System and method for treating waste gas in chemical product production process

Abstract

The present application relates to the technical field of control system, and specifically relates to a waste gas treatment system and method in the production process of chemical products, which comprises a three-dimensional analysis unit, a waste gas diffusion surface and a diffusion direction of a target device are determined; a modeling prediction unit, a prediction model of waste gas flow is constructed, and waste gas emission data of the target device under different operating parameters is predicted based on the prediction model of waste gas flow; a regulation and control processing unit, each air inlet corresponds to a management area; waste gas emission data is determined based on the operating parameters of each target device in the current target workshop, the spatial form of waste gas discharged by the waste gas diffusion surface is simplified into a truncated cone shape, the parameter value of the truncated cone is determined based on the waste gas diffusion velocity and the area corresponding to each waste gas diffusion surface and recorded as a spatial diffusion model, and the projection area of the spatial diffusion model on the horizontal plane is recorded as a waste gas distribution area; the air inlet corresponding to the management area is opened based on the waste gas distribution area.

Description

Technical Field

[0001] This invention relates to the field of control system technology, and specifically to a waste gas treatment system and method in the production process of chemical products. Background Technology

[0002] Waste gases generated during chemical production processes are typically collected and treated using fume hoods, exhaust ducts, and centralized treatment devices. However, in actual production, the emission patterns of waste gases are quite complex. On the one hand, different production equipment often has multiple waste gas vents, and the emission direction, velocity, and intensity of each vent vary. On the other hand, waste gas emissions also fluctuate dynamically with changes in process parameters such as feed rate, reaction temperature, pressure, and stirring conditions, resulting in a significant non-uniform distribution of waste gas within the workshop.

[0003] For the above scenarios, existing collection systems typically fix the intake port based on equipment placement and empirical parameters, and exhaust the waste gas through uniform airflow or simple linkage control. This approach is insufficient to effectively cover areas with high concentrations of waste gas, resulting in poor collection efficiency. Furthermore, although sensors can be installed at the equipment outlet or in the pipeline to obtain gas flow rate or concentration data and adjust the suction distribution, tar, dust, or crystalline substances often contained in chemical waste gas tend to deposit on the sensor surface, leading to decreased measurement accuracy or even equipment failure. This affects the reliability of the monitoring data and still impacts the waste gas collection efficiency. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a waste gas treatment system and method in the production process of chemical products, which can effectively solve the problem that it is difficult to dynamically predict the generation and diffusion process of waste gas and actively control the intake port in the existing technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a waste gas treatment system and method for chemical product manufacturing processes, comprising at least: The three-dimensional analysis unit records the production equipment that generates waste gas during the execution of production tasks as target equipment, obtains the three-dimensional model of the target equipment, analyzes the three-dimensional model of the target equipment, determines the waste gas emission surface and emission direction of the target equipment, and determines each target equipment in the three-dimensional coordinate system. The modeling and prediction unit records the equipment operating parameters and exhaust gas emission data of the target equipment during the historical production process, analyzes the historical data and constructs a prediction model for exhaust gas flow, and predicts the exhaust gas emission data of the target equipment under different operating parameters based on the prediction model for exhaust gas flow. The control and processing unit divides the target workshop into multiple management areas based on the distribution of the air intakes, with each air intake corresponding to one management area; Based on the operating parameters of each target device in the current target workshop, and combined with the prediction model of exhaust gas flow, the corresponding exhaust gas emission data for each exhaust gas emission surface are determined. The spatial form of exhaust gas discharged from the exhaust gas escaping surface is simplified to a truncated cone shape. The parameter values ​​of the truncated cone are determined based on the exhaust gas escaping velocity and area corresponding to each exhaust gas escaping surface and recorded as a spatial diffusion model. The projection area of ​​the spatial diffusion model on the horizontal plane is recorded as the exhaust gas distribution area. When any management area overlaps with the exhaust gas distribution area, the corresponding air intake of that management area is opened.

[0006] Furthermore, the process for determining the exhaust gas dispersion surface and dispersion direction of the target equipment is as follows: Obtain a 3D model of the target device and mark each exhaust gas outlet in it. Record the exhaust gas outlet as the gas flow port. Draw a virtual space model corresponding to each gas flow port. Record the 3D model of the target device as the solid space model. Record the multiple edges on the solid space model that coincide with the virtual space model as the overlapping edges. Select multiple overlapping edges to form a closed contour as the contact surface contour. The contact surface contour corresponds to the curved surface that contacts the gas flow port with the external space. Multiple planes are drawn through any three points on the contact surface contour and are called undetermined planes. The projection length of the contact surface contour onto each undetermined plane is recorded as the projection parameter of the corresponding undetermined plane. The undetermined plane corresponding to the maximum projection parameter is selected and recorded as the air contact surface corresponding to the contact surface contour. This air contact surface is the air contact surface of the gas flow port. The direction of dispersion is indicated as perpendicular to the exhaust gas dispersion surface and pointing in the direction of airflow.

[0007] Furthermore, the contact surface profile satisfies the following conditions: Condition 1: The outline of the contact surface intersects with the external space; Condition 2: One side of the contact surface outline is a virtual space model, and the other side is the external space.

[0008] Furthermore, the process for predicting exhaust gas emission data is as follows: The system has a preset monitoring cycle. Multiple equipment operating parameters that affect the exhaust emissions of the target equipment are recorded as emission impact data, and the airflow velocity of the target equipment within the exhaust gas escaping surface is recorded as target emission data. The emission impact data and target emission data of the target equipment are recorded periodically within the monitoring cycle, and the correlation function between multiple target emission data at the same time is analyzed. Using any exhaust gas escape surface as the target prediction surface, a target emission data prediction model corresponding to the target prediction surface is constructed and denoted as the exhaust gas prediction model. The dependent variable of the exhaust gas prediction model is the emission impact data. Based on the current emission impact data and the exhaust gas prediction model, the predicted values ​​of the corresponding target emission data are generated. Based on the known target emission data and the correlation function, other target emission data are predicted.

[0009] Furthermore, the process of constructing the exhaust gas prediction model is as follows: Preprocess emission impact data and target emission data to obtain a continuous and stable time series sample set; Calculate the correlation coefficient between each emission impact data and the target emission data, and select emission impact data with an absolute value of correlation coefficient greater than a preset threshold as initial features; based on the time-series characteristics of the chemical production process, introduce the time-series lag term of each initial feature and the sliding window statistical feature, and combine the selected initial features, time-series lag term and sliding window statistical feature to form the model input feature set; Using the model input feature set as the independent variable and the target emission data as the dependent variable, multiple basic prediction models are constructed using regression algorithms. The sample set is divided into a training set and a test set in chronological order. The basic prediction model is trained using the training set, so that the basic prediction model learns the mapping relationship between input features and target emission data.

[0010] The trained basic prediction model is validated using a test set. The root mean square error and mean absolute error between the predicted and measured values ​​are calculated. The accuracy evaluation value of the basic prediction model is obtained by weighted calculation. Based on the accuracy evaluation value and the difference between the predicted and measured values, the exhaust gas prediction model is selected.

[0011] Furthermore, the model selection criteria include: Condition A: The accuracy assessment value is greater than or equal to the preset accuracy threshold; Condition B: All predicted values ​​are greater than or equal to their corresponding measured values.

[0012] Furthermore, the spatial diffusion model is determined as follows: The exhaust gas velocity at the current moment corresponding to the exhaust gas escaping surface is recorded as the initial velocity. The escaping fitting circle of the exhaust gas escaping surface is drawn. The escaping fitting circle is the smallest circle including the exhaust gas escaping surface. The terminal velocity is preset. A truncated cone is constructed with the escaping fitting circle as the upper base. The height of the truncated cone is calculated based on the initial velocity, the terminal velocity, and the diameter of the escaping fitting circle. The diameter of the lower base of the truncated cone is calculated based on the preset extension half angle and the height of the truncated cone. The truncated cone is recorded as a spatial diffusion model.

[0013] Furthermore, the product of the height of the truncated cone and the terminal velocity is proportional to the product of the diameter of the diffusion fitting circle and the initial velocity, with the proportionality coefficient being a preset jet attenuation constant.

[0014] The waste gas treatment method in the chemical product production process includes the following steps: Step 1: Mark each exhaust gas outlet based on the 3D model of the target equipment, and construct an exhaust gas generation model in a spatial coordinate system that includes the exhaust gas emission surface and its emission direction, in combination with the distribution location of the target equipment in the target workshop. Step 2: Simultaneously collect emission impact data and target emission data during the monitoring period to build a basic prediction model; verify and select the exhaust gas prediction model through the test set, and combine the correlation function between the target emission data of each fugitive surface to realize the prediction of emission data of all fugitive surfaces; Step 3: Based on the exhaust gas emission velocity output by the exhaust gas prediction model as the initial velocity, construct a spatial diffusion model to simulate the distribution pattern of the initial momentum-dominant part of the exhaust gas according to the ratio of the initial velocity to the preset terminal velocity and the jet attenuation constant. Step 4: Divide the target workshop into multiple management areas based on the distribution of air intakes. Determine the waste gas distribution area based on the current operating parameters of the target equipment and the waste gas prediction model. When any management area overlaps with the waste gas distribution area, control the air intake corresponding to that management area to open.

[0015] Furthermore, the inhalation control process is as follows: A basic adjustment cycle is preset. The predicted exhaust gas velocity in the past basic adjustment cycle is continuously obtained. The maximum value is taken as the reference exhaust velocity in the next basic adjustment cycle. A spatial diffusion model is constructed based on the reference exhaust velocity and is called the reference diffusion model. All reference diffusion models are represented in a spatial coordinate system. The total volume of the reference diffusion model in each management area is recorded as the absorption index in that management area. The air intake has multiple preset opening positions, and each opening position corresponds to an absorption range. The opening level of each inhaler is adjusted based on the absorption index until the next adjustment is needed.

[0016] The technical solution provided by this invention has the following advantages compared with the known prior art: 1. This invention constructs a waste gas generation model based on a three-dimensional model of the target equipment, identification of the waste gas escape surface, and the actual layout of the workshop. It also establishes a waste gas emission prediction model by combining historical operating parameters. This enables visualization, parameterization, and dynamic prediction of the location, escape direction, and emission trend of waste gas generation in chemical workshops. Compared with the existing technology that relies solely on experience or single-point monitoring, this invention overcomes the shortcomings of coarse representation of waste gas sources and difficulty in timely and accurate grasp of emission status. It solves the problem of difficulty in finely modeling and predicting the waste gas generation process.

[0017] 2. This invention matches the predicted waste gas distribution area with the management area corresponding to the intake port, and controls the relevant intake ports to open as needed, realizing the zoned linkage and directional collection of intake ports. Compared with the existing technology of fixed arrangement of intake ports, uniform extraction or blind opening, it avoids the problems of mismatched waste gas collection range, local waste gas escape and high energy consumption, and solves the problem that the waste gas treatment system is difficult to efficiently control based on the actual distribution of waste gas. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0019] Figure 1 This is an overall module block diagram of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] The present invention will be further described below with reference to embodiments.

[0022] See Figure 1 The waste gas treatment system in the chemical product production process shall include at least: The three-dimensional analysis unit designates the chemical product manufacturing and processing workshop that requires waste gas treatment as the target workshop, and the production equipment that generates waste gas during the execution of production tasks as the target equipment. It obtains a three-dimensional model of the target equipment (based on the equipment construction drawings provided by the equipment manufacturer and constructed using three-dimensional modeling software). Based on the distribution of the target equipment in the target workshop and the corresponding three-dimensional model, it constructs a waste gas generation model within the target workshop.

[0023] It should be noted that waste gas in the production process of chemical products refers to a mixed fluid generated in the entire chemical production process, existing in gaseous form, containing one or more chemical substances that exceed environmental tolerance, pose potential hazards to human health or the ecological environment, no longer have the original technological value, and must be captured, purified and controlled by engineering means.

[0024] Compared to the limitations of existing technologies that only focus on pollution sources, the exhaust gas generation model integrates the three-dimensional model of the target equipment with its actual layout. The exhaust gas generation model can be used to accurately simulate the real direction and distribution of exhaust gas in the workshop, providing a scientific basis for subsequent optimization of the intake port location and parameter adjustment.

[0025] Specifically, the process of constructing the waste gas generation model is as follows: The three-dimensional model of the target equipment is analyzed to determine the exhaust gas escaping surface and escaping direction of the target equipment. A three-dimensional coordinate system is constructed with the center point of the target workshop as the origin, which is called the spatial coordinate system. The three-dimensional model and exhaust gas escaping surface of each target equipment are drawn in the spatial coordinate system, and the escaping direction corresponding to the exhaust gas escaping surface is indicated.

[0026] By constructing a waste gas generation model, it is possible to simulate and predict the waste gas generation during the production and processing process in the target workshop, transforming the originally abstract waste gas diffusion problem into a visualized spatial physical model, thereby enabling analysis and management from the source of waste gas generation.

[0027] More specifically, the process for determining the exhaust gas venting surface and venting direction of the target equipment is as follows: Obtain a 3D model of the target device and mark each exhaust gas outlet in it. That is, under normal working conditions, the exhaust gas generated by the target device will be discharged from each exhaust gas outlet. Perform independent analysis on each exhaust gas outlet and determine the air contact surface corresponding to each exhaust gas outlet as the exhaust gas venting surface. The venting direction is represented as perpendicular to the exhaust gas venting surface and pointing in the direction of airflow (exhaust gas).

[0028] It should be noted that the exhaust gas outlet is a virtual shape composed of multiple contour lines (as opposed to a solid; for example, if there is a trough, the trough-shaped space inside the trough is a virtual shape, and conversely, the trough itself is a solid). In other words, the exhaust gas outlet corresponds to a spatial region between the external space and the internal space, which is the space that exhaust gas must pass through to be discharged from the internal space of the target equipment into the outside air.

[0029] Furthermore, the process for determining the air contact surface is as follows: The exhaust gas outlet is designated as the gas flow port. A virtual space model is drawn for each gas flow port. The 3D model of the equipment corresponding to the gas flow port is designated as the solid space model. Multiple edges on the solid space model that coincide with the virtual space model are designated as coincident edges. Multiple coincident edges are selected to form a closed contour, which is designated as the contact surface contour. The selected contact surface contour corresponds to the curved surface where the gas flow port contacts the external space (the contact surface contour is the edge of this curved surface). The contact surface contour satisfies the following conditions: Condition 1: The outline of the contact surface intersects with the external space; Condition 2: One side of the contact surface outline is a virtual space model, and the other side is the external space.

[0030] It should be noted that, in performing gas space analysis on the equipment, this invention mainly divides it into internal space and external space. The portion of the internal space that contacts the external space is the virtual space model corresponding to the gas flow port. Gas completes the spatial transformation from inside to outside through the virtual space. Obviously, the virtual space model and the external space must have an intersection surface, allowing the gas to completely enter the outside after passing through this intersection surface. The contact surface contour obtained through the above screening process can be used to determine the intersection surface, thereby determining more relevant emission data based on this intersection surface. In another specific embodiment, the contact surface contour is determined manually by the operator, that is, the contact surface contour representing the edge of the gas outlet is drawn on the three-dimensional model of the equipment.

[0031] Multiple planes are drawn through any three points on the contact surface contour and are denoted as undetermined planes. The projection length of the contact surface contour onto each undetermined plane is obtained and denoted as the projection parameter of the corresponding undetermined plane. The undetermined plane corresponding to the maximum projection parameter is selected and denoted as the air contact surface corresponding to the contact surface contour. This air contact surface is the air contact surface of the gas flow port.

[0032] It should be noted that this process automatically defines the contact surface between the gas inlet and the external space through geometric calculations, transforming the abstract exhaust gas emission boundary into quantifiable planar geometric parameters. Through boundary analysis of the virtual space and the solid model, combined with projection extremum screening, the key interface through which gas completely enters the external space can be accurately located. This provides a unified and repeatable geometric benchmark for subsequent exhaust gas diffusion simulation, emission flux calculation, and capture efficiency analysis, thereby improving the scientific rigor and accuracy of exhaust gas system modeling.

[0033] The modeling and prediction unit records the equipment operating parameters and exhaust gas emission data during the historical production process of the target equipment, analyzes the historical data and constructs a prediction model for exhaust gas flow, and predicts the exhaust gas emission data of the target equipment under different operating parameters.

[0034] Specifically, the process for predicting exhaust gas emission data is as follows: The monitoring cycle is preset (specifically set to one day). Multiple equipment operating parameters that affect the exhaust emissions of the target equipment are recorded as emission impact data, and the airflow velocity (measured by a wind speed sensor installed at the air contact surface) of the target equipment is recorded as target emission data. The emission impact data and target emission data of the target equipment are recorded periodically within the monitoring cycle (the recording cycle is a preset value), and the correlation function between multiple target emission data at the same time is analyzed. It should be noted that since the target emission data originates from the same production equipment or the same process unit and is driven by the same equipment operating parameters (i.e. emission impact data) and internal pressure fluctuations, temperature changes, and other operating conditions, the airflow velocity at each escaping surface often exhibits homogeneity, synchronicity, or coupling in the time series. That is, when the emission velocity of one escaping surface changes due to process disturbances, the emission velocities of other escaping surfaces will also change in a regular and interconnected manner.

[0035] Using any exhaust gas escape surface as the target prediction surface, a target emission data prediction model corresponding to the target prediction surface is constructed and denoted as the exhaust gas prediction model (i.e., the prediction model of exhaust gas flow). The dependent variable of the exhaust gas prediction model is the emission impact data. Based on the current emission impact data and the exhaust gas prediction model, the predicted values ​​of the corresponding target emission data are generated. Based on the known target emission data and the correlation function, other target emission data are predicted.

[0036] It should be noted that the emission impact data is specifically determined by the staff. For example, in one specific embodiment, the emission impact data is set as the feed rate, pressure, stirring speed, and reaction temperature. By constructing an exhaust gas prediction model, the predicted target emission data (airflow velocity) can be output in real time when the current emission impact data is input. This allows for the output of exhaust gas emission data without relying on sensor equipment, and can be used for feedforward control or early warning in exhaust gas treatment systems. It is especially suitable for target equipment that cannot be monitored by sensors. For example, chemical exhaust gases often contain tar, polymer monomers, or crystalline dust (such as ammonium chloride and naphthalene). After long-term use, these substances are prone to adhere to sensor probes, electric valve cores, or the inner walls of pipes, affecting sensor readings.

[0037] More specifically, the process for determining the correlation function is as follows: By synchronously collecting target emission data at each vent surface during multiple complete production cycles, performing correlation analysis on multiple data sequences at the same time, calculating the Pearson correlation coefficient between any two vent surfaces, or using regression analysis methods (such as linear regression or multinomial regression) to fit the functional relationship between them, the mathematical mapping relationship between emission rates at each vent surface can be determined (the above methods are all existing technologies and will not be elaborated on here).

[0038] Let the target emission data be respectively , where i is the index of the target emission data, i=0,1,2,…,j, and j represents the total number of target emission data. Corresponding to the exhaust gas prediction model, there are j correlation functions representing respectively and The mathematical mapping relationship between (where i = 0, 1, 2, ..., j) is used to predict target emission data. It can predict all target emission data. .

[0039] More specifically, the process of building the exhaust gas prediction model is as follows: During the monitoring period, the collected emission impact data and target emission data are preprocessed, including outlier removal, missing value filling and smoothing filtering, to obtain a continuous and stable time series sample set. Calculate the correlation coefficient between each emission impact data and the target emission data, and select emission impact data with an absolute value of correlation coefficient greater than a preset threshold as initial features; based on the time-series characteristics of the chemical production process (considering the lag and inertia effect of the chemical production process), introduce the time-series lag term of each initial feature and the sliding window statistical feature (the time-series lag term includes historical values ​​of several recording periods before the current time, and the sliding window statistical feature includes the mean and standard deviation within the preset time window), and combine the selected initial features, time-series lag term and sliding window statistical feature to form the model input feature set; It should be noted that, since waste gas emissions during chemical production are often influenced by the cumulative effects of historical operating conditions, the emission rate at the current moment depends not only on the emission impact data at the current moment but also on the equipment status at several past moments. Therefore, it is necessary to introduce time-series lag terms into the model input features. For each selected initial feature, its historical values ​​for several recording periods prior to the current moment (i.e., lag order) are constructed as additional features. The lag order can be determined using autocorrelation function analysis or preset based on process experience. For example, if the lag order is set to 3, then for the current moment t, the values ​​of this feature at moments t-1, t-2, and t-3 are introduced.

[0040] Meanwhile, to characterize the fluctuations and trends in exhaust gas emissions, a sliding window statistical feature is introduced. For each initial feature and its lag term, its statistics are calculated within a preset time window length. These statistics include, but are not limited to, the mean and standard deviation of the data within the window. The time window length can be set according to the process characteristics. For example, if the window length is set to 10 recording cycles, then for the current time t, the mean and standard deviation of the feature value over 10 cycles from t-9 to time t are calculated and used as additional input features for the current time.

[0041] The model input feature set is used as the independent variable and the target emission data is used as the dependent variable. Multiple basic prediction models are constructed using regression algorithms. The regression algorithm is selected from one of multiple linear regression, random forest, support vector regression or long short-term memory network. Furthermore, the sample set is divided into a training set and a test set in chronological order. The basic prediction model is trained using the training set, and the model hyperparameters are adjusted through cross-validation so that the basic prediction model learns the mapping relationship between input features and target emission data.

[0042] The trained basic prediction model is validated using a test set. The root mean square error and mean absolute error between the predicted and measured values ​​are calculated. The accuracy evaluation value of the basic prediction model is obtained by weighted calculation. Based on the accuracy evaluation value and the difference between the predicted and measured values, the exhaust gas prediction model is selected.

[0043] It should be noted that the accuracy assessment value is inversely proportional to the error value.

[0044] More specifically, the model selection criteria include: Condition A: The accuracy assessment value is greater than or equal to the preset accuracy threshold; Condition B: All predicted values ​​(corresponding to the test set) are greater than or equal to their corresponding measured values.

[0045] The control and processing unit divides the target workshop into multiple management areas based on the distribution of the air intakes. Each air intake corresponds to one management area. Based on the operating parameters of each target device in the current target workshop, and combined with the prediction model of the exhaust gas flow, the exhaust gas distribution area is determined. When any management area overlaps with the exhaust gas distribution area, the air intake corresponding to that management area is controlled to open.

[0046] Specifically, the process for determining the waste gas distribution area is as follows: The exhaust gas flow prediction model determines the exhaust gas velocity (i.e., target emission data) corresponding to each exhaust gas escaping surface at the current moment. A simplified escaping model is preset, which simplifies the spatial shape of the exhaust gas discharged from the exhaust gas escaping surface into a frustum cone shape to simulate the distribution shape of the exhaust gas when it is discharged from the exhaust gas escaping surface. Based on the simplified escaping model, combined with the exhaust gas escaping velocity and area corresponding to each exhaust gas escaping surface, the spatial diffusion model corresponding to each horizontal escaping region is determined. The projection area of ​​the spatial diffusion model on the horizontal plane is recorded as the exhaust gas distribution region.

[0047] It should be noted that the process of constructing the simplified escape model is as follows: During the continuous gas jet process, the gas is divided into a "free diffusion part" and an "initial momentum-dominated part". The free diffusion part refers to the part that moves randomly under the influence of molecular diffusion and environmental turbulence, while the initial momentum-dominated part refers to the part that moves with a clear direction under the control of the outlet kinetic energy. The distribution of the initial momentum-dominated part can be effectively tracked based on data (under the assumptions of a fixed outlet, infinite space, isothermal, and incompressible fluid, it has self-modeling properties, that is, the dimensionless velocity distribution curves on different cross sections are similar). Imagine a jet composed of countless thin disks perpendicular to the axis. By limiting the boundary velocity of the thin disks to give them clear boundaries, the distribution of the initial momentum-dominant part approximates a rotating body with a wide bottom and a narrow top formed by a rotating parabola. Furthermore, the rotating body is simplified into a truncated cone, thus obtaining a rough representation of the distribution of the initial momentum-dominant part. It is assumed that inside the boundary, the fluid is mainly dominated by momentum and undergoes directional motion; outside the boundary, free diffusion is dominant. The expansion angle of the truncated cone (usually 10° to 16°) is a fixed value, the upper base of the truncated cone depends on the exhaust gas escaping surface area, and the height of the truncated cone depends on the exhaust gas escaping velocity.

[0048] It should be noted that real jet flow involves solving the Navier-Stokes equations, which is computationally complex. In contrast, the 3D model, utilizing the geometric relationships of a trapezoid (such as the upper base and expansion angle), can quickly estimate the distance and coverage area that the exhaust gas can reach, providing an engineering-acceptable approximate solution. By simplifying the model, the main gas movement path dominated by outlet momentum can be clearly identified. In exhaust gas treatment, this means that to "suck away" the exhaust gas, the optimal location of the intake port should cover this conical region as much as possible. Conversely, if the intake port is far from this region, it will mainly draw in surrounding air (the free diffusion portion), resulting in very low efficiency in capturing high-concentration exhaust gas dominated by momentum.

[0049] More specifically, the process of determining the spatial diffusion model is as follows: The exhaust gas velocity (i.e., the target emission data) corresponding to the exhaust gas escape surface at the current moment is recorded as the initial velocity. The escape fitting circle of the exhaust gas escape surface is drawn. The escape fitting circle is the smallest circle including the exhaust gas escape surface. The terminal velocity is preset (representing the velocity at the far end of the spatial diffusion model from the exhaust gas escape surface). A frustum of a cone is constructed with the escape fitting circle as the upper base and recorded as the spatial diffusion model. The height of the spatial diffusion model is calculated based on the initial velocity, the terminal velocity, and the diameter of the escape fitting circle. The diameter of the lower base of the spatial diffusion model is calculated based on the preset extension half angle (half of the extension angle) and the height of the spatial diffusion model (Pythagorean theorem).

[0050] Among them, the high-precision calculation formula of the spatial diffusion model is: , Indicates the initial flow velocity. Indicates the terminal flow velocity. The diameter of the effluent fitting circle is represented by K, which is a preset jet attenuation constant (in one specific embodiment, the value is between 5.0 and 7.0).

[0051] It should be noted that under the influence of the initial flow velocity, the flow velocity of the exhaust gas gradually decreases as the diffusion distance increases. When it decreases to a certain level (i.e., the terminal flow velocity), the exhaust gas can be considered to have switched to free diffusion. After that, the diffusion mode of the exhaust gas is mainly characterized by disordered diffusion to the low concentration area. The process of decreasing flow velocity before this is regular. Therefore, the ratio between the initial flow velocity and the terminal flow velocity, and the ratio between the height of the spatial diffusion model and the diameter of the escape fitting circle are proportional and can be represented by a fixed constant value (jet attenuation constant) for simulation.

[0052] By defining a spatial diffusion model, the boundaries of the waste gas distribution area can be analyzed in detail, thereby enabling control and adjustment of the relevant intake ports. This makes the waste gas treatment effect of the intake ports more targeted and improves the waste gas treatment efficiency.

[0053] The waste gas treatment method in the chemical product production process includes the following steps: Step 1: Mark each exhaust gas outlet based on the 3D model of the target equipment. Filter out the contact surface contour by analyzing the overlapping edges of the solid space model and the virtual space model. Determine the plane corresponding to the maximum projection parameter of the contact surface contour onto each undetermined plane as the air contact surface. Combine the distribution location of the target equipment in the target workshop to construct an exhaust gas generation model in a spatial coordinate system that includes the exhaust gas emission surface and its emission direction. Step 2: Simultaneously collect emission impact data and target emission data during the monitoring period. Select emission impact data with an absolute value of correlation coefficient greater than a preset threshold as initial features. Introduce time lag terms and sliding window statistical features to form the model input feature set. Use a regression algorithm to build a basic prediction model. Validate the selected exhaust gas prediction model through a test set, ensuring that the accuracy evaluation value is greater than or equal to the preset threshold and all predicted values ​​are greater than or equal to the measured values. Combine the correlation function between target emission data of each fugitive surface to achieve prediction of emission data for all fugitive surfaces. Step 3: Based on the exhaust gas escape velocity output by the exhaust gas prediction model as the initial velocity, draw the smallest circle containing the exhaust gas escape surface as the escape fitting circle. With the escape fitting circle as the upper base, the preset extension half angle as the waist slope, and the height of the frustum calculated according to the ratio of the initial velocity to the preset end velocity and the jet attenuation constant, construct a spatial diffusion model to simulate the distribution pattern of the initial momentum-dominant part of the exhaust gas. Step 4: Divide the target workshop into multiple management areas based on the distribution of air intakes. Determine the exhaust gas distribution area (i.e., the projection of the spatial diffusion model on the horizontal plane) based on the current operating parameters of the target equipment and the exhaust gas prediction model. When any management area overlaps with the exhaust gas distribution area, control the air intake corresponding to that management area to open.

[0054] Furthermore, the inhalation control process is as follows: A basic adjustment cycle is preset (5 minutes in a specific embodiment). The predicted exhaust gas velocity in the past basic adjustment cycle is continuously acquired, and the maximum value is taken as the reference velocity in the next basic adjustment cycle. A spatial diffusion model is constructed based on the reference velocity and is called the reference diffusion model. All reference diffusion models are represented in a spatial coordinate system. The total volume of the reference diffusion model in each management area is recorded as the absorption index in that management area. The air intake has multiple preset opening positions, and each opening position corresponds to an absorption range. The opening position of each inhaler is adjusted based on the absorption index, and the opening position of the inhaler is maintained for a basic adjustment cycle until the opening position is readjusted again.

[0055] By introducing a basic adjustment cycle and a reference escape velocity, the intake port control, which originally required real-time dynamic adjustment, is transformed into periodic steady-state control, ensuring the model's tolerance to operating condition fluctuations. Since the intake port position remains constant in each basic adjustment cycle, frequent adjustments to the intake port adjustment mechanism due to real-time tracking of exhaust gas emission velocity are effectively avoided. Under the premise of ensuring effective collection of exhaust gas, the stability and feasibility of the system are significantly improved, achieving a balance between control stability and collection efficiency.

[0056] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A waste gas treatment system for chemical product manufacturing processes, characterized in that, include: The three-dimensional analysis unit identifies production equipment that generates waste gas during the execution of production tasks as target equipment, acquires a three-dimensional model of the target equipment, analyzes the three-dimensional model of the target equipment, and determines the waste gas emission surface and emission direction of the target equipment. The modeling and prediction unit records the equipment operating parameters and exhaust gas emission data of the target equipment during the historical production process, analyzes the historical data and constructs a prediction model for exhaust gas flow, and predicts the exhaust gas emission data of the target equipment under different operating parameters based on the prediction model for exhaust gas flow. The control and processing unit divides the target workshop into multiple management areas based on the distribution of the air intakes, with each air intake corresponding to one management area; Based on the operating parameters of each target device in the current target workshop, and combined with the prediction model of exhaust gas flow, the corresponding exhaust gas emission data for each exhaust gas emission surface are determined. The spatial form of exhaust gas discharged from the exhaust gas escaping surface is simplified to a truncated cone shape. The parameter values ​​of the truncated cone are determined based on the exhaust gas escaping velocity and area corresponding to each exhaust gas escaping surface and recorded as a spatial diffusion model. The projection area of ​​the spatial diffusion model on the horizontal plane is recorded as the exhaust gas distribution area. When any management area overlaps with the exhaust gas distribution area, the corresponding air intake of that management area is opened.

2. The waste gas treatment system in the chemical product production process according to claim 1, characterized in that, The process for determining the exhaust gas venting surface and venting direction of the target equipment is as follows: Obtain a 3D model of the target device and mark each exhaust gas outlet in it. Record the exhaust gas outlet as the gas flow port. Draw a virtual space model corresponding to each gas flow port. Record the 3D model of the target device as the solid space model. Record the multiple edges on the solid space model that coincide with the virtual space model as the overlapping edges. Select multiple overlapping edges to form a closed contour as the contact surface contour. The contact surface contour corresponds to the curved surface that contacts the gas flow port with the external space. Multiple planes are drawn through any three points on the contact surface contour and are called undetermined planes. The projection length of the contact surface contour onto each undetermined plane is recorded as the projection parameter of the corresponding undetermined plane. The undetermined plane corresponding to the maximum projection parameter is selected and recorded as the air contact surface corresponding to the contact surface contour. This air contact surface is the air contact surface of the gas flow port. The direction of dispersion is indicated as perpendicular to the exhaust gas dispersion surface and pointing in the direction of airflow.

3. The waste gas treatment system in the chemical product production process according to claim 2, characterized in that, The contact surface profile must satisfy the following conditions: Condition 1: The outline of the contact surface intersects with the external space; Condition 2: One side of the contact surface outline is a virtual space model, and the other side is the external space.

4. The waste gas treatment system in the chemical product production process according to claim 1, characterized in that, The process for predicting exhaust gas emission data is as follows: The system has a preset monitoring cycle. Multiple equipment operating parameters that affect the exhaust emissions of the target equipment are recorded as emission impact data, and the airflow velocity of the target equipment within the exhaust gas escaping surface is recorded as target emission data. The emission impact data and target emission data of the target equipment are recorded periodically within the monitoring cycle, and the correlation function between multiple target emission data at the same time is analyzed. Using any exhaust gas escape surface as the target prediction surface, a target emission data prediction model corresponding to the target prediction surface is constructed and denoted as the exhaust gas prediction model. The dependent variable of the exhaust gas prediction model is the emission impact data. Based on the current emission impact data and the exhaust gas prediction model, the predicted values ​​of the corresponding target emission data are generated. Based on the known target emission data and the correlation function, other target emission data are predicted.

5. The waste gas treatment system in the chemical product production process according to claim 4, characterized in that, The process of building the exhaust gas prediction model is as follows: Preprocess emission impact data and target emission data to obtain a continuous and stable time series sample set; Calculate the correlation coefficient between each emission impact data and the target emission data, and select emission impact data whose absolute value of the correlation coefficient is greater than a preset threshold as initial features; Based on the time-series characteristics of chemical production processes, time-series lag terms of each initial feature and sliding window statistical features are introduced. The selected initial features, time-series lag terms, and sliding window statistical features together constitute the model input feature set. Using the model input feature set as the independent variable and the target emission data as the dependent variable, multiple basic prediction models are constructed using regression algorithms. The sample set is divided into a training set and a test set in chronological order. The basic prediction model is trained using the training set, so that the basic prediction model learns the mapping relationship between input features and target emission data. The trained basic prediction model is validated using a test set. The root mean square error and mean absolute error between the predicted and measured values ​​are calculated. The accuracy evaluation value of the basic prediction model is obtained by weighted calculation. Based on the accuracy evaluation value and the difference between the predicted and measured values, the exhaust gas prediction model is selected.

6. The waste gas treatment system in the chemical product production process according to claim 5, characterized in that, Model selection criteria include: Condition A: The accuracy assessment value is greater than or equal to the preset accuracy threshold; Condition B: All predicted values ​​are greater than or equal to their corresponding measured values.

7. The waste gas treatment system in the chemical product production process according to claim 1, characterized in that, The process of determining the spatial diffusion model is as follows: The exhaust gas velocity at the current moment corresponding to the exhaust gas escaping surface is recorded as the initial velocity. The escaping fitting circle of the exhaust gas escaping surface is drawn. The escaping fitting circle is the smallest circle including the exhaust gas escaping surface. The terminal velocity is preset. A truncated cone is constructed with the escaping fitting circle as the upper base. The height of the truncated cone is calculated based on the initial velocity, the terminal velocity, and the diameter of the escaping fitting circle. The diameter of the lower base of the truncated cone is calculated based on the preset extension half angle and the height of the truncated cone. The truncated cone is recorded as a spatial diffusion model.

8. The waste gas treatment system in the chemical product production process according to claim 7, characterized in that, The product of the height of the truncated cone and the terminal velocity is proportional to the product of the diameter of the diffusion fitting circle and the initial velocity, with the proportionality coefficient being a preset jet attenuation constant.

9. A method for treating waste gas during the production of chemical products, applied to the waste gas treatment system during the production of chemical products as described in any one of claims 1 to 8, characterized in that, Includes the following steps: Step 1: Mark each exhaust gas outlet based on the 3D model of the target equipment, and construct an exhaust gas generation model in a spatial coordinate system that includes the exhaust gas emission surface and its emission direction, in combination with the distribution location of the target equipment in the target workshop. Step 2: Simultaneously collect emission impact data and target emission data during the monitoring period to build a basic prediction model; verify and select the exhaust gas prediction model through the test set, and combine the correlation function between the target emission data of each fugitive surface to realize the prediction of emission data of all fugitive surfaces; Step 3: Based on the exhaust gas emission velocity output by the exhaust gas prediction model as the initial velocity, construct a spatial diffusion model to simulate the distribution pattern of the initial momentum-dominant part of the exhaust gas according to the ratio of the initial velocity to the preset terminal velocity and the jet attenuation constant. Step 4: Divide the target workshop into multiple management areas based on the distribution of air intakes. Determine the waste gas distribution area based on the current operating parameters of the target equipment and the waste gas prediction model. When any management area overlaps with the waste gas distribution area, control the air intake corresponding to that management area to open.

10. The waste gas treatment method in the chemical product production process according to claim 9, characterized in that, The intake port control process is as follows: A basic adjustment cycle is preset. The predicted exhaust gas velocity in the past basic adjustment cycle is continuously obtained. The maximum value is used as the reference exhaust velocity in the next basic adjustment cycle. A spatial diffusion model is constructed based on the reference exhaust velocity and is called the reference diffusion model. All reference diffusion models are represented in a spatial coordinate system. The total volume of the reference diffusion model in each management area is recorded as the absorption index in that management area. The air intake has multiple preset opening positions, and each opening position corresponds to an absorption range. Adjust the opening level of each inhaler based on the absorption index.

Images