improving the acid mist absorption effect, and reducing tail gas acid mist emissions.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic structural diagram of an intelligent control system for an acid mist absorption tower based on big data is provided in an embodiment of the present application.

[0011] Figure 2 A schematic diagram of the process of constructing a big database in an intelligent control system of an acid mist absorption tower based on big data provided in an embodiment of the present application.

[0012] Figure 3 A schematic diagram of the process of extracting key features of absorption control in an intelligent control system of an acid mist absorption tower based on big data provided in an embodiment of the present application.

[0013] Description of the reference numerals: operation data collection unit 10, relationship fitting unit 20, model optimization unit 30, model deployment unit 40. DETAILED DESCRIPTION

[0014] The embodiment of the present application provides an intelligent control system for an acid mist absorption tower based on big data, thereby solving the technical problem that the prior art lacks adaptive adjustment means for real-time operating conditions and is unable to cope with complex and changeable operating environments, resulting in unstable control effects and slow control response speeds of the acid mist absorption tower, thereby affecting the operational stability of the acid mist absorption tower. The embodiment of the present application achieves the technical effect of improving the control accuracy and control response speed of the acid mist absorption tower, thereby enhancing the operational stability of the acid mist absorption tower.

[0015] like Figure 1 As shown, the embodiment of the present application provides an intelligent control system for an acid mist absorption tower based on big data, the system comprising:

[0016] The operation data collection unit 10 is used to collect the operation data of the acid mist absorption tower through sensors and monitoring equipment, including gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and absorption liquid pH value, and build a large database, including a local monitoring database and a family collaborative monitoring database.

[0017] Specifically, the local monitoring database is a database that stores monitoring data from the same acid mist absorption tower and is used to analyze and extract key control features. The same-family collaborative monitoring database is a database that stores monitoring data from similar or same-type acid mist absorption towers and is used for model testing and optimization.

[0018] The operation data collection unit 10 collects a number of key operation data through sensors and monitoring equipment to build a large database. These operation data include key parameters such as gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration and pH value of the absorption liquid. Through the collection of these data, the operating status of the acid mist absorption tower can be understood in real time. For example, changes in gas flow and liquid flow reflect the load condition of the absorption tower, while the tail gas acid mist concentration and the pH value of the absorption liquid are directly related to the absorption effect. The operation data collected by the sensors and monitoring equipment are stored in the local monitoring database. At the same time, the same family collaborative monitoring database is used to collect and share monitoring data from similar equipment, thereby enriching the data set and facilitating collaborative analysis and optimized control across devices. Finally, all monitoring data are integrated to establish a comprehensive large database as the basis for subsequent analysis and modeling.

[0019] By combining the local monitoring database with the collaborative monitoring database of the same clan, multi-dimensional monitoring data can be obtained, providing a comprehensive data foundation for subsequent data analysis and model building, and improving the accuracy and reliability of subsequent data analysis.

[0020] The relationship fitting unit 20 is used to extract key features of absorption control according to the local monitoring database, fit the influence relationship between the absorption evaluation results and the control parameters, and build a preliminary adaptive control model.

[0021] Specifically, the absorption evaluation results refer to an assessment of the acid mist absorber's operating performance, such as the exhaust gas purification rate and whether the acid mist concentration meets emission standards. Control parameters are those that adjust the acid mist absorber's operating status, such as the liquid flow rate adjustment value and the frequency of absorption liquid replacement. The preliminary adaptive control model, derived from data relationship fitting, adjusts the control strategy based on monitoring data to optimize the acid mist absorption effect.

[0022] The relationship fitting unit 20 extracts key features based on the local monitoring database and constructs a control model. First, by analyzing the monitoring data (such as the tail gas acid mist concentration and the pH value of the absorption liquid, etc.), the characteristic parameters that have a significant relationship with the absorption effect are determined. Then, by fitting the relationship between the absorption evaluation results and the control parameters, a preliminary adaptive control model is established. For example, by analyzing the relationship between the acid mist concentration and the pH value of the absorption liquid, it can be found that the acid mist absorption efficiency is highest within a certain specific pH value range. Using these fitting relationships, it is possible to predict how the control parameters affect the absorption effect and construct a preliminary adaptive control model. This model preliminarily establishes the connection between the acid mist absorption effect and the control parameters, providing a basic framework for subsequent model optimization.

[0023] The model optimization unit 30 is used to construct a test data set and an incremental learning set based on the same family collaborative monitoring database, use the incremental learning set to perform incremental learning on the preliminary adaptive control model, and use the test data set to verify the preliminary adaptive control model until the convergence target is met, thereby obtaining an adaptive control model.

[0024] Specifically, the incremental learning set is a dataset used for incremental learning of the preliminary adaptive control model, helping it adapt to new data and environmental changes. The test dataset is a dataset used to verify the performance of the preliminary adaptive control model, ensuring its effectiveness in practical applications.

[0025] Data is extracted from the same family collaborative monitoring database to construct a test data set and an incremental learning set. For example, monitoring data under operating conditions not covered in the local monitoring database are extracted from the same family collaborative monitoring database to generate an incremental learning set. At the same time, a small amount of monitoring data is selected under various operating conditions to construct a test data set. The preliminary adaptive control model is continuously trained and updated using the incremental learning set to enhance the model's adaptability to environmental changes, and the test data set is used to verify whether the preliminary adaptive control model after incremental learning meets the expected performance, especially its stability under normal and abnormal conditions. For example, if the preliminary adaptive control model cannot stably control the acid mist concentration under certain special conditions, the test data set can help identify the problem and adjust the model until the convergence target is met. This convergence target is set so that the performance of the model (such as prediction error) reaches a predetermined standard. Through continuous learning and verification, the preliminary adaptive control model gradually enhances its adaptability to various operating environments and operating conditions, and ultimately obtains a high-precision adaptive control model.

[0026] A test data set and incremental learning set were constructed to optimize the preliminary adaptive control model, enabling the model to better cope with various complex situations in the actual operation of the acid mist absorption tower, improving the accuracy and stability of the model, and providing a more reliable model foundation for achieving precise control.

[0027] The model deployment unit 40 is used to disassemble the input, operation, and output logical relationship of the adaptive control model, and deploy a modular control system including an input module, a logic operation module, and an output module. The input module is used to connect the acquisition path. The logic operation module has the adaptive control model built in to perform control operation analysis on real-time monitoring data. The output module is used to feed back the control parameters output by the logic operation module to the control center for parameter control.

[0028] Specifically, the modular control system is a system composed of an input module, a logic operation module and an output module, which is used to realize the actual control function of the adaptive control model. The model deployment unit 40 disassembles the optimized adaptive control model into input, operation and output modules. These modules form a complete control system that can analyze and process real-time monitoring data, generate corresponding control strategies, and control and adjust the acid mist absorption tower. In the modular control system, the input module is responsible for receiving data collected from sensors and monitoring equipment, and transmitting this data to the control system. The logic operation module executes the core control algorithm and analyzes and calculates the real-time data according to the adaptive control model. The output module feeds back the control parameters processed by the logic operation module to the control center or equipment for actual operation adjustments to ensure that the acid mist absorption tower is always in the optimal operating state under different working conditions.

[0029] The modular design enables efficient operation and flexibility of the control system, which can quickly respond to real-time data changes of the acid mist operation tower, ensuring the stable operation and efficient performance of the acid mist absorption tower.

[0030] Further, such as Figure 2 As shown, the operation data collection unit 10 is further configured to perform the following steps:

[0031] Step P11: Establishing a node mapping relationship between the gas flow rate, liquid flow rate, pressure, temperature, tail gas acid mist concentration, absorption liquid pH value and absorption process.

[0032] Step P12: Obtain the node state timing difference of the absorption process, and time-align the gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and absorption liquid pH value based on the node state timing difference to build a local monitoring database.

[0033] Step P13: Obtain the basic parameters and constraint parameters of the local absorption process. The basic parameters are the equipment parameters and control parameter range of the acid mist absorption tower, and the constraint parameters are the local absorption acid mist type, absorption liquid parameters, monitoring parameters, and emission evaluation standards.

[0034] Step P14: configuring the same family screening conditions according to the basic parameters and constraint parameters, searching the shared database through the same family screening conditions, and obtaining the same family collaborative monitoring database.

[0035] Specifically, the node mapping relationship refers to the correspondence between the monitoring parameters such as gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and pH value of the absorption liquid during the operation of the acid mist absorption tower and each link (node) in the absorption process. By analyzing the process flow of the acid mist absorption tower, for each operating node in the absorption tower, the monitoring parameters corresponding to the node are determined. For example, when the acid mist enters the absorption tower, the size of the gas flow will affect the distribution of the acid mist in the tower, thereby establishing a mapping relationship between the gas flow and the inlet node of the acid mist absorption tower; similarly, for the pH value of the absorption liquid, at the node where the absorption liquid reacts with the acid mist, its value directly determines the efficiency of the reaction, and then establishes a corresponding mapping relationship. By traversing each operating node of the acid mist absorption tower and determining the monitoring parameters corresponding to each node, multiple node mapping relationships are established.

[0036] Due to the varying operating speeds or response times of different nodes, the state data of different nodes differs in chronological order, known as node state timing differences. Based on these node state timing differences, time alignment is performed on collected data such as gas flow, liquid flow, pressure, temperature, exhaust acid mist concentration, and absorption liquid pH. This unifies the timestamps of different nodes and monitoring data, ensuring temporal consistency across all data. For example, during the operation of an acid mist absorption tower, the acid mist enters the absorption tower five minutes before absorption liquid pH monitoring begins. When constructing the local monitoring database, this five-minute time difference needs to be taken into account, and the gas flow data collected when the acid mist enters the absorption tower and the absorption liquid pH data collected five minutes later should be arranged in the correct chronological order. By performing time alignment based on node state timing differences, the data in the constructed local monitoring database is temporally coherent and consistent. This makes subsequent data analysis more accurate and reliable, avoids data misinterpretation or erroneous analysis due to time asynchrony, and ensures accurate mining of useful information from the data.

[0037] Obtain the equipment parameters of the acid mist absorption tower (such as tower height, diameter, and internal structure) and control parameter ranges (such as the range of gas flow rate during normal operation and the adjustable range of liquid flow rate). These parameters are defined as basic parameters, which form the fundamental basis for the operation and control of the acid mist absorption tower. For example, obtain the equipment parameters such as the height and diameter of the acid mist absorption tower, as well as the ranges of various control parameters during normal operation, from the equipment manufacturer's equipment manual. Then, obtain parameters such as the local absorption acid mist type (such as sulfuric acid mist, hydrochloric acid mist), absorption liquid parameters (such as the absorption liquid's composition and initial concentration), monitoring parameters (such as the specified gas flow rate and tail gas acid mist concentration to be monitored), and emission evaluation standards (such as the upper limit of tail gas acid mist emission concentration specified by the local environmental protection department). These parameters are defined as constraint parameters. These constraint parameters constrain and regulate the acid mist absorption process and control. The local absorption acid mist type, specified monitoring parameters, and emission evaluation standards can be obtained from environmental protection documents, and the absorption liquid parameters can be obtained from the absorption liquid formula. By obtaining basic parameters and constraint parameters, the characteristics of the acid mist absorption tower itself as well as its operation and environmental protection requirements are clarified, which helps to screen out similar data that are relevant and comparable to local absorption towers, thereby building a more valuable similar collaborative monitoring database.

[0038] The same family screening conditions are conditions formulated based on basic parameters and constraint parameters for screening the data of acid mist absorption towers of the same family in the shared database. Based on the basic parameters and constraint parameters obtained previously, the same family screening conditions are configured using database query statements. For example, if the equipment parameters of the local acid mist absorption tower are a diameter of 3 meters and the treatment of sulfuric acid mist, then the same family screening conditions may be the data of acid mist absorption towers with a diameter between 2.5 meters and 3.5 meters and the treatment of sulfuric acid mist. Execute the query statement to search in the shared database to obtain the same family collaborative monitoring database. By configuring the same family screening conditions and searching the shared database to obtain the same family collaborative monitoring database, the local acid mist absorption tower can draw on the data of other absorption towers of the same type, thereby enriching the data source, providing more data support for a more comprehensive analysis of the acid mist absorption process and the construction of a more accurate control model, and further improving the reliability and adaptability of the entire acid mist absorption tower intelligent control system.

[0039] Further, such as Figure 3 As shown, the relationship fitting unit 20 is further configured to perform the following steps:

[0040] Step P21: performing absorption evaluation based on the exhaust acid mist concentration in the local monitoring database and constructing a multi-level evaluation threshold.

[0041] Step P22: Based on the multi-level evaluation threshold, the local monitoring database is classified into sample levels to construct a multi-level effect evaluation data cluster.

[0042] Step P23: For the multi-level effect evaluation data cluster, the fluctuation relationship of each monitoring parameter of the gas flow, liquid flow, pressure, temperature, and absorption liquid pH value is integrated across levels to construct a cross-level fluctuation data set of a single monitoring parameter. The cross-level fluctuation data set of a single monitoring parameter is a monitoring data set at different evaluation levels in which a single monitoring parameter fluctuates while other monitoring parameters do not fluctuate or fluctuate less than a threshold.

[0043] Step P24: Based on the cross-level fluctuation data set of the single monitoring parameter, fit the influence relationship of the fluctuation of the single monitoring parameter on the evaluation level to obtain the fluctuation influence coefficient.

[0044] Step P25: performing key feature identification and extraction based on the fluctuation influence coefficient to obtain key feature parameters.

[0045] Specifically, the multi-level evaluation threshold is a set of boundary values ​​of multiple different levels based on the exhaust acid mist concentration, which is used to classify and evaluate the acid mist absorption effect. The multi-level evaluation threshold is constructed by analyzing the distribution of exhaust acid mist concentration data in the local monitoring database. Data statistical analysis methods can be used, such as drawing a histogram or frequency distribution curve of the exhaust acid mist concentration to observe the concentration trend and dispersion of the data. Taking the local monitoring database of an acid mist absorption tower as an example, after statistical analysis of a large amount of exhaust acid mist concentration data, it was found that the concentration values ​​are mainly concentrated between 0 and 5%. According to the actual absorption requirements and environmental protection standards, this range is divided into different levels. For example, the exhaust acid mist concentration below 1% is set as an excellent threshold, 1% to 3% is set as a good threshold, and 3% to % is set as a qualified threshold, thereby obtaining a multi-level evaluation threshold.

[0046] According to the constructed multi-level evaluation threshold, a data classification algorithm is used to judge each set of data in the local monitoring database, classify it into corresponding levels, and construct a multi-level effect evaluation data cluster. For example, for each set of sample data in the local monitoring database, the conditional judgment statement in the data processing software can be used to classify the samples. If the tail gas acid mist concentration is lower than 1%, it is marked as an excellent level. If it is between 1% and 3%, it is marked as a good level, and so on, thereby constructing a multi-level effect evaluation data cluster. This multi-level effect evaluation data cluster refers to grouping relevant monitoring data (gas flow, liquid flow, pressure, temperature, absorption liquid pH value, etc.) according to different evaluation levels to form multiple data sets, each of which corresponds to a different absorption effect level. Through sample classification, local monitoring data can be classified, stored and analyzed according to different absorption effect levels, providing more accurate and detailed data support for subsequent relationship fitting.

[0047] For each monitoring parameter (gas flow, liquid flow, pressure, temperature, pH value of absorption liquid), in each level of the multi-level effect evaluation data cluster, by setting the fluctuation range of other monitoring parameters (such as fluctuation less than a small threshold, such as 5%), the data of single monitoring parameter fluctuation while other parameters are relatively stable are screened out, and a cross-level fluctuation data set is constructed for each monitoring parameter. For example, at different levels such as excellent, good, and qualified, the data of gas flow fluctuation alone, liquid flow, pressure, temperature, pH value of absorption liquid and other parameters that do not fluctuate or fluctuate less than the threshold are integrated to generate a gas flow cross-level fluctuation data set. Constructing a cross-level fluctuation data set of a single monitoring parameter helps to analyze the impact of each monitoring parameter fluctuation on the acid mist absorption effect (different evaluation levels) separately. Excluding the interference of other parameters can more accurately study the independent role of each parameter in the acid mist absorption process, providing a more detailed and targeted data basis for the subsequent fitting of fluctuation influence relationships.

[0048] The fluctuation impact coefficient is a quantitative coefficient that describes the degree to which the fluctuation of a single monitoring parameter affects the evaluation grade. Based on a constructed cross-level fluctuation dataset for a single monitoring parameter, mathematical modeling methods, such as linear regression analysis (if the fluctuation relationship is approximately linear) or nonlinear fitting (if the fluctuation relationship is more complex), are used. With the single monitoring parameter as the independent variable and the corresponding absorption effect evaluation grade as the dependent variable, the influence relationship between the monitoring parameter and the evaluation grade is fitted to determine the fluctuation impact coefficient corresponding to each monitoring parameter. Taking gas flow rate as an example, the fluctuation value of the gas flow rate is used as the independent variable, and the corresponding evaluation grade change (which can be quantified, such as 3 for excellent, 2 for good, and 1 for qualified) as the dependent variable. A fitting analysis is performed to obtain a relationship model between the two. Parameters in the model, such as the correlation coefficient, can be used as the fluctuation impact coefficient. Obtaining the fluctuation impact coefficient allows quantification of the degree to which the fluctuation of a single monitoring parameter affects the evaluation grade, allowing comparison of the magnitude of the fluctuation impact of different monitoring parameters, providing a quantitative basis for identifying key characteristic parameters.

[0049] Based on the fluctuation influence coefficient, the monitoring parameters that have a significant impact on the acid mist absorption effect are identified and determined as key characteristic parameters. The key characteristic parameters can be determined using a threshold judgment method. For example, statistical analysis software is used to perform a distribution analysis of the fluctuation influence coefficient. Based on the distribution characteristics of the fluctuation influence coefficient, a threshold is determined. For example, the threshold can be set to a certain quantile of the data, such as the 95th percentile. When the fluctuation influence coefficient of a monitoring parameter meets this threshold requirement, it is identified as a key characteristic parameter.

[0050] Identifying key characteristic parameters can focus on the factors that have the greatest impact on the acid mist absorption effect. In subsequent steps such as building a control model, these key parameters can be focused on, thereby simplifying the model building process and improving the accuracy and efficiency of the model.

[0051] Furthermore, constructing a cross-level fluctuation dataset of a single monitoring parameter in step P23 further includes:

[0052] Step P231: Take each monitoring parameter as a single monitoring parameter target in turn, and configure screening logic conditions, which are used to characterize the minimum fluctuation threshold range of the single monitoring parameter and the maximum fluctuation threshold range of other monitoring parameters.

[0053] Step P232: Utilize the screening logic conditions to perform screening and matching in the effect evaluation data clusters of each evaluation level, and construct a cross-level fluctuation data set of the single monitoring parameter.

[0054] Specifically, the screening logic condition is a rule setting used to specify the minimum threshold range of a single monitoring parameter fluctuation and the maximum threshold of other monitoring parameters fluctuation when constructing a cross-level fluctuation dataset of a single monitoring parameter. The screening logic condition can be expressed as a single parameter fluctuation meeting: |Pi (t1)-P i (t2)|>ΔP thresh , other parameter fluctuations satisfy |P j (t1)-P j (t2)|>ΔP small Among them, P i (t1) is the single monitoring parameter value at time t1, which is P i (t2) is the single monitoring parameter value at time t2, ΔP thresh is the minimum threshold of fluctuation of a single monitoring parameter, P j (t1) is the value of other monitoring parameters at time t1, and P j (t2) is the value of other monitoring parameters at time t2, ΔP small is the maximum threshold of fluctuation of a single monitoring parameter, ΔP thresh Greater than ΔP small For example, when gas flow is used as a single monitoring parameter, the minimum threshold range of gas flow fluctuation ΔP is set. thresh The maximum fluctuation threshold ΔP of other monitoring parameters (such as liquid flow, pressure, temperature, and pH value of the absorption liquid) is set at 10%. small 5%, that is, the fluctuation of other monitoring parameters cannot exceed this range.

[0055] According to the configured filtering logic conditions, data is searched and filtered within the effect evaluation data clusters at each evaluation level to identify data that meets the conditions. All data that meets the conditions is then merged to obtain a cross-level fluctuation dataset for a single monitoring parameter. By using filtering logic conditions to filter and match data within the effect evaluation data clusters at each evaluation level to construct a dataset, it is possible to more accurately obtain data where a single monitoring parameter fluctuates while other parameters are relatively stable, thus constructing an accurate cross-level fluctuation dataset for a single monitoring parameter.

[0056] Furthermore, the relationship fitting unit 20 is further configured to perform the following steps:

[0057] Step P26: Establishing a control relationship between the key characteristic parameters and the control parameters.

[0058] Step P27: Fitting the characteristic parameter influence relationship between the key characteristic parameters and the absorption evaluation results according to the local monitoring database.

[0059] Step P28: Based on the key characteristic parameters, the characteristic parameter influence relationship and the control relationship are coupled to model to obtain the preliminary adaptive control model.

[0060] Specifically, control parameters refer to parameters that can be adjusted to affect the state of the acid mist absorber, such as gas flow, gas pressure, gas temperature, and spray liquid volume. Key characteristic parameters are key monitoring parameters extracted from the monitoring parameters of the acid mist absorber. Control relationships describe the interaction between key characteristic parameters and control parameters. Establishing a control relationship between key characteristic parameters and control parameters determines how to influence key characteristic parameters by adjusting control parameters, thereby optimizing the performance of the absorber.

[0061] The characteristic parameter influence relationship is a quantitative relationship that quantifies the impact of key characteristic parameters on absorption evaluation results. Based on historical data from a local monitoring database, data processing software, such as MATLAB or the NumPy and Scipy libraries in Python, is used to fit the relationship between key characteristic parameters and absorption evaluation results, determining the characteristic parameter influence relationship between the key characteristic parameters and the absorption evaluation results. Taking the absorption liquid pH value and tail gas acid mist concentration as an example, tail gas acid mist concentration data corresponding to different absorption liquid pH values ​​is obtained from the local monitoring database. An appropriate fitting method is then selected based on the characteristics of the data. If the data exhibit a linear relationship, a linear fitting method can be used to obtain an equation of the form y = mx + c (where y is the absorption evaluation result, x is the key characteristic parameter, and m and c are fitting coefficients). If the relationship is nonlinear, a nonlinear fitting method, such as polynomial fitting or exponential fitting, is used. Fitting the characteristic parameter influence relationship between key characteristic parameters and absorption evaluation results helps quantify the degree of influence of key characteristic parameters on absorption performance, providing a basis for a comprehensive understanding of the acid mist absorption process. Furthermore, this influence relationship can be incorporated into the construction of an adaptive control model, enabling the model to more accurately reflect the actual acid mist absorption situation.

[0062] The coupling modeling of the characteristic parameter influence relationship and the control relationship yields a preliminary adaptive control model. This model comprehensively considers the impact of key characteristic parameters on the absorption evaluation results and the relationship between key characteristic parameters and control parameters, reflecting the operating mechanism of the acid mist absorber more comprehensively and accurately, thereby achieving adaptive control of the acid mist absorber.

[0063] Furthermore, step P26 further includes:

[0064] Step P261: Obtain monitoring positioning of key characteristic parameters.

[0065] Step P262: Based on the gas-liquid absorption operation path of the acid mist absorption tower, obtain the control relationship between the monitoring parameters and the control parameters in the path.

[0066] Step P263: Position the gas-liquid absorption operation path according to the monitoring positioning, and obtain the control relationship of the positioning path as the control relationship between the key characteristic parameters and the control parameters.

[0067] Specifically, when establishing the control relationship between the key characteristic parameters and the control parameters, first, determine the monitoring location of the key characteristic parameters, that is, the specific position or range of the key characteristic parameters in the entire monitoring system of the acid mist absorption tower. By exchanging the monitoring system of the acid mist absorption tower, the monitoring layout design and data collection points of the acid mist absorption tower are obtained, thereby determining the monitoring location of the key characteristic parameters. For example, for a system that uses a distributed sensor network to monitor the acid mist absorption tower, the sensor number and installation position are one-to-one corresponding to the monitored data. By querying the installation position corresponding to the sensor number (such as the middle of the tower body 2 meters away from the bottom), the monitoring location of the key characteristic parameter (such as the pressure value monitored here) can be determined.

[0068] The gas-liquid absorption process path is the actual process route of the gas-liquid absorption reaction in the acid mist absorption tower. It includes the physical and chemical process paths involved in a series of processes, including the acid mist entering the tower, contacting the absorption liquid, reacting, and exhaust gas discharge. Based on the specific design and operation process of the gas-liquid absorption tower, the interaction between gas flow and absorption liquid flow is analyzed and established. Focusing on the gas-liquid contact surface, the flow path and mixing process of gas and liquid affect the absorption efficiency of acid mist. Through these path analyses, it is possible to determine which monitoring parameters have a clear correlation with the control parameters, and then establish the control relationship between the monitoring parameters and control parameters in the gas-liquid absorption process path.

[0069] The gas-liquid absorption operation path is located according to the monitoring positioning, and the gas-liquid absorption operation process involved in the specific area or stage of the acid mist absorption tower is determined. Then, from the control relationship between the monitoring parameters and the control parameters in the obtained gas-liquid absorption operation path, the control relationship related to the positioning path is screened out and determined as the control relationship between the key characteristic parameters and the control parameters. For example, if the monitoring location of the key characteristic parameters is the reaction area in the middle of the acid mist absorption tower, then the control relationship between the key characteristic parameters (such as the liquid flow rate in the middle area) and the control parameters (such as the replenishment rate of the absorption liquid in the middle area) in this middle reaction area is extracted from the control relationship of the entire gas-liquid absorption operation path.

[0070] By obtaining the control relationship between key characteristic parameters and control parameters, we can focus more on the control relationship of specific areas or stages related to the key characteristic parameters, improve the pertinence and accuracy of the control relationship, and provide a more reasonable and effective control relationship basis for the subsequent construction of an adaptive control model.

[0071] Furthermore, step P262 further includes:

[0072] Step P262-1: Determine the absorption operation path based on the structure of the acid mist absorption tower and the operation monitoring data.

[0073] Step P262-2: Fit the gas-liquid propagation path based on the gas-liquid transmission principle and gas dynamics principle.

[0074] Step P262-3: Merge the absorption operation path with the gas-liquid propagation path to obtain the gas-liquid absorption operation path.

[0075] Step P262-4: Analyze the control relationship between the monitoring parameters and the control parameters in each path node based on the gas-liquid absorption operation path.

[0076] Specifically, the control relationship between the monitoring parameters and the control parameters in the path is obtained. First, the specific path of the gas-liquid absorption operation is determined based on the structural information of the acid mist absorption tower and the monitoring data collected during the operation. The operation monitoring data includes various real-time monitoring working condition data, such as gas flow, tail gas acid mist concentration, liquid flow, temperature, etc. The flow paths of the gas and the absorption liquid can be analyzed through these data. For example, the structural drawings of the acid mist absorption tower can be used to determine the various components inside the tower body and their layout, such as the position and function of the packing layer, distributor, collector, etc. Then, according to the process flow, the flow trajectory of the gas and absorption liquid in the acid mist absorption tower, that is, the absorption operation path, is determined. Determining the absorption operation path helps to have a deeper understanding of the working principle and process of the acid mist absorption tower, and provides a basis for the subsequent gas-liquid propagation path fitting.

[0077] The gas-liquid transmission path is a theoretical path for the propagation of acid mist and absorption liquid in the tower based on the gas-liquid transmission principle and the gas dynamics principle. According to the gas-liquid transmission principle and the gas dynamics principle, a mathematical model for the propagation of acid mist and absorption liquid in the tower is established. For example, using the gas-liquid mass transfer equation N = k L a(C g -C l )(where N is the mass transfer flux, k L is the liquid phase mass transfer coefficient, a is the specific surface area, C g is the gas phase concentration, C l is the liquid concentration), and the gas kinetic equation (where ΔP is the pressure drop, f is the friction factor, L is the tower height, D is the tower diameter, ρ is the gas density, and v is the gas velocity). Combined with the physical properties of the acid mist and the absorption liquid (such as density and viscosity), a set of equations describing the gas-liquid two-phase flow is constructed. These equations are then solved by numerical simulation or analytical methods to obtain the propagation path of the gas and liquid in the tower. In actual operation, computational fluid dynamics (CFD) software, such as Fluent and CFX, can be used for numerical simulation to obtain a more accurate gas-liquid propagation path.

[0078] The absorption operation path is integrated with the gas-liquid propagation path. Calculations are performed using mathematical models and data fusion algorithms (such as the least squares method and Kalman filtering), combining the theoretical derivation of the gas-liquid flow path with actual monitoring data. For example, if the gas flow rate in a certain area of ​​the absorption tower is low and the liquid flow rate is high, this information can be combined to evaluate the contact efficiency of the gas and liquid in this area, and the gas-liquid distribution within the absorption tower can be optimized based on the simulation results of the gas-liquid propagation path. Through path fusion, a more comprehensive and accurate gas-liquid absorption operation path can be obtained.

[0079] At each node along the gas-liquid absorption process, analyze the relationship between monitoring parameters and control parameters. This can be done using data analysis tools. For example, at the acid mist inlet node, analyze the relationship between the acid mist flow rate and the initial absorption liquid supply. At the reaction zone node, analyze the impact of parameters like temperature and pressure on the absorption reaction rate to determine the relationship between these parameters and the control parameters.

[0080] Exemplarily, based on the aforementioned path analysis, the control parameters of each node are determined as follows: Gas inlet area: flow rate, temperature, and pressure. Packing layer area: liquid-gas ratio, pH value, and gas-liquid contact time. Gas-liquid separator area: droplet removal efficiency and gas flow state. Tail gas outlet area: tail gas concentration and emission standards. There is a close relationship between the monitoring parameters of each node (such as flow rate, temperature, pH value, acid mist concentration, etc.) and the control parameters (such as liquid-gas ratio, spray liquid volume, fan power, etc.). For example, gas-liquid ratio and absorption efficiency: the gas-liquid ratio (L / G) directly affects the absorption efficiency. Flow rate and pressure drop: an increase in gas flow rate will lead to an increase in the pressure drop in the tower, which in turn affects the flow efficiency of the gas. The balance of the system needs to be maintained by adjusting the fan power or liquid flow rate. Temperature and reaction rate: an increase in temperature usually accelerates the absorption reaction of acidic components in the gas, but it may also cause the absorption liquid to fail or increase in volatility. Through data analysis tools, these control relationships are fitted according to the historical data in the local monitoring database to generate a mathematical model describing these relationships.

[0081] By analyzing the control relationship between the monitoring parameters and control parameters in each path node, precise regulation of different path nodes can be achieved, thereby improving the overall working efficiency of the acid mist absorption tower and its adaptability under dynamic working conditions.

[0082] Furthermore, the model optimization unit 30 is further configured to perform the following steps:

[0083] Step P31: performing a difference comparison between the local monitoring database and the same-family collaborative monitoring database to determine a difference monitoring data sample.

[0084] Step P32: Obtain the sample evaluation types of the local monitoring database, perform quantitative distribution on the sample evaluation types, obtain sample evaluation types whose distribution is less than the mean, and generate scarce sample features.

[0085] Step P33: Using the scarce sample features as an index, search in the same family collaborative monitoring database to obtain compensation data samples.

[0086] Step P34: Construct the incremental learning set based on the difference monitoring data samples and the compensation data samples.

[0087] Step P35: extracting multiple operating condition samples from the same family collaborative monitoring database, including normal operating conditions, abnormal operating conditions, and boundary condition conditions, and constructing the test data set based on the multiple operating condition samples.

[0088] Specifically, a data comparison algorithm is used to compare data in the local monitoring database with that in the collaborative monitoring database of the same family, identifying differences between the two and determining discrepant monitoring data samples. For example, in a database management system, SQL queries can be written to compare data records in the same table structure in the two databases. Alternatively, data mining or data analysis software can be used to read the data from the two databases into data structures and then compare the corresponding fields in each row of data to determine differences.

[0089] The sample evaluation type is the result of classifying the monitoring data in the local monitoring database, such as different operating states, working conditions or performance. Scarce sample features refer to the sub-features of monitoring data that account for a relatively small proportion in the sample classification, such as certain special working conditions or special operating states. The samples in the local monitoring database are classified and marked according to the pre-set evaluation criteria to obtain multiple sample evaluation types. The number of each sample evaluation type is then counted and its mean is calculated. Next, the sample evaluation types whose number distribution is less than the mean are found. For these sample evaluation types, their data characteristics, such as the range of data values, the distribution form of data, etc., can be further analyzed to generate scarce sample features. For example, the samples in the local monitoring database of operating temperature can be classified to obtain sample data at different operating temperatures, and the total number of sample data can be divided by the number of classified sample categories (such as divided into 5 groups), and the mean of the sample data is calculated. The sample categories whose sample number is less than the sample data mean are screened out, and the operating temperatures of these sample categories are extracted as scarce sample features.

[0090] Using the generated scarce sample features as an index, a search is performed in the collaborative monitoring database of the same family to find monitoring data with similar features to the scarce sample, which is used as a compensatory data sample. By finding compensatory data samples, the system makes up for the lack of scarce samples in the local database, improves the coverage of training data, and thus reduces the model bias caused by missing data. For example, if the scarce sample feature is operating data under high temperature (greater than 50°C), operating data under high temperature (greater than 50°C) will be searched from the collaborative monitoring database of the same family to supplement it.

[0091] The difference monitoring data samples and compensation data samples are merged together to construct an incremental learning set. The incremental learning set can provide more diverse data for the preliminary adaptive control model, enabling the preliminary adaptive control model to learn more features and patterns, thereby improving the performance of the model.

[0092] Multi-condition samples are sample data extracted from the same family collaborative monitoring database, covering different working conditions such as normal working conditions, abnormal working conditions, and boundary condition working conditions. Among them, boundary conditions can be extremely high or extremely low liquid-gas ratios, etc.; abnormal working conditions can be sudden events such as blockages and equipment failures. First, sample data under different working conditions such as normal working conditions, abnormal working conditions, and boundary condition working conditions are identified and extracted from the same family collaborative monitoring database. For example, for normal working conditions, it can be judged based on indicators such as the value range and change trend of the data; for abnormal working conditions, it can be judged based on whether it exceeds the normal range, whether there is a sudden change, etc.; for boundary condition working conditions, it can be judged based on whether it is close to the limit value, etc. These multi-condition samples are then combined to construct a test data set. The multi-condition test data set helps to comprehensively evaluate the performance of the preliminary adaptive control model under different working conditions, thereby discovering the advantages and disadvantages of the model and further optimizing the model.

[0093] In summary, the intelligent control system for acid mist absorption tower based on big data provided by the embodiments of the present application has the following technical effects:

[0094] The embodiment of the present application realizes the intelligent control of the acid mist absorption tower by constructing and optimizing an adaptive control model based on big data and deploying it as a modular control system. First, real-time data is collected through high-precision sensors and monitoring equipment, and a local and collaborative database containing multi-dimensional monitoring data is constructed to ensure comprehensive and accurate data support. Then, the relationship fitting unit 20 is used to conduct an in-depth analysis of the data, and through multi-level effect evaluation and cross-level integration of fluctuation relationships, key characteristic parameters are extracted, and a precise control relationship between them and the control parameters is established, thereby constructing a preliminary adaptive control model. On this basis, the model optimization unit 30 continuously optimizes the control model through incremental learning and the construction of multiple working condition samples, ensuring that the adaptive control model can adjust the control parameters in real time according to different operating states and environmental changes and ensure the stable operation of the acid mist absorption tower. Finally, a modular design is adopted to deploy the control system to ensure that the control system is flexible and efficient under different operating environments.

[0095] Overall, the embodiment of the present application realizes intelligent control of the acid mist absorption tower through the combination of this series of technical means, significantly improves the control accuracy and adaptability to various operating conditions, can more accurately control the acid mist absorption process, effectively reduce exhaust acid mist emissions, ensure the stable operation and purification effect of the acid mist absorption tower, and improve environmental benefits.

[0096] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent control system for acid mist absorption tower based on big data, characterized in that: include: Operation data collection unit, used to collect the operation data of the acid mist absorption tower through sensors and monitoring equipment, including gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and absorption liquid pH value, and build a large database, including a local monitoring database and a homologous collaborative monitoring database; a relationship fitting unit, configured to extract key features of absorption control based on the local monitoring database, fit the influence relationship between absorption evaluation results and control parameters, and construct a preliminary adaptive control model; a model optimization unit, configured to construct a test data set and an incremental learning set based on the same family collaborative monitoring database, perform incremental learning on the preliminary adaptive control model using the incremental learning set, and verify the preliminary adaptive control model using the test data set until a convergence target is met, thereby obtaining an adaptive control model; A model deployment unit is used to disassemble the input, operation, and output logical relationships of the adaptive control model, and deploy a modular control system including an input module, a logic operation module, and an output module. The input module is used to connect to an acquisition path. The logic operation module has the adaptive control model built in to perform control operation analysis on real-time monitoring data. The output module is used to feed back the control parameters output by the logic operation module to a control center for parameter control. The execution steps of the model optimization unit include: Comparing the local monitoring database with the same-family collaborative monitoring database to determine a difference monitoring data sample; Obtaining sample evaluation types from the local monitoring database, performing quantitative distribution on the sample evaluation types, obtaining sample evaluation types with distributions smaller than the mean, and generating scarce sample features; Using the scarce sample features as an index, searching the same-family collaborative monitoring database to obtain a compensation data sample; Constructing the incremental learning set according to the difference monitoring data samples and the compensation data samples; Extract multiple working condition samples from the same family collaborative monitoring database, including normal working conditions, abnormal working conditions, and boundary condition working conditions, and construct the test data set based on the multiple working condition samples.

2. The acid mist absorption tower intelligent control system based on big data according to claim 1, characterized in that, The execution steps of the operation data collection unit include: Establishing a node mapping relationship between the gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, absorption liquid pH value and absorption process; Obtaining the node state timing difference of the absorption process, and time-aligning the gas flow, liquid flow, pressure, temperature, tail gas acid mist concentration, and absorption liquid pH value based on the node state timing difference to build a local monitoring database; Obtaining basic parameters and constraint parameters of the local absorption process, wherein the basic parameters are the equipment parameters and control parameter ranges of the acid mist absorption tower, and the constraint parameters are the local absorption acid mist type, absorption liquid parameters, monitoring parameters, and emission evaluation standards; The same family screening conditions are configured according to the basic parameters and constraint parameters, and a shared database search is performed using the same family screening conditions to obtain the same family collaborative monitoring database.

3. The intelligent control system for acid mist absorption tower based on big data according to claim 1, characterized in that, The execution steps of the relationship fitting unit include: Performing absorption evaluation based on the tail gas acid mist concentration in the local monitoring database and constructing a multi-level evaluation threshold; Classifying samples of the local monitoring database based on the multi-level evaluation threshold to construct a multi-level effect evaluation data cluster; For the multi-level effect evaluation data cluster, the fluctuation relationship of each monitoring parameter of the gas flow, liquid flow, pressure, temperature, and absorption liquid pH value is integrated across levels to construct a cross-level fluctuation data set of a single monitoring parameter, wherein the cross-level fluctuation data set of a single monitoring parameter is a monitoring data set at different evaluation levels in which a single monitoring parameter fluctuates while other monitoring parameters do not fluctuate or fluctuate less than a threshold value; According to the cross-level fluctuation data set of the single monitoring parameter, fitting the influence relationship of the fluctuation of the single monitoring parameter on the evaluation level to obtain the fluctuation influence coefficient; Key feature identification and extraction are performed based on the fluctuation influence coefficient to obtain key feature parameters.

4. The intelligent control system for acid mist absorption tower based on big data according to claim 3, characterized in that, The execution steps of the relationship fitting unit include: Taking each monitoring parameter as a single monitoring parameter target in turn, configuring a screening logic condition, wherein the screening logic condition is used to characterize the minimum fluctuation threshold range of the single monitoring parameter and the maximum fluctuation threshold of other monitoring parameters; The screening logic conditions are used to perform screening and matching in the effect evaluation data clusters of each evaluation level to construct a cross-level fluctuation data set of the single monitoring parameter.

5. The intelligent control system for acid mist absorption tower based on big data according to claim 3, characterized in that, The execution steps of the relationship fitting unit include: Establishing a control relationship between the key characteristic parameters and the control parameters; Fitting the characteristic parameter influence relationship between the key characteristic parameters and the absorption evaluation results according to the local monitoring database; Based on the key characteristic parameters, the characteristic parameter influence relationship and the control relationship are coupled and modeled to obtain the preliminary adaptive control model.

6. The intelligent control system for acid mist absorption tower based on big data according to claim 5, characterized in that: The execution steps of the relationship fitting unit include: Obtain monitoring positioning of key characteristic parameters; Based on the gas-liquid absorption operation path of the acid mist absorption tower, the control relationship between the monitoring parameters and the control parameters in the path is obtained; The gas-liquid absorption operation path is positioned according to the monitoring positioning, and a control relationship of the positioning path is obtained as the control relationship between the key characteristic parameters and the control parameters.

7. The intelligent control system for acid mist absorption tower based on big data according to claim 6, characterized in that: The execution steps of the relationship fitting unit include: Determining an absorption operation path based on the structure of the acid mist absorption tower and operation monitoring data; Fit the gas-liquid propagation path based on the gas-liquid transmission principle and gas dynamics principle; Merging the absorption operation path with the gas-liquid propagation path to obtain the gas-liquid absorption operation path; The control relationship between the monitoring parameters and the control parameters in each path node is analyzed based on the gas-liquid absorption operation path.

Images