Method for optimizing baffle in jet impinging-negative pressure reactor based on CFD-genetic algorithm

Abstract

The invention discloses a method for optimizing a baffle in a jet impinging-negative pressure reactor based on a CFD-genetic algorithm, which comprises the following steps: S1, taking the minimization of the pressure drop and uniformity of fluid in the reactor as a target, and selecting geometric structure parameters of the baffle as design variables; s2, dividing a design variable parameter range interval, and generating a sample point set covering a parameter space; s3, simulating the fluid mechanics condition of a sample point baffle structure, and obtaining physical parameters to construct a basic data set; s4, establishing an optimization target and design variable mathematical regression model containing linear, nonlinear and interaction items, and carrying out statistical test to verify the precision and reliability of the model; s5, adopting an NSGA-II algorithm to carry out iterative optimization on the baffle parameters, and obtaining a Pareto optimal solution set within a preset number of times through population initialization operation; and S6, comparing the flow field characteristic parameters before and after optimization, verifying the mass transfer efficiency of the optimized baffle through an experiment, and determining an optimal parameter combination. The flow field turbulence degree and the mass transfer efficiency in the reactor are improved, and the pressure drop and the uniformity are effectively balanced.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical equipment optimization, and in particular to a method for optimizing inner baffles of a jet impact-negative pressure reactor based on a CFD-genetic algorithm. Background Art

[0002] Jet impingement-negative pressure reactors (JI-NPRs) are widely used in gas-liquid mass transfer and separation processes in the chemical industry due to their efficient mixing properties. However, their internal flow field is easily affected by the structural parameters of the baffles, resulting in high energy dissipation and uneven fluid distribution, making it difficult to further improve mass transfer efficiency. Traditional methods that rely on discrete experimental data to adjust structural parameters consume a lot of time and resources and cannot meet the requirements of multi-objective optimization.

[0003] The design of traditional reactor baffle structures mainly relies on trial and error methods and empirical formulas. The number, position and size of the baffles are adjusted through repeated experiments, and the parameters are determined based on limited flow field test data. Although this method can improve the flow field distribution to a certain extent, it has significant defects: on the one hand, discrete experiments cannot cover the entire parameter space and are prone to missing the optimal solution; on the other hand, single-objective optimization (such as focusing only on pressure drop or uniformity) is difficult to balance the multi-objective requirements in actual working conditions, resulting in the optimized structure may cause new problems such as increased energy loss or deterioration of mixing effect. In addition, the empirical formula lacks an accurate description of complex flow fields (such as vortex structure evolution and phase mass transfer coupling), making it difficult to reveal the interaction mechanism between parameters, which limits the in-depth improvement of reactor performance.

[0004] Existing studies have attempted to introduce computational fluid dynamics (CFD) and optimization algorithms to improve design efficiency. For example, some scholars have used CFD simulation to analyze the impact of baffle structure on turbulent characteristics, and combined with the response surface method (RSM) to establish a parameter-performance mapping model, but this method relies on a large amount of sample data, and the computational cost surges when dealing with high-dimensional variables. Other studies have used genetic algorithms (GA) for multi-objective optimization, but traditional GAs are prone to falling into local optimality when dealing with complex constraints, and the non-dominated sorting efficiency is low. In addition, existing technologies often focus on a single optimization objective or simplify the flow field model, and lack a systematic analysis of the multi-scale vortex structure and the gas-liquid two-phase coupled mass transfer process in the reactor, resulting in limited applicability of the optimization results in actual working conditions, making it difficult to achieve an effective connection from simulation to engineering application. Summary of the Invention

[0005] Based on the above technical problems, the present application discloses a method for optimizing the inner baffle of a jet impact-negative pressure reactor based on CFD-genetic algorithm, comprising:

[0006] S1. Taking minimizing the pressure drop and uniformity of the fluid in the jet impact-negative pressure reactor as optimization goals, at least two geometric structural parameters of the baffle are selected as design variables;

[0007] S2. Using Latin hypercube sampling, divide the parameter range of the design variable into intervals to generate a set of sample points covering the parameter space;

[0008] S3. Using computational fluid dynamics software to perform fluid dynamics simulation on the baffle structure corresponding to each sample point, obtain physical parameter data related to the optimization target, and construct a basic data set;

[0009] S4. Establishing a mathematical regression model of the optimization objective and design variables based on the basic data set, incorporating linear terms, nonlinear terms, and interaction terms, and verifying the fitting accuracy and reliability of the model through statistical test methods;

[0010] S5. Use the non-dominated sorting genetic algorithm NSGA-Ⅱ to iteratively optimize the baffle structure parameters. Through population initialization, genetic operation and non-dominated sorting strategy, the Pareto optimal solution set is obtained within the preset number of iterations.

[0011] S6. Compare the flow field characteristic parameters in the reactor before and after optimization, verify the mass transfer efficiency of the optimized baffle structure through actual working condition experiments, and determine the optimal parameter combination.

[0012] Preferably, at least two geometric structural parameters of the baffle are selected in S1 as design variables, specifically: according to the structural characteristics of the jet impact-negative pressure reactor and the fluid mechanics optimization requirements, the radial distance of the radial installation position of the baffle from the central axis of the reactor, and the wing width of the baffle along the fluid flow direction are used as geometric parameters, and by analyzing the influence of the number of baffles on the flow field distribution in the reactor, combined with the previous research basis, the number of baffles is determined to be a reasonable number that can effectively break the eccentric flow and secondary reflux and promote the generation of vortex structure, so that the radial distance and wing width are used as adjustable design variables.

[0013] Preferably, the Latin hypercube sampling method is used in S2 to divide the parameter range of the design variables into intervals and generate a set of sample points. Specifically, for the preset parameter range of each selected baffle geometric structure parameter, the value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability, each sub-interval corresponds to a probability unit, and a sample value is randomly selected in each sub-interval. By combining the sample values ​​of each design variable, a set of sample points that are evenly distributed and non-repetitive in the parameter space is generated.

[0014] Preferably, the value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability. Specifically, for the value range of each design variable, the length of each sub-interval is determined by calculating the ratio of the total length of the variable interval to the number of preset sub-intervals, so that each sub-interval occupies an equal probability share in the entire value range, ensuring that the sample values ​​in each sub-interval have the same probability of being selected, and achieving a uniform probability distribution of the sample points in the parameter space.

[0015] Preferably, in S3, computational fluid dynamics software is used to perform fluid mechanics simulation on the baffle structure corresponding to each sample point and obtain physical parameter data, specifically: the baffle geometric structure parameters of each sample point are imported into the computational fluid dynamics software, based on the actual size and boundary conditions of the reactor, including inlet flow velocity, outlet pressure, and fluid physical properties, a turbulence model and a discrete format are used to numerically solve the gas-liquid two-phase flow field in the reactor. After the calculation is completed, the fluid pressure drop value and uniformity index corresponding to each sample point are extracted, including the velocity distribution variance and the turbulent kinetic energy distribution uniformity parameter, to construct a basic data set containing design variables and optimization objectives.

[0016] Preferably, the turbulence model and discrete format are used to numerically solve the gas-liquid two-phase flow field in the reactor and extract parameters, specifically: according to the flow characteristics of the gas-liquid two-phase flow in the reactor, the k-ε turbulence model or the k-ω turbulence model is selected to describe the turbulent pulsation effect, and the control equation group is established through the continuity equation, the momentum conservation equation and the energy conservation equation. The finite volume method is used to discretize the calculation domain into grid units, and the second-order upwind format or the central difference format is applied to each control equation for spatial discretization. The time term is iteratively solved using an implicit or explicit format; after the calculation converges, the pressure difference between the inlet and outlet sections of the reactor is extracted as the fluid pressure drop value Δp through the post-processing module, and the velocity distribution mean square error or the turbulent kinetic energy distribution uniformity parameter of the flow field at each sample point is calculated based on the Euler-Euler two-fluid model as a uniformity index. The velocity distribution mean square error formula is: where u i is the velocity of each grid point, is the average speed, forming a data set containing the corresponding relationship between design variables and optimization objectives.

[0017] Preferably, in said S4, a mathematical regression model of optimization target and design variables is established based on the basic data set and the reliability is verified, specifically: a multiple regression analysis method is adopted, with the design variables as independent variables and the optimization target as the dependent variable, a mathematical regression model including linear terms, nonlinear terms and interaction terms of the independent variables is constructed, the model coefficients are fitted by the parameter estimation method, and the correlation coefficient R is used to calculate the optimal value. 2 and adjusted correlation coefficient The degree to which the model explains the data was evaluated, the overall significance of the model was tested using variance analysis, and the statistical characteristics of the error term were verified through residual analysis, so that its fitting accuracy and statistical reliability were consistent with the basic assumptions of regression analysis.

[0018] Preferably, the non-dominated sorting genetic algorithm NSGA-Ⅱ is used in S5 to iteratively optimize the baffle structure parameters, specifically: by randomly generating an initial population containing design variables, including the baffle radial distance and wing width, and the population size is determined according to the number of sample points; by performing genetic operations through selection operators, crossover operators and mutation operators, a progeny population is generated; after merging the parent and progeny populations, a fast non-dominated sort is performed based on the optimization target value of each individual, and the crowding degree of each individual is calculated to maintain population diversity; through the elite retention strategy, individuals with higher fitness are screened out to form a new population, and the above iterative process is repeated until the preset number of iterations is reached, and finally all non-dominated individuals are extracted from the population to form a Pareto optimal solution set.

[0019] Preferably, the elite retention strategy is used to screen out individuals with higher fitness to form a new population. Specifically, in the mixed population after the parent generation and the offspring generation are merged, non-dominated individuals with higher levels are preferentially selected according to the non-dominated sorting result. If the number of individuals of the same level exceeds the size of the new population, the distribution density of the individuals in the target space is calculated by the crowding comparison operator, and individuals with higher crowding are retained to maintain population diversity until the size of the new population reaches a preset number, thereby ensuring that individuals with high fitness are preferentially retained and inherited to the next generation during the iteration process.

[0020] Preferably, the flow field characteristic parameters in the reactor before and after optimization are compared in S6 and the mass transfer efficiency is verified, specifically: the baffle structure parameters before and after optimization are respectively imported into the computational fluid dynamics software, and the pressure drop values, velocity distribution mean square deviation and turbulent kinetic energy distribution uniformity parameters of the two sets of flow fields are simulated and obtained, and the optimization effect is evaluated by comparing and analyzing the change trends of each parameter; at the same time, under actual working conditions, the optimized baffle structure is installed in the jet impact-negative pressure reactor, and the mass transfer efficiency is calculated by measuring the change in substance concentration at the inlet and outlet of the reactor, and an error analysis is performed with the simulation results, and the optimal parameter combination that takes into account both pressure drop and uniformity is determined by combining the flow field characteristics and experimental data.

[0021] Compared with the prior art, the technical solution of this application has the following technical effects:

[0022] By combining the Latin hypercube sampling method with computational fluid dynamics simulation, the present invention can generate a uniformly distributed set of sample points while covering the parameter space of the design variables, thus avoiding the blindness and discreteness of the traditional trial-and-error method in exploring the parameter space; by dividing the value interval of each design variable with equal probability and random sampling, the uniform distribution of the sample points in the parameter space is ensured, so that the constructed basic data set can fully reflect the mapping relationship between the baffle structure parameters (such as radial distance, wing width) and the optimization objectives (pressure drop, uniformity), providing reliable data support for the establishment of the subsequent mathematical regression model, effectively improving the model's ability to describe complex flow field characteristics, and avoiding distortion of the optimization results due to sample bias.

[0023] The mathematical regression model constructed based on CFD simulation in this invention incorporates the linear terms, nonlinear terms and interaction terms of the design variables, which can accurately describe the complex coupling effects between the baffle structure parameters. 2 , analysis of variance), the model can quantitatively evaluate the individual and interactive effects of various parameters on pressure drop and uniformity, overcoming the limitations of traditional empirical formulas in describing nonlinear relationships between parameters. For example, the model reveals the mechanisms by which the quadratic effects and higher-order interactions of radial distance and wing width influence fluid turbulent diffusion and mass transfer efficiency. This provides a clear mathematical guide for multi-objective optimization, transforming the optimization process from a "black box" operation relying on experience to a scientific design based on mathematical statistics.

[0024] The introduction of the non-dominated sorting genetic algorithm NSGA-Ⅱ of the present invention realizes the multi-objective global optimization of the baffle structure parameters. Through population initialization, genetic operations (selection, crossover, mutation) and fast non-dominated sorting strategy, the algorithm can simultaneously optimize the pressure drop and uniformity objectives during the iteration process, effectively solving the multi-objective conflict problem. The elite retention strategy and the crowding calculation mechanism ensure the retention of population diversity and high-fitness individuals, avoid the algorithm from falling into local optimality, and thus obtain a uniformly distributed Pareto optimal solution set within a preset number of iterations. This method does not require pre-setting the objective function weights, and can provide multiple trade-off solutions for engineering selection, meeting the differentiated requirements for pressure drop and uniformity under different working conditions, and significantly improving the practicality and flexibility of the optimization results.

[0025] By comparing the flow field characteristic parameters (such as turbulent kinetic energy, velocity distribution, and vortex structure) before and after optimization and combining them with actual working condition experiments, the present technical solution can verify the mass transfer efficiency improvement effect of the optimized baffle structure. The CFD simulation results show that the optimized baffle structure can effectively break the eccentric flow and secondary reflux in the reactor, promote the splitting of large eddies and the generation of small eddies, enhance the degree of fluid turbulence and the uniformity of turbulent kinetic energy distribution, and reduce energy dissipation. Actual experimental data show that the efficiency of the optimized reactor is significantly improved in mass transfer processes such as ammonia nitrogen removal, verifying the reliability of the simulation results. This closed-loop design process of "simulation-optimization-experimentation" ensures the effective connection of the technical solution from theoretical modeling to engineering application, and provides a replicable technical path for the energy-saving and efficient design of chemical reactors.

[0026] 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 so that it can be implemented in accordance with the contents of the specification, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following is a detailed description of the preferred embodiment of the present application in conjunction with the accompanying drawings.

[0027] Based on the detailed description of the specific embodiments of the present application in conjunction with the accompanying drawings below, those skilled in the art will become more aware of the above and other objects, advantages and features of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0029] Figure 1 This is a flow chart of the baffle optimization method for a jet impact-negative pressure reactor based on CFD-genetic algorithm of the present invention;

[0030] Figure 2 The Pareto optimal solution set point diagram of models A, B, C and D in the embodiment of the present invention;

[0031] Figure 3 A comparison diagram of the homogeneity between the four prediction models and the original model in the embodiment of the present invention;

[0032] Figure 4 Volume fraction cloud diagram and trace distribution of the optimized structure and the original device structure in the embodiment of the present invention;

[0033] Figure 5 is the change of turbulent kinetic energy dissipation, turbulent kinetic energy, and velocity along the mainstream direction in the embodiment of the present invention;

[0034] Figure 6 The cloud diagrams of the streamwise vortex and spanwise vortex distribution of the optimized structure and the original device structure in the embodiment of the present invention are shown;

[0035] Figure 7 This is a comparison chart of ammonia nitrogen removal rates between the optimal structure and the original structure in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the following description, specific details such as specific configurations and components are provided only to help fully understand the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, for clarity and brevity, the description of known functions and structures has been omitted in the embodiments.

[0037] It should be understood that references throughout this specification to "one embodiment" or "this embodiment" mean that a particular feature, structure, or characteristic associated with the embodiment is included in at least one embodiment of the present application. Therefore, the appearance of "one embodiment" or "this embodiment" throughout this specification does not necessarily refer to the same embodiment. Furthermore, these particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0038] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0039] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" in this article describes another type of association object relationship, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0040] The term "at least one" in this article is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, at least one of A and B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0041] It should also be noted that, in this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprises," or any other variations thereof are intended to cover non-exclusive inclusion.

[0042] Example 1

[0043] This embodiment mainly describes the optimization method of the inner baffle of the jet impact-negative pressure reactor based on CFD-genetic algorithm. Figure 1 Shown, including:

[0044] S1. Taking minimizing the pressure drop and uniformity of the fluid in the jet impact-negative pressure reactor as optimization goals, at least two geometric structural parameters of the baffle are selected as design variables;

[0045] S2. Use Latin hypercube sampling to divide the parameter range of the design variable into intervals and generate a set of sample points covering the parameter space;

[0046] S3. Use computational fluid dynamics software to perform fluid dynamics simulation on the baffle structure corresponding to each sample point, obtain physical parameter data related to the optimization target, and construct a basic data set;

[0047] S4. Establish a mathematical regression model of the optimization objective and design variables based on the basic data set, incorporate linear terms, nonlinear terms, and interaction terms, and verify the fitting accuracy and reliability of the model through statistical test methods;

[0048] S5. Use the non-dominated sorting genetic algorithm NSGA-Ⅱ to iteratively optimize the baffle structure parameters. Through population initialization, genetic operation and non-dominated sorting strategy, the Pareto optimal solution set is obtained within the preset number of iterations.

[0049] S6. Compare the flow field characteristic parameters in the reactor before and after optimization, verify the mass transfer efficiency of the optimized baffle structure through actual working condition experiments, and determine the optimal parameter combination.

[0050] Furthermore, at least two geometric structural parameters of the baffle are selected in S1 as design variables, specifically: based on the structural characteristics of the jet impact-negative pressure reactor and the fluid mechanics optimization requirements, the radial distance of the radial installation position of the baffle from the central axis of the reactor, and the wing width of the baffle along the fluid flow direction are used as geometric parameters. By analyzing the influence of the number of baffles on the flow field distribution in the reactor, combined with the previous research basis, the number of baffles is determined to be a reasonable number that can effectively break the eccentric flow and secondary recirculation and promote the generation of vortex structure, so that the radial distance and wing width are used as adjustable design variables.

[0051] Furthermore, the Latin hypercube sampling method is used in S2 to divide the parameter range of the design variables into intervals and generate a set of sample points. Specifically, for the preset parameter range of each selected baffle geometric structure parameter, the value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability. Each sub-interval corresponds to a probability unit. A sample value is randomly selected in each sub-interval. By combining the sample values ​​of each design variable, a set of sample points that is evenly distributed and non-repetitive in the parameter space is generated.

[0052] Furthermore, the value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability. Specifically, for the value range of each design variable, the length of each sub-interval is determined by calculating the ratio of the total length of the variable interval to the number of preset sub-intervals, so that each sub-interval occupies an equal probability share in the entire value range, ensuring that the sample values ​​in each sub-interval have the same probability of being selected, and realizing a uniform probability distribution of the sample points in the parameter space.

[0053] Furthermore, computational fluid dynamics software is used in S3 to perform fluid mechanics simulation on the baffle structure corresponding to each sample point and obtain physical parameter data. Specifically, the geometric structure parameters of the baffle at each sample point are imported into the computational fluid dynamics software. Based on the actual size and boundary conditions of the reactor, including inlet flow velocity, outlet pressure, and fluid physical properties, the turbulence model and discrete format are used to numerically solve the gas-liquid two-phase flow field in the reactor. After the calculation is completed, the fluid pressure drop value and uniformity index corresponding to each sample point are extracted, including the velocity distribution variance and the turbulent kinetic energy distribution uniformity parameter, to construct a basic data set containing design variables and optimization objectives.

[0054] Furthermore, the turbulence model and discrete format are used to numerically solve the gas-liquid two-phase flow field in the reactor and extract parameters. Specifically, according to the flow characteristics of the gas-liquid two-phase flow in the reactor, the k-ε turbulence model or the k-ω turbulence model is selected to describe the turbulent pulsation effect. The control equation group is established through the continuity equation, momentum conservation equation and energy conservation equation. The finite volume method is used to discretize the computational domain into grid cells. The second-order upwind format or central difference format is applied to each control equation for spatial discretization. The time term is iteratively solved using an implicit or explicit format. After the calculation converges, the pressure difference between the inlet and outlet sections of the reactor is extracted as the fluid pressure drop value Δp through the post-processing module. The velocity distribution mean square error or the turbulent kinetic energy distribution uniformity parameter of the flow field at each sample point is calculated based on the Euler-Euler two-fluid model as a uniformity index. The velocity distribution mean square error formula is: where u i is the velocity of each grid point, is the average speed, forming a data set containing the corresponding relationship between design variables and optimization objectives.

[0055] Furthermore, in S4, a mathematical regression model of optimization objectives and design variables was established based on the basic data set and its reliability was verified. Specifically, a multiple regression analysis method was used, with the design variables as independent variables and the optimization objectives as dependent variables, to construct a mathematical regression model containing linear terms, nonlinear terms and interaction terms of independent variables. The model coefficients were fitted by the parameter estimation method, and the correlation coefficient R was used to calculate the optimal regression model. 2 and adjusted correlation coefficient The degree to which the model explains the data was evaluated, the overall significance of the model was tested using variance analysis, and the statistical characteristics of the error term were verified through residual analysis, so that its fitting accuracy and statistical reliability were consistent with the basic assumptions of regression analysis.

[0056] Furthermore, the non-dominated sorting genetic algorithm NSGA-Ⅱ is used in S5 to iteratively optimize the baffle structure parameters. Specifically, an initial population containing design variables, including the baffle radial distance and wing width, is randomly generated, and the population size is determined according to the number of sample points; genetic operations are performed through selection operators, crossover operators, and mutation operators to generate offspring populations; after merging the parent and offspring populations, a fast non-dominated sort is performed based on the optimization target value of each individual, and the crowding degree of each individual is calculated to maintain population diversity; individuals with higher fitness are selected through the elite retention strategy to form a new population, and the above iterative process is repeated until the preset number of iterations is reached, and finally all non-dominated individuals are extracted from the population to form a Pareto optimal solution set.

[0057] Furthermore, an elite retention strategy is used to select individuals with higher fitness to form a new population. Specifically, in the mixed population after the parent and offspring are merged, non-dominated individuals with higher ranks are preferentially selected according to the non-dominated sorting results. If the number of individuals of the same rank exceeds the size of the new population, the distribution density of individuals in the target space is calculated by the crowding comparison operator, and individuals with higher crowding are retained to maintain population diversity until the size of the new population reaches the preset number, thereby ensuring that individuals with high fitness are preferentially retained during the iteration process and inherited to the next generation.

[0058] Furthermore, in S6, the flow field characteristic parameters in the reactor before and after optimization are compared and the mass transfer efficiency is verified. Specifically, the baffle structure parameters before and after optimization are respectively imported into the computational fluid dynamics software, and the pressure drop values, velocity distribution mean square deviation and turbulent kinetic energy distribution uniformity parameters of the two sets of flow fields are simulated to obtain the optimization effect by comparing and analyzing the changing trends of each parameter. At the same time, under actual working conditions, the optimized baffle structure is installed in the jet impact-negative pressure reactor, and the mass transfer efficiency is calculated by measuring the changes in the substance concentration at the inlet and outlet of the reactor, and the error analysis is performed with the simulation results. The optimal parameter combination that takes into account both pressure drop and uniformity is determined by combining the flow field characteristics and experimental data.

[0059] This embodiment describes in detail how to generate sample points that evenly cover the parameter space through LHS, and combines CFD simulation to construct a high-precision response surface model containing linear, nonlinear and interaction terms to overcome the blindness of traditional experimental design; with the help of NSGA-Ⅱ's global search capability and elite retention strategy, it simultaneously optimizes the pressure drop and uniformity goals, obtains the Pareto optimal solution set, and solves the multi-objective conflict problem. This application deeply combines vortex structure analysis with multi-objective optimization algorithms, and realizes the transition from "passive adjustment" to "active design" by revealing the influence of baffle parameters on the degree of flow field turbulence and turbulent kinetic energy distribution, providing a new paradigm for reactor structure optimization that combines theoretical depth and engineering practicality.

[0060] Based on Example 1, this implementation describes in detail the optimization effect of NSGA-II of the present application, specifically:

[0061] In this study, NSGA-Ⅱ was used to solve the multi-objective optimization problem. The regression fitting formula obtained by RSM design was used as the fitness function. The NSGA-Ⅱ algorithm was used to analyze the fitness function of the internal fluid of the JI-NPR reactor and obtain the Pareto optimal solution. With the baffle wing width h and radial distance r as input values, and the reactor pressure drop and uniformity as output values, the NSGA-Ⅱ algorithm controlled the parameter population size to 36, the maximum number of generations to 1000, and the Pareto solution set ratio to 0.4. A Pareto optimal solution set containing 36 individuals was obtained, as shown in the following figure: Figure 2 As shown, Figure 2All Pareto frontier solutions are shown. These solutions are affected by two functions. Therefore, when one objective function reaches the optimal value, the performance of the other objective function is poor. This study takes four Pareto solutions as examples and names the four models as A, B, C and D. For example, point A has the smallest pressure drop but the smallest uniformity, that is, the mixing inside the device is the most uniform; point D has the largest uniformity but the largest pressure drop, that is, the driving force for gas-liquid two-phase mass transfer is the largest. Moving from point A to point B, the objective function pressure drop value increases by 0.026%, and the objective function uniformity value increases by 72.4%. It can be seen that the mixing effect between fluids deteriorates by increasing the mass transfer driving force. Moving from point C to point D, the objective function pressure drop value increases by 0.183%, and the objective function uniformity value increases by 39.4%. The corresponding design variables are shown in Table 1, and the optimized model structure is compared with the original device structure;

[0062] Table 1 Design variable values ​​for the four optimal points

[0063]

[0064] As shown in Table 2, there is a comparison between the predicted values ​​of the four Pareto points and the CFD simulation values. It can be seen that the error value of the pressure drop characteristic is less than 0.2%, and the error value of the uniformity characteristic is less than 5%; the error values ​​of the predicted values ​​and the simulation values ​​of the four prediction points are all less than 5%. Therefore, the model meets the requirements. The pressure drop of the original device model is 9600Pa, and the uniformity is 0.09. Compared with the original device structure, the optimized pressure drop increased by 1.53%, 1.55%, 1.69% and 1.88% respectively, and the uniformity decreased by 75%, 56.9%, 31.6% and 4.6% respectively. The change law of the two objectives of pressure drop and uniformity of the JI-NPR reactor with the design variables of the Pareto optimal solution is beneficial to the structural design of the reactor;

[0065] Table 2 Comparison of the predicted values ​​of the four Pareto points and the CFD simulation values

[0066]

[0067] This example uses the RSM regression fitting formula as the fitness function and the NSGA-Ⅱ algorithm to perform multi-objective optimization on the structural parameters of the JI-NPR reactor baffle, and successfully obtains a Pareto optimal solution set containing 36 individuals. The results show that different Pareto solutions show significant trade-off characteristics between the pressure drop and uniformity objectives: for example, point A has the smallest pressure drop but the lowest uniformity (the most uniform mixing), and point D has the largest uniformity but the highest pressure drop (the largest mass transfer driving force). The errors between the predicted values ​​and CFD simulation values ​​corresponding to each optimal solution are all controlled within 5%, verifying the reliability of the model. Compared with the original device, the pressure drop after optimization increased by 1.53%-1.88%, and the uniformity decreased by 4.6%-75%. By adjusting the wing width and radial distance, the mass transfer driving force and mixing uniformity are effectively balanced, providing a parameter combination applicable to multiple scenarios for the reactor structure design, and achieving a breakthrough from single-objective optimization to multi-objective global optimization.

[0068] Based on Example 1, this implementation describes in detail the performance and flow field analysis of the present application before and after optimization, specifically:

[0069] like Figure 3 As shown in the figure, the uniformity performance of the reactor without baffle structure is compared with that of the reactor with optimal structure. It can be seen that the uniformity of the reactor without baffle structure (orange scattered point E) is higher than that of the reactor with optimal structure, and the difference shows a trend of first increasing and then decreasing. It can be seen that the baffle structure is conducive to enhancing the degree of disorder of the flow field inside the reactor. As a result, the internal flow field presents a nonlinear spatiotemporal disorder state, thereby improving the mass transfer effect between the gas and liquid phases. In addition, from Figure 3 This can also be seen in the cloud diagram. After entering the jet orifice, the fluid impacts and moves upward under the influence of the negative pressure at the top. The upward flow is impeded by the baffle structure. From the impact point until the fluid passes through Z = 450 mm, its cross-sectional area expands. Passing the baffle structure disrupts the original flow structure, resulting in non-periodic, irregular motion.

[0070] Taking the Pareto A model as an example, the volume fraction cloud and trace distribution of the optimized device structure and the original device structure are compared. Figure 4It can be seen from the volume fraction cloud map and the trace distribution map that the liquid phase of the original device is concentrated at the central axis. The baffle structure breaks the original fluid flow trajectory, improves the original eccentric flow and overflow problems, and reduces damage to the equipment. In addition, the baffle structure causes the fluid to shift toward the boundary area, and the liquid phase distribution area is expanded. The degree of chaos inside the device is increased, and the atomization effect is enhanced. After the fluid flows through the baffle, a vortex structure is formed, which then sucks the fluid into this area to form a vortex. This vortex structure enhances turbulent diffusion, promotes the diffusion of molecules between phases, and significantly improves the mixing and mass transfer efficiency. In addition, the dead angle area at the top of the device has also been improved. Due to the action of the vortex, the fluid is fully mixed, effectively avoiding the situation of uneven fluid mixing.

[0071] like Figure 5 The curves of turbulent kinetic energy dissipation TEP, turbulent kinetic energy TEK, and velocity V along the vertical direction of the JI-NPR reactor are given when the jet velocity is 4m / s and the top negative pressure is 10000Pa. Among them, baffle structures that increase turbulence are installed at Z=450mm and 720mm. Figure 5 It can be seen that in the jet impact area, the turbulent kinetic energy dissipation of the original model is higher than that of the optimized model, indicating that the original model caused a large amount of energy dissipation during the impact process. However, in the negative pressure separation area, the turbulent kinetic energy dissipation increased significantly after the fluid passed through the baffle. This shows that the baffle structure causes the vortex structure to break up and disperse, which is beneficial to the mass transfer separation between phases; in addition, it can be clearly seen that the turbulent kinetic energy of the optimized structural device is significantly higher than that of the original device. This shows that the addition of the baffle structure greatly improves the irregular movement of the fluid inside the reactor. The turbulent kinetic energy of the fluid inside the reactor increases, the nonlinear and complex movement of the fluid is enhanced, while the turbulent kinetic energy dissipation between the fluids decreases, and the energy dissipated in the form of heat is reduced; this shows that more energy is used for the mass transfer separation process between the gas and liquid phases. From the velocity distribution diagram, it can be seen that the velocity distribution difference between the original device structure and the optimized device structure in the slow flow area and the jet impact area is small, and the baffle structure has no significant effect on the velocity here. In the negative pressure separation area, because the baffle structure makes the flow cross-section smaller, it brings resistance to the fluid flow. The fluid must overcome a larger pressure difference, thereby accelerating the flow rate; from Figure 5 As can be seen in the figure, the velocity of the fluid increases significantly after passing through the baffle. The increased flow rate significantly increases the reaction rate inside the reactor, thereby improving production efficiency.

[0072] like Figure 6 As shown, Figure 6The cloud diagrams of the flow vortex and spanwise vortex distribution before and after the device is optimized. The positive and negative values ​​of the flow vortex only represent its direction. It can be clearly seen that more flow vortex structures are formed in the negative pressure separation zone during the fluid flow of the optimized device. This shows that the optimized device can effectively enhance the mixing effect of the fluid and improve the mass transfer and separation efficiency. The flow vortex of the original device is mainly located in the boundary layer area, the jet impact area and the top outlet position. The optimized device adds a baffle structure at Z = 450 mm, which hinders the fluid flow through the baffle, causing irregular movement of the fluid. The fluid gathers at the bottom to form a flow vortex structure. When the fluid flows through the baffle at Z = 720 mm, the blocking effect of the baffle changes the flow pattern of the fluid, resulting in uneven distribution of fluid velocity and the formation of a complex vortex structure area behind the baffle.

[0073] The change in the spanwise vortex of the fluid is caused by the velocity gradient. The spanwise vortex of the fluid is mainly distributed at the jet impact point, the top negative pressure area and the slow flow area. The stronger the spanwise vortex of the fluid, the more uniform the fluid mixing. It makes the fluid mix more fully inside and increases the contact area and contact time between the gas and liquid phases. During the fluid jet impact process, the velocity changes significantly. Specifically, it rapidly decreases from 4m / s at the jet mouth to 0m / s at the impact point. After passing through the jet, the fluid has an obvious vertical downward flow, which is mainly affected by gravity. In the negative pressure separation area, after the fluid flows through the baffle structure, it is significantly affected by the baffle, and a vortex structure is formed behind the baffle. This process causes the fluid velocity to decrease sharply and forms a large velocity gradient. At the top outlet, due to the effect of negative pressure, the fluid presents disordered non-steady-state characteristics. It is mainly manifested in a significant increase in the randomness and instability of the fluid flow, resulting in an increase in the unevenness of the fluid distribution.

[0074] This example describes in detail the performance and flow field analysis before and after optimization. By comparing the flow field characteristics of a baffle-free structure with those after optimization, the study reveals the mechanism by which baffles enhance reactor performance. The results show that the baffle structure significantly enhances flow field turbulence, overcomes the eccentric flow and secondary reflux issues of the original device, and expands the liquid phase distribution from the central axis to the boundary regions, expanding the mixing range and enhancing the atomization effect. Vortex structure analysis reveals that the optimized device forms more streamwise and spanwise vortices in the negative pressure separation zone, promoting turbulent diffusion and interphase mass transfer while reducing fluid retention in the top dead zone. Turbulent kinetic energy and velocity distributions reveal that while the baffle increases energy dissipation in the impact zone, it increases flow velocity and turbulent kinetic energy in the negative pressure zone by reducing the flow cross-section, reducing energy loss as heat and improving mass transfer efficiency. Experimental verification demonstrates that the optimized ammonia nitrogen removal rate increases by 37.4% compared to the original structure, confirming the significant enhancement of the baffle structure on the mass transfer and separation process.

[0075] Based on Example 1, this embodiment describes in detail whether the improved device structure of this application can promote the removal of ammonia nitrogen in ammonia nitrogen wastewater, specifically:

[0076] To validate the simulation results, we experimentally verified whether the improved device structure could promote the removal of ammonia nitrogen from ammonia nitrogen wastewater. The ammonia nitrogen removal rate in ammonia nitrogen wastewater is measured by the ammonia nitrogen content in the solution. Therefore, the difference between the NH4+ content in the solution before and after the reaction is calculated and then the ratio is calculated to the NH4+ content in the solution before the reaction. This value is called the ammonia nitrogen removal rate, and its functional expression is: Where η is the ammonia nitrogen extraction rate (%), After the reaction Content (μg / mL), Before the reaction Content (μg / mL);

[0077] From the above simulation results, it can be seen that the ammonia nitrogen removal rate is closely related to the structure of the JI-NPR reactor device. Baffle structures are set at Z=450mm and 720mm in the reactor. By adjusting the number of baffles, radial distance and wing width, the flow state of the fluid inside the reactor can be effectively controlled. Thereby achieving the purpose of improving the ammonia nitrogen removal efficiency of the reactor. In order to verify the reliability of the simulation results, the experiment also adopted the same operating conditions as the simulation. The outlet pressure is -10000Pa and the jet velocity is 4m / s. The experiment used 45L simulated ammonia nitrogen wastewater. During the experiment, a sample was collected at the sampling port every 1h. The collected samples were processed to obtain the ammonia nitrogen removal rate at the corresponding moment. The experiment was repeated three times, and the average value was obtained. The figure shows a comparison of the ammonia nitrogen removal rate of the optimal structure and the original structure. After statistical analysis, the standard deviation of the data of the three repeated experiments was small, indicating that the data repeatability is good and the experimental results have a high degree of credibility. From Figure 7 As can be seen from the figure, the addition of a baffle structure to the JI-NPR reactor significantly improves ammonia nitrogen removal efficiency. The optimized JI-NPR reactor with the baffle structure achieves a 60% ammonia nitrogen removal efficiency after 7 hours. After 10 hours of continuous operation, the optimized reactor structure achieves a 70% ammonia nitrogen removal rate. The original structure achieves a removal rate of less than 50%, resulting in a 37.4% improvement in ammonia removal efficiency.

[0078] This embodiment describes in detail the actual effectiveness of the optimized baffle structure verified by the ammonia nitrogen wastewater removal experiment. The experiment adopted the same operating conditions as the simulation (outlet pressure -10000Pa, jet velocity 4m / s), and compared the performance before and after optimization with the ammonia nitrogen removal rate as an indicator. The results showed that the ammonia nitrogen removal rate of the original structure reactor was less than 50% after 10h of continuous operation, while the removal rate of the optimized reactor reached 60% at 7h and increased to 70% at 10h, and the deammonification efficiency was increased by 37.4%. The low standard deviation of the three repeated experiments showed that the data had good repeatability, which verified the reliability of the simulation results. The experiment showed that the optimized baffle structure significantly enhanced the mass transfer contact area and time between the gas and liquid phases by regulating the flow field vortex structure and pressure drop characteristics, realizing the effective transformation from simulation optimization to engineering application, and providing an experimental basis for the energy-saving and efficient operation of the JI-NPR reactor in actual wastewater treatment.

[0079] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention can be modified and varied in various ways. Any changes, modifications, replacements, integrations and parameter changes to these embodiments through conventional substitutions or that can achieve the same functions without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. A CFD-genetic algorithm-based baffle optimization method for jet impact-negative pressure reactor, characterized in that: include: S1. Taking minimizing the pressure drop and uniformity of the fluid in the jet impact-negative pressure reactor as optimization goals, at least two geometric structural parameters of the baffle are selected as design variables; S2. Using Latin hypercube sampling, divide the parameter range of the design variable into intervals to generate a set of sample points covering the parameter space; S3. Using computational fluid dynamics software to perform fluid dynamics simulation on the baffle structure corresponding to each sample point, obtain physical parameter data related to the optimization target, and construct a basic data set; S4. Establishing a mathematical regression model of the optimization objective and design variables based on the basic data set, incorporating linear terms, nonlinear terms, and interaction terms, and verifying the fitting accuracy and reliability of the model through statistical test methods; S5. Use the non-dominated sorting genetic algorithm NSGA-Ⅱ to iteratively optimize the baffle structure parameters. Through population initialization, genetic operation and non-dominated sorting strategy, the Pareto optimal solution set is obtained within the preset number of iterations. S6. Compare the flow field characteristic parameters in the reactor before and after optimization, verify the mass transfer efficiency of the optimized baffle structure through actual working condition experiments, and determine the optimal parameter combination.

2. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: At least two geometric structural parameters of the baffle are selected in S1 as design variables, specifically: based on the structural characteristics and fluid mechanics optimization requirements of the jet impact-negative pressure reactor, the radial distance of the radial installation position of the baffle from the central axis of the reactor, and the wing width of the baffle along the fluid flow direction are used as geometric parameters. By analyzing the influence of the number of baffles on the flow field distribution in the reactor, combined with the previous research basis, the number of baffles is determined to be a reasonable number that can effectively break the eccentric flow and secondary reflux and promote the generation of vortex structure, so that the radial distance and wing width are used as adjustable design variables.

3. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: In S2, the Latin hypercube sampling method is used to divide the parameter range of the design variables into intervals and generate a set of sample points. Specifically, for the preset parameter range of each selected baffle geometric structure parameter, the value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability. Each sub-interval corresponds to a probability unit. A sample value is randomly selected in each sub-interval. By combining the sample values ​​of each design variable, a set of sample points that is evenly distributed and non-repetitive in the parameter space is generated.

4. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 3, characterized in that: The value interval of each design variable is divided into several non-overlapping sub-intervals according to the principle of equal probability. Specifically, for the value range of each design variable, the length of each sub-interval is determined by calculating the ratio of the total length of the variable interval to the number of preset sub-intervals, so that each sub-interval occupies an equal probability share in the entire value range, ensuring that the sample values ​​in each sub-interval have the same probability of being selected, and achieving a uniform probability distribution of the sample points in the parameter space.

5. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: In the S3, computational fluid dynamics software is used to perform fluid dynamics simulation on the baffle structure corresponding to each sample point and obtain physical parameter data. Specifically, the geometric structure parameters of the baffle at each sample point are imported into the computational fluid dynamics software. Based on the actual size and boundary conditions of the reactor, including inlet flow velocity, outlet pressure, and fluid physical properties, a turbulence model and a discrete format are used to numerically solve the gas-liquid two-phase flow field in the reactor. After the calculation is completed, the fluid pressure drop value and uniformity index corresponding to each sample point are extracted, including the velocity distribution variance and the turbulent kinetic energy distribution uniformity parameter, to construct a basic data set containing design variables and optimization objectives.

6. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 5, characterized in that: The method uses a turbulence model and a discrete format to numerically solve the gas-liquid two-phase flow field in the reactor and extract parameters, specifically: according to the flow characteristics of the gas-liquid two-phase flow in the reactor, selects a k-ε turbulence model or a k-ω turbulence model to describe the turbulent pulsation effect, establishes a control equation group through the continuity equation, the momentum conservation equation and the energy conservation equation, uses the finite volume method to discretize the computational domain into grid cells, applies a second-order upwind format or a central difference format to each control equation for spatial discretization, and uses an implicit or explicit format to iteratively solve the time term; After the calculation converges, the pressure difference between the inlet and outlet sections of the reactor is extracted as the fluid pressure drop value Δp through the post-processing module. Based on the Euler-Euler two-fluid model, the velocity distribution mean square deviation or turbulent kinetic energy distribution uniformity parameter of the flow field at each sample point is calculated as a uniformity index. The velocity distribution mean square deviation formula is: where u i is the velocity of each grid point, is the average speed, forming a data set containing the corresponding relationship between design variables and optimization objectives.

7. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: In said S4, a mathematical regression model of optimization objectives and design variables is established based on the basic data set and its reliability is verified. Specifically, a multiple regression analysis method is adopted, with the design variables as independent variables and the optimization objectives as dependent variables, to construct a mathematical regression model containing linear terms, nonlinear terms and interaction terms of the independent variables, and the model coefficients are fitted by the parameter estimation method. The correlation coefficient R is used to calculate the optimal regression model. 2 and adjusted correlation coefficient The degree to which the model explains the data was evaluated, the overall significance of the model was tested using variance analysis, and the statistical characteristics of the error term were verified through residual analysis, so that its fitting accuracy and statistical reliability were consistent with the basic assumptions of regression analysis.

8. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: In S5, a non-dominated sorting genetic algorithm (NSGA-II) is used to iteratively optimize the baffle structural parameters. Specifically, an initial population containing design variables, including the baffle radial distance and wing width, is randomly generated, and the population size is determined according to the number of sample points; a genetic operation is performed by selecting an operator, a crossover operator, and a mutation operator to generate a progeny population; after merging the parent and progeny populations, a fast non-dominated sort is performed based on the optimization target value of each individual, and the crowding degree of each individual is calculated to maintain population diversity; Through the elite retention strategy, individuals with higher fitness are selected to form a new population. The above iterative process is repeated until the preset number of iterations is reached. Finally, all non-dominated individuals are extracted from the population to form the Pareto optimal solution set.

9. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 8, characterized in that: The elite retention strategy is used to select individuals with higher fitness to form a new population. Specifically, in the mixed population after the parent generation and the offspring generation are merged, non-dominated individuals with higher ranks are preferentially selected according to the non-dominated sorting result. If the number of individuals of the same rank exceeds the size of the new population, the distribution density of the individuals in the target space is calculated by the crowding comparison operator, and individuals with higher crowding are retained to maintain population diversity until the size of the new population reaches a preset number, thereby ensuring that individuals with high fitness are preferentially retained during the iteration process and inherited to the next generation.

10. The CFD-genetic algorithm-based jet impact-negative pressure reactor inner baffle optimization method according to claim 1, characterized in that: In the S6, the flow field characteristic parameters in the reactor before and after optimization are compared and the mass transfer efficiency is verified. Specifically, the baffle structure parameters before and after optimization are respectively imported into the computational fluid dynamics software, and the pressure drop values, velocity distribution mean square deviation and turbulent kinetic energy distribution uniformity parameters of the two groups of flow fields are simulated and obtained, and the optimization effect is evaluated by comparing and analyzing the change trends of each parameter; at the same time, under actual working conditions, the optimized baffle structure is installed in the jet impact-negative pressure reactor, and the mass transfer efficiency is calculated by measuring the changes in the substance concentration at the inlet and outlet of the reactor, and the error analysis is performed with the simulation results, and the optimal parameter combination that takes into account both pressure drop and uniformity is determined by combining the flow field characteristics and experimental data.

Images