Hydrophobic valve based on TEMOF-NSGA-II algorithm optimization

Abstract

The application relates to the technical field of hydrophobic valves, and proposes a hydrophobic valve based on TEMOF-NSGA-II algorithm optimization, which comprises a hydrophobic valve body and a synergistic module, and the module is a static integrated structure of a Venturi tube section, a cyclone steady flow cavity section and a casing type heat energy recovery section connected in sequence. The throat area of the Venturi tube section, the installation angle of the guide vane in the cyclone steady flow cavity section and the heat transfer coefficient of the casing type heat energy recovery section are determined through a specific multi-objective optimization process. The process first performs global exploration through a proxy model, then switches to a high-fidelity model based on an initial solution set for local refinement, and finally selects an optimal compromise solution from a Pareto optimal solution set, the corresponding parameters of which are optimal parameter combinations for guiding manufacturing. Therefore, through the design paradigm, optimal static parameters are determined before manufacturing, so that the hydrophobic valve can inherently have excellent performance suitable for wide fluctuation working conditions, thereby solving the inherent contradiction in traditional design.

Description

Technical Field

[0001] This invention relates to the field of steam trap technology, and in particular to a steam trap optimized based on the TEMOF-NSGA-II algorithm. Background Technology

[0002] Traditional disc-type steam traps, as key components of steam systems, directly impact energy consumption, equipment reliability, and operating costs in industrial production. However, existing technologies have long faced an irreconcilable contradiction: on the one hand, pursuing high condensation efficiency and reliability often relies on complex dynamic control valves, whose high cost, complex control systems, and inherent mechanical failure risks limit their widespread application; on the other hand, while simple, robust, and durable static steam traps offer less maintenance, their performance is fixed after design and cannot adapt to the wide and fluctuating operating conditions of steam systems. This leads to a sharp drop in efficiency outside the design point, and systemic problems such as high vibration and noise, rapid cavitation wear at the valve orifice, and significant waste of sensible heat from high-temperature condensate are common. The root of these problems lies in the fact that traditional design methods heavily rely on engineers' experience and limited trial and error, failing to find a globally optimal solution among multiple conflicting performance objectives. The final product is often a compromise, with significant deficiencies in its overall performance.

[0003] A utility model patent application with authorization publication number CN222864670U discloses a Venturi steam trap with bypass, comprising a valve body with a hollow inner cavity, and a slide-type valve core disposed within the hollow cavity. The slide-type valve core includes a movable disc base, a movable disc, a fixed disc, and a pressure cap, which are sequentially mounted on the valve body from the inside out. The movable disc has multiple first flow channel holes; the fixed disc has multiple second flow channel holes. When the steam trap is closed, the first and second flow channel holes are in a staggered closed state; when the steam trap is open, the first and second flow channel holes are in a aligned and connected state. One of the first flow holes is aligned and connected with the second flow channel hole to form a drainage channel, and a Venturi nozzle is threadedly connected to the drainage channel.

[0004] For example, the aforementioned technical solution improves the overall performance of the steam trap through structural modifications. However, these structural improvements require repeated verification and correction. The main problem with traditional steam traps and their improvement technologies stems from the limitations of their design methods. Existing technologies heavily rely on engineers' personal experience, approximate formulas, and a process of trial and error when determining valve structural parameters. Designers first pre-determine a structural scheme based on experience, then manufacture physical samples for testing, manually adjust parameters based on test results, and then begin a new round of trial production and verification. This method is not only time-consuming and costly, but the final design is often only a compromise solution that performs adequately under specific operating conditions, failing to guarantee consistently excellent performance under complex real-world working conditions.

[0005] This outdated design philosophy directly led to a series of obvious defects in the product during actual use. In terms of reliability, the valves often generated significant vibration and noise; critical internal components such as the valve core and seat were prone to premature failure due to erosion and wear, resulting in poor sealing and steam leakage, thus shortening the overall service life. Regarding energy efficiency, the large amount of heat energy contained in the discharged high-temperature condensate was usually wasted without effective utilization. Even with the addition of simple heat recovery devices, the recovery efficiency was often low due to incompatibility with the dynamic drainage characteristics of the steam trap, and it might even interfere with the normal drainage function of the valve by increasing back pressure. Summary of the Invention

[0006] In view of this, the present invention proposes a steam trap based on the TEMOF-NSGA-II algorithm optimization, which adopts a "deterministic pre-optimization" paradigm and changes the traditional "experience-trial and error" design method. This ensures from the source that the steam trap's structure inherently possesses the optimal performance potential to cope with complex working conditions before manufacturing, rather than performing well only near specific working points like existing products. This solves the design problems of existing steam traps.

[0007] The technical solution of this invention is implemented as follows: On the one hand, the present invention provides a steam trap optimized based on the TEMOF-NSGA-II algorithm, including a steam trap body and an efficiency enhancement module fixedly connected to the outlet end of the steam trap body; The efficiency enhancement module is a static integrated structure, which includes a Venturi tube section, a swirl flow stabilizing cavity section and a sleeve-type heat recovery section connected in series. The throat area A of the venturi section and the installation angle of the guide vanes in the swirl flow stabilization cavity section. The heat transfer coefficient U of the shell-and-tube heat recovery section is determined based on the following optimization design process; The execution process of the TEMOF-NSGA-II optimization algorithm includes: S1. Construct a multi-objective optimization problem with conflicting objectives of minimizing flow-induced vibration, minimizing flow pressure loss, and maximizing heat recovery efficiency. The design variables are throat area A and installation angle. and heat transfer coefficient U; S2. The TEMOF-NSGA-II two-stage framework is used to solve this optimization problem: In the first stage, a surrogate model is used to explore the global potential region and output the initial Pareto solution set; In the second stage, the initial Pareto solution set is used as the starting point, and the high-fidelity simulation model is switched to perform local refinement in the potential region until the convergence condition based on the hypervolume index is met, and the final Pareto optimal solution set is obtained. S3. Select the best compromise solution from the final Pareto optimal solution set, and select the corresponding throat area A and installation angle. And the heat transfer coefficient U, as the optimal combination of parameters.

[0008] Based on the above technical solutions, preferably, in step S1, the multi-objective optimization problem is specifically defined as minimizing the objective vector. ,in, , and This refers to the throat area of ​​the Venturi tube segment; And θ is the installation angle of the guide vanes in the swirling flow stabilization cavity section; and The overall heat transfer coefficient of the shell-and-tube heat recovery section; The vibration index is defined based on the root mean square of the pulsating pressure on the flow channel wall. This is a pressure loss indicator defined based on the total pressure drop of the efficiency enhancement module; The heat recovery efficiency index is defined based on the heat exchange performance of the shell-and-tube heat recovery section. This indicates minimization.

[0009] Based on the above technical solutions, preferably, in step S2, the surrogate model is a Kriging model, which is trained by obtaining high-fidelity simulation data through Latin hypercube sampling, and is used to quickly predict the approximate value of the objective function under different design variables.

[0010] Based on the above technical solutions, preferably, in step S2, the convergence condition based on the hypervolume index is: in L consecutive generations of evolution, the hypervolume of the final Pareto solution set... The relative improvement is less than the set threshold, resulting in over-volume. Defined as the volume enclosed by the solution set and the reference point in the target space.

[0011] Based on the above technical solutions, preferably, in step S2, the TEMOF-NSGA-II algorithm uses the cross-distribution index during the first-stage global exploration. ∈[5,15] and the variation distribution index ∈[5,15], the cross-distribution index used in the second stage of local refinement. ∈[50,100] and the variation distribution index ∈[50,100].

[0012] Based on the above technical solutions, preferably, in the local refinement stage of step S2, the mutation operation used by the TEMOF-NSGA-II algorithm has a mutation distribution index. ∈[50,100], such that the probability distribution exist The mutation operation reaches its peak at 0, and follows the following probability distribution: , in, It is the variation distribution index; This represents the variation perturbation value.

[0013] Based on the above technical solutions, preferably, in step S3, the best compromise solution is selected by calculating the weighted Chebyshev distance from each solution in the final Pareto optimal solution set to the ideal point, and the solution with the smallest distance is selected.

[0014] Based on the above technical solutions, preferably, the Venturi tube section is configured to perform initial throttling and kinetic energy dissipation on the fluid to suppress cavitation; the swirl flow stabilization cavity section is configured to convert the fluid into a swirling flow of optimized intensity to promote vapor-liquid separation and reduce vibration and noise; and the shell-and-tube heat recovery section is configured to achieve efficient recovery of sensible heat with minimal flow resistance.

[0015] Based on the above technical solutions, preferably, the Venturi tube section has a gradually converging section, a throat, and a diffuser section determined according to the optimal throat area A. The convergence angle of the gradually converging section, the length and diameter of the throat, and the diffusion angle of the diffuser section work together with the optimal throat area A to optimize the flow field.

[0016] Based on the above technical solutions, preferably, the swirl flow stabilization cavity is a cylindrical cavity with two ends having optimal installation angles. Fixed guide vanes; The shell-and-tube heat recovery section includes an inner tube and an outer tube arranged coaxially, as well as a shell-and-tube heat exchanger located in the inner and outer tubes. The inner and outer tubes form a counter-current heat exchange channel. The optimal heat transfer coefficient U is determined by the material, wall thickness, spacing, and flow channel length of the shell-and-tube heat recovery section.

[0017] The steam trap optimized by the TEMOF-NSGA-II algorithm of this invention has the following advantages over the prior art: (1) This invention changes the traditional design concept of single or simple superimposed functional components of steam traps by constructing an integrated efficiency-enhancing module with global algorithm optimization. The module consists of three-level structural units connected in series, and all key dimensions are precisely determined by the optimization algorithm. The first level is a fixed Venturi tube, whose unique throat area and contraction-diffusion angle are precisely calculated. Its function is to optimally throttle and decelerate the high-speed fluid from the source of the flow field. The effect is to reduce the fluid kinetic energy, fundamentally suppress the generation of cavitation effect and erosion wear at the valve port, and greatly extend the valve life. The second level is a fixed swirl flow stabilizing cavity, which is equipped with guide vanes with optimized angles. Its function is to guide the fluid to form a stable swirl. The effect is to effectively break bubbles, promote droplet coalescence, and significantly reduce noise and pipe vibration. The third level is an integrated heat recovery jacket, whose structural parameters are optimally matched with the internal flow channel. Its function is to efficiently recover the sensible heat in the condensate while ensuring smooth drainage. The effect is to improve the overall energy efficiency of the system while avoiding the increase in back pressure due to excessive cooling. These three-tiered structures are not simply combined, but rather form a synergistically optimized whole that collectively achieves vibration damping, energy saving, and long lifespan. Thus, through a "deterministic pre-optimization" design paradigm, the optimal static parameters are determined before manufacturing, enabling the steam trap to inherently possess superior performance adaptable to a wide range of fluctuating operating conditions, thereby resolving the inherent contradictions in traditional designs.

[0018] (2) The core innovation of this invention in terms of algorithm lies in the two-stage optimization framework of TEOF-NSGA-II. TEOF is a "coarse-to-fine" intelligent optimization strategy, and NSGA-II is a multi-objective evolutionary optimization algorithm. This framework transforms engineering design from an experience-based model to a data-driven model. Its workflow begins with the mathematical modeling of three key static parameters of the steam trap (Venturi throat area A, guide vane angle θ, and heat transfer coefficient U), defining them as design variables in a high-dimensional search space. The primary task of the algorithm is to simultaneously minimize vibration, pressure loss, and maximize heat recovery efficiency through multi-objective optimization, which overcomes the inherent defect of traditional methods in handling multiple conflicting performance indicators. To achieve this goal, the algorithm operates in two continuous and coordinated phases: the first phase is global exploration, which uses a low-computational-cost surrogate model to quickly approximate the real fluid dynamics and heat transfer processes, driving the NSGA-II algorithm to perform an initial search within a broad design space. Its purpose is to efficiently locate potentially high-performance areas and avoid getting trapped in local optima. The second phase is local refinement, where the algorithm uses the excellent design schemes selected in the first phase as the initial population and then calls a high-precision, high-fidelity simulation model for fine-tuning. Its purpose is to deeply explore within the narrowed high-performance area, ensuring the accuracy and reliability of the final parameters. Throughout the process, non-dominated sorting and congestion calculation, as the core mechanisms of the NSGA-II algorithm, ensure that the algorithm can find a series of Pareto optimal solutions that achieve the best trade-offs among multiple performance objectives, rather than a single compromise. Finally, the algorithm outputs a specific set of optimal parameter combinations (A, θ, U), which serves as the direct basis for manufacturing, enabling the produced static structural steam traps to possess good performance in handling complex operating conditions without any online adjustments.

[0019] (3) In this invention, after the condensate drain body completes the basic "vapor blocking and drainage" action, the mixture of high-temperature condensate and a small amount of flash vapor first enters the Venturi tube section. This structure accelerates the fluid through its precisely optimized tapering section, forming a low-pressure zone at the throat. Its function is not only to reduce the fluid pressure energy, but more importantly, to promote the flash evaporation of some liquid water. The vapor-liquid two-phase flow decreases in velocity and recovers pressure in the expanding section. This process can effectively consume excess kinetic energy and provide a preliminary stable flow field for the subsequent process. Immediately afterwards, the fluid flows tangentially into the swirling flow stabilization chamber section. Guided by the optimized angle of the guide vanes, a strong swirling flow is formed. Its function is to use centrifugal force to throw the denser droplets toward the wall and collect them into a liquid film, while the less dense vapor concentrates in the central axis area, realizing the preliminary separation and deep coalescence of vapor and liquid. The effect is to significantly reduce flow noise and stabilize the flow rate. Finally, the pretreated fluid enters the flow channel of the shell-and-tube heat recovery section. Here, heat is transferred to the external cooling medium through the wall. Its function is to efficiently recover the sensible heat of the condensate and condense some of the flash vapor, reducing the total volume and temperature of the outflowing medium, thus completing the entire process of hydrophobicity, flow stabilization and energy recovery. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a three-dimensional view of the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention; Figure 2 This is a top view of the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention; Figure 3 For the present invention Figure 2 Sectional view along the AA direction; Figure 4 For the present invention Figure 3 Enlarged view of the structure at point A in the middle; Figure 5 This is a three-dimensional view of the enhancement module of the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention. Figure 6 An exploded view of the enhancement module of the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention; Figure 7 This is a system operation diagram of the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention; Figure 8The flowchart of the optimized TEMOF-NSGA-II algorithm of this invention is shown below; In the diagram: 1. Steam trap body; 101. Outlet; 2. Enhancement module; 21. Venturi tube section; 211. Converging section; 212. Throat; 213. Diffusion section; 22. Swirl flow stabilizing cavity section; 221. Guide vane; 23. Shell-and-tube heat recovery section; 231. Inner tube; 232. Outer tube; 233. Shell-and-tube heat exchanger; 234. Fin; 235. Baffle; 201. Through hole. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this 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 this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] In the description of the embodiments of the present invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of the present invention based on the specific circumstances.

[0024] In the description of the embodiments of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention.

[0025] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0026] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0027] The following disclosure provides numerous different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. Additionally, examples of various specific processes and materials are provided in this invention; however, those skilled in the art will recognize the applicability of other processes and / or the use of other materials.

[0028] like Figures 1-6 As shown, the steam trap optimized by the TEMOF-NSGA-II algorithm of the present invention includes a steam trap body 1 and an efficiency enhancement module 2.

[0029] Among them, the steam trap body 1 is a disc steam trap, and its end is seamlessly integrated with the efficiency enhancement module 2; The efficiency enhancement module 2 is a static integrated structure, which includes a Venturi tube section 21, a swirl flow stabilizing cavity section 22 and a sleeve-type heat recovery section 23 connected in series. The Venturi tube section 21 is configured to initially throttle and dissipate kinetic energy of the fluid to suppress cavitation; the swirling flow stabilizing cavity section 22 is configured to convert the fluid into a swirling flow of optimized intensity to promote vapor-liquid separation and reduce vibration and noise; the shell-and-tube heat recovery section 23 is configured to achieve efficient recovery of sensible heat with minimal flow resistance. Specifically, in the structure of the efficiency enhancement module 2, the first stage is the Venturi tube section 21, whose precisely designed throat area and contraction-diffusion angle are responsible for the initial pressure reduction and flow stabilization of the discharged fluid. Its core function is to suppress cavitation corrosion from the source by controlling the flow rate, thereby greatly extending the valve core life. The second stage is the swirling flow stabilization chamber section 22, which is equipped with guide vanes at a fixed angle. Its function is to guide the fluid to generate a stable swirling flow to promote vapor-liquid separation and droplet coalescence, and significantly reduce noise and pipeline vibration. The third stage is the shell-and-tube heat recovery section 23, which uses a shell-and-tube heat exchanger for heat exchange. The heat of the discharged medium is preheated and flows through the cooling working fluid in its jacket, thereby achieving efficient heat recovery. These three static stages are physically solidified into a single module. By connecting with the valve body, the traditional single drainage function of the steam trap is upgraded to a high-performance terminal integrating vibration damping, long service life, and energy saving without relying on any moving parts.

[0030] Among them, the Venturi tube section 21 is the inlet and flow control core of the entire static enhancement module 2. Its structure is not a traditional standard Venturi tube, but a fixed flow channel with a specific profile precisely calculated by a global optimization algorithm. It is directly connected to the outlet of the steam trap body 1, and its structure consists of three parts in sequence: a gradually converging section 211, a throat 212 with the smallest cross-sectional area, and a gradually expanding diffusion section 213.

[0031] The converging section 211 smoothly guides the acceleration of the steam-water mixture from the steam trap, converting pressure energy into kinetic energy. The throat 212 is the core of the entire structure; its diameter and length are optimized by an algorithm to ensure that the fluid reaches an optimal critical velocity—sufficient to effectively dissipate excess kinetic energy but below the threshold that would trigger severe cavitation corrosion. The subsequent diffuser section 213 decelerates the high-speed fluid, restoring some kinetic energy to pressure energy, thus achieving a smooth pressure transition and avoiding impact on downstream components. This meticulously designed structure ultimately achieves optimal management of fluid energy from the flow field source. Through this one-time, fixed structural design, it actively suppresses cavitation, reduces vibration, and protects downstream equipment, laying a solid foundation for the long-term stable operation of the entire steam trap.

[0032] Specifically, the area of ​​the throat 212 is calculated to obtain the optimal parameters. The design of the converging section 211, the throat 212, and the diffuser section 213 in the Venturi tube section 21 is determined based on the optimal throat area A. The convergence angle of the converging section 211, the length and diameter of the throat 212, and the diffusion angle of the diffuser section 213, together with the optimal throat area A, jointly achieve the optimization of the flow field.

[0033] The second-stage swirl flow stabilization chamber 22 is a key flow stabilization and separation unit at the venturi tube outlet. The swirl flow stabilization chamber 22 is a cylindrical cavity with optimal installation angles at both ends. Fixed guide vane 221; When the pre-treated vapor-liquid two-phase flow from the front-end Venturi tube section 21 enters this cavity tangentially, the fixed guide vanes 221 force the fluid to generate a stable swirling flow with a preset intensity. During this intense rotational motion, denser droplets are thrown towards the periphery of the cavity by centrifugal force and collide with the wall, coalescing into larger droplets that are then guided downstream along the wall. Meanwhile, the less dense vapor concentrates in the central axis region of the cavity, thus achieving efficient vapor-liquid separation. This process effectively breaks up any cavitation bubbles that may remain in the Venturi tube section 21 and greatly smooths the fluid pulsation. The effect is a significant reduction in fluid flow noise and suppression of pipe vibration. It ensures the flow capacity of the fluid channel and, with the help of the guide vanes 221, achieves efficient rectification and swirling control of the incoming flow. This is suitable for stabilizing and separating the vapor-water mixture after the steam trap, as well as for subsequent heat recovery processes, providing a stable swirling environment for the fluid processing flow.

[0034] The third-stage shell-and-tube heat recovery section 23 serves as the core structure to achieve efficient heat recovery. This device consists of concentrically mounted inner tubes 231 and outer tubes 232. The inner tube 231 acts as the main flow channel, directly receiving and passing through high-temperature condensate and a small amount of flash vapor processed by the preceding swirl-flow stabilizing chamber section 22. The annular jacket between the outer tube 232 and the inner tube 231 serves as the flow channel for the cooling medium. The inner tube 231 and outer tube 232 form a counter-current heat exchange channel. The material, wall thickness, spacing, and flow channel length of the shell-and-tube heat recovery section 23 are used to determine the optimal heat transfer coefficient U, which is the heat transfer coefficient of the shell-and-tube heat recovery section. Its working mechanism is counter-current heat exchange; the high-temperature medium flows through the inner tube 231, transferring its sensible heat through the inner tube 231 wall to the low-temperature cooling water flowing counter-currently in the jacket. This counter-current arrangement can form and maintain a large logarithmic mean temperature difference, thereby driving a strong heat transfer process and achieving higher heat recovery efficiency. This device serves as the terminal of the entire efficiency enhancement process, effectively transferring the heat of the emitted medium, thereby significantly improving the overall thermal economy of the system.

[0035] like Figures 1-6 As shown, a shell-and-tube heat exchanger 233 is installed in the inner tube 231 and the outer tube 232; Specifically, the outer tube 232 is provided with a number of fins 234. The fins, together with the inner wall of the outer tube 232 and the drain valve body 1, form a number of flow channels. At the same time, the outer tube 232 is provided with through holes 201 to connect the flow channels. Specifically, a baffle 235 is provided at the end of the outer pipe 232 to close one end of the flow channel, so that cooling water can flow in reverse between the inner pipe 231 and the outer pipe 232. The cooling water first exchanges heat with the medium transported in the valve through the inner pipe 231, and then exchanges heat with the medium flowing in the shell-and-tube heat exchanger 233 to achieve two-stage heat exchange. The shell-and-tube heat exchanger 233 is used to flow a medium with a better specific heat capacity to achieve efficient heat exchange. Finally, the cooling water flows through the flow channel formed between the outer pipe 232 and the drain valve body 1, where it exchanges heat with the valve body, thereby achieving full recovery and utilization of heat.

[0036] like Figure 3 As shown, in some embodiments, the drain valve body 1 has an outlet 101 corresponding to the converging section 211. In this example, the baffle 235 is not set to close the flow channel. In this way, after the cooling medium exchanges heat with the medium flowing in the shell-and-tube heat exchanger 233, it can continue to flow between the Venturi tube section 21 and the drain valve body 1, thereby exchanging heat with the Venturi tube section 21. This is beneficial to further recover heat and at the same time, to exchange heat and cool down the throat 212 with a high flow rate. This ensures heat recovery and also helps to protect the overall structural stability of the efficiency enhancement module 2.

[0037] Specifically, the throat area A of the Venturi tube section 21 and the installation angle of the guide vanes in the swirl flow stabilization cavity section 22. The heat transfer coefficient U of the shell-and-tube heat recovery section 23 is determined based on the following optimization design process; The execution process of the TEMOF-NSGA-II optimization algorithm includes: S1. Construct a multi-objective optimization problem with conflicting objectives of minimizing flow-induced vibration, minimizing flow pressure loss, and maximizing heat recovery efficiency. The design variables are throat area A and installation angle. and heat transfer coefficient U; S2. The TEMOF-NSGA-II two-stage framework is used to solve this optimization problem: In the first stage, a surrogate model is used to explore the global potential region and output the initial Pareto solution set; In the second stage, the initial Pareto solution set is used as the starting point, and the high-fidelity simulation model is switched to perform local refinement in the potential region until the convergence condition based on the hypervolume index is met, and the final Pareto optimal solution set is obtained. S3. Select the best compromise solution from the final Pareto optimal solution set, and select the corresponding throat area A and installation angle. And the heat transfer coefficient U, as the optimal combination of parameters.

[0038] In step S1, the multi-objective optimization problem is specifically defined as minimizing the objective vector. ,in, , and This refers to the throat area of ​​Venturi section 21; And θ is the installation angle of the guide vanes inside the swirl flow stabilization cavity section 22; and The overall heat transfer coefficient of the shell-and-tube heat recovery section 23; The vibration index is defined based on the root mean square of the pulsating pressure on the flow channel wall. The pressure loss index is defined based on the total pressure drop of the efficiency enhancement module 2; The heat recovery efficiency index is defined based on the heat exchange efficiency of the shell-and-tube heat recovery section 23.

[0039] Specifically, the algorithm defines the physical parameters to be optimized as a vector, representing the position of a candidate design in the search space. This vector represents the position of a candidate design in the search space. The three key adjustable physical parameters of the engineering system (throat area A, guide vane angle, etc.) are then used. The heat transfer coefficient (U) is mathematically represented as a set of design variables, and the physical feasible range of each variable is specified.

[0040] , In the above formula This represents a design variable vector, which represents a complete design scheme. and This represents the throat area of ​​the venturi tube, which is the most critical flow control dimension. And θ is the installation angle of the guide vane, which determines the swirling intensity of the fluid. and It is the overall heat transfer coefficient of the heat exchanger.

[0041] Constraints: ,

[0042] in and , and , and Equals are the upper and lower limits of the physical feasible domain for each variable, determined by factors such as processing capacity, safety, and physical laws.

[0043] In order to transform complex engineering decision problems into computable mathematical problems, a target vector needs to be constructed.

[0044] , In the above formula, It is a vibration index function defined based on the root mean square of the pulsating pressure on the flow channel wall, used to quantify the intensity of mechanical vibrations induced by fluid flow and minimize them. This reduces vibration and noise, extending equipment lifespan. It is a pressure loss index defined based on the total pressure drop of enhancement module 2, used to quantify the mechanical energy loss when fluid flows through the system and minimize it. This is to reduce the energy consumption of the system. This is a heat recovery efficiency index based on the heat exchange performance of the shell-and-tube heat recovery section 23. It is used to quantify the effectiveness of heat recovery by the heat exchanger by minimizing - To achieve equivalent maximization Improve energy efficiency This indicates minimization.

[0045] Then, an evaluation model was constructed: , in, It is the global trend constant, which can be understood as the best global estimate (average) of the objective function value when there is no sample information. In the Kriging model, it is a constant to be estimated. For Gaussian process terms, it represents the global trend. The local bias on the basis has a mean of 0, which means that the expected value of the bias is zero; its covariance structure defines spatial correlation.

[0046] Generally, the following conditions are met: , In the above formula, It is process variance, a measure The intensity of fluctuation is the maximum possible magnitude of change in the objective function value around the global trend. It is the correlation function, which is the soul of the Kriging model. It quantifies the similarity of function values ​​at two different design points. The closer the distance, the stronger the correlation (R→1).

[0047] The proxy model in this solution is the Kriging model, which is trained using high-fidelity simulation data obtained through Latin hypercube sampling. It is used to quickly predict approximate values ​​of the objective function under different design variables.

[0048] An initial population of size N is then randomly generated and enters the NSGA-II core loop.

[0049] Fast non-dominated sorting, predicted values ​​based on surrogate model The population is stratified. Based on the dominance formula:

[0050] , In the above formula, and Use the target index, then determine the individual. Dominate , Indicates the first The surrogate model predictions for each optimization objective. Indicates the first The surrogate model predicts the values ​​of the optimization objective. Non-dominated individuals form the first frontier F1. They are then removed, and new non-dominated individuals are found from the remaining individuals to form F2, and so on.

[0051] Crowding distance calculation: For individuals within the same frontal F1 region, calculate their crowding distance. To measure the distribution density. For individual i, in each objective function superior.

[0052]

[0053] in, and It is the first in the current cutting-edge Fl The maximum and minimum values ​​of each target. and Let i represent the neighboring individuals of individual i on target k, and set the crowding distance of boundary individuals to infinity.

[0054] Selection, crossover, and mutation steps: Selection: A binary tournament selection method is used. Individuals with lower non-dominant rank (frontier number) are prioritized; if ranks are the same, those with greater crowding distance are prioritized. Large individuals, to balance convergence and diversity.

[0055] Cross: Apply simulated binary cross to the selected parent. For the parent... The offspring are calculated for the i-th dimension variable as follows:

[0056] , , in, and This represents the values ​​of two offspring individuals in the d-th dimension variable. and This represents the values ​​of two parent individuals in the d-th dimension variable. Represents the crossover step size factor, which is... Distribution generation. In this stage, a smaller cross-distribution index is set to produce offspring with greater differences, enhancing global exploration capabilities.

[0057] Mutation: Apply polynomial mutation to the offspring. For variables After mutation, it becomes:

[0058] , in, This represents the mutated variable value. Represents the original variable value. For variable asynchronous long factor, by distribution generate, and These represent the upper and lower bounds of the variable's values, respectively. At this stage, a relatively small variation distribution index is also set to allow for a larger range of variation, thus promoting exploration.

[0059] Elite retention and merging involves merging the parent population Pt and the offspring population Qt into Rt. Rt is then subjected to a fast non-dominated sort and crowding distance calculation. The top N best individuals are then selected to form the next generation population Pt+1. This elite strategy ensures that superior individuals are not lost.

[0060] The first stage output, and the final output of the algorithm, is the initial Pareto solution set. This includes M widely distributed solutions that perform well under surrogate model evaluation. These solutions mark the most promising regions in the design space.

[0061] The second stage is local refinement. The goal of this stage is to use an accurate physical model to perform in-depth optimization within the potential region discovered in the first stage, in order to obtain a high-precision final Pareto optimal solution.

[0062] High-fidelity vibration objective function Based on the theory of flow-induced vibration, it can be concretized as the root mean square (RMS) integral of the pulsating pressure on the flow channel wall over a specific frequency band: , In the above formula, It is the pulsating pressure (Pa) at a key point on the flow channel wall, which is a physical quantity directly output by CFD transient calculation; It is the total time for statistical purposes, and it needs to be long enough to ensure statistical stability.

[0063] This formula quantifies the intensity of pressure fluctuations caused by vortex shedding, turbulence, etc. within the flow channel. The smaller the value, the more stable the flow field and the smaller the vibration excitation.

[0064] High-fidelity pressure loss objective function Defined as the total voltage drop flowing through the enhancement module: , In the above formula, It is the average total pressure (Pa) at the inlet section of the Venturi tube. It is the average total pressure (Pa) at the outlet section of the heat recovery jacket. The total pressure drop directly reflects the mechanical energy loss during the flow process. The smaller the value, the smaller the flow resistance and the lower the operating energy consumption.

[0065] The objective function for high-fidelity heat recovery efficiency is accurately calculated using the heat exchanger efficiency-number of heat transfer units method. , In the above formula, This represents the actual heat exchange (W). To maximize the possible heat exchange (W). It is the fluid with the lower heat capacity among cold and hot fluids. and It represents the inlet temperature (K) of the hot fluid (discharge medium) and the cold fluid (cooling medium). Efficiency η characterizes how close the actual performance of the heat exchanger is to its theoretical maximum performance; the larger the value (closer to 1), the higher the heat recovery efficiency.

[0066] Similarly, the objective function is constructed as follows: , In the formula, It is a vibration index function defined based on the root mean square of the pulsating pressure on the flow channel wall. It is a pressure loss index defined based on the total pressure drop of the efficiency enhancement module 2. It is a heat recovery efficiency index defined based on the heat exchange performance of the shell-and-tube heat recovery section 23. This indicates minimization.

[0067] The convergence criterion formula requires a more stringent convergence criterion in the second stage to determine whether the optimization can be terminated.

[0068] The improvement in hypervolume is an indicator of the overall performance of the Pareto front. Its improvement is an important convergence criterion. oversize of generation population Defined as:

[0069] , In the above formula, It is the Lebesgue measure, which calculates the volume of the hypercube in the target space. This is a reference point for the m-th objective function, which should be worse than all currently found solutions. , , Solution "Interval representation" over three objectives.

[0070] Among them, in the continuous L generations of evolution, the supervolume of the final Pareto solution set The relative improvement is less than the set threshold, resulting in over-volume. Defined as the volume enclosed by the solution set and the reference point in the target space.

[0071] Convergence criterion: If the hypervolume exceeds the limit in L consecutive generations... The relative improvement is less than the threshold. If the result is positive, then it is considered convergent. , in, Describes the k-th generation population super volume, This indicates the threshold for relative improvement in volume. Represents the kL generation population The hypervolume is used to compare the hypervolume changes before and after iteration. Solution set spacing: measures the uniformity of the Pareto solution distribution in the target space.

[0072] , in, , is the minimum distance from solution i to other solutions. It is all The average value. The smaller the value, the more uniform the distribution of the solution set. When the value stabilizes at a low level, it indicates that the shape of the leading edge has become stable.

[0073] In this scheme, in step S2, the TEMOF-NSGA-II algorithm uses the cross-distribution index during the first-stage global exploration. ∈[5,15] and the variation distribution index ∈[5,15], to enhance global search capability and avoid getting trapped in local optima; in the second stage of local refinement, the cross-distribution index is used. ∈[50,100] and the variation distribution index ∈[50,100], to achieve fine development of potential solutions.

[0074] In the local refinement stage of step S2, the TEMOF-NSGA-II algorithm employs a mutation operation with a mutation distribution exponent. ∈[50,100], such that the probability distribution exist The mutation operation reaches its peak at 0, and follows the following probability distribution: , in, It is the variation distribution index; The value represents the variation perturbation. Use larger Values ​​that make the probability distribution exist The peak value is reached at =0.

[0075] Specifically, in the second stage, the goal of crossover and mutation operations is to conduct a fine search around excellent individuals, so the formula parameter settings are different from those in the first stage.

[0076] Local development parameters for simulating binary cross (SBX); The probability distribution function of SBX is: , In the above formula, It is the cross-distribution index, and at this stage, a larger one is used. Value. This makes the probability distribution exist The proximity becomes very sharp, meaning that the generated offspring c has a very high probability of being very close to its parent. This effectively limits the search scope, enabling detailed local development.

[0077] The probability distribution function of polynomial mutation is: , In the above formula, It is the variation distribution index. Use a larger one. Value. This leads to the distribution. exist The peak value is reached at 0, indicating the random perturbation caused by the mutation operation. It is highly likely to be a minimum value close to zero. Therefore, the mutated individual c′ is very similar to the original individual c, achieving fine-tuning and facilitating a precise search near the optimal solution.

[0078] Final Decision and Parameter Output: After the convergence condition is met, the algorithm obtains the final Pareto solution set. An optimal compromise solution needs to be selected from this set as the final optimal static parameters. This patent employs a decision-making method based on weighted Chebyshev distance. This method can effectively handle differences in target dimensions and scale. The decision-making steps are as follows:

[0079] Normalized target value: for For each solution, calculate its normalized objective value: , In the above formula, and yes The Middle The extreme value of an objective.

[0080] Calculate the weighted Chebyshev distance: For each solution, calculate its weighted Chebyshev distance to the ideal point: , in, Let x represent the weighted Chebyshev distance from the solution x to the ideal point. The smaller the distance, the closer the solution is to the ideal point. This represents the ideal value of the m-th objective. This represents the normalization function of solution x on the m-th objective. It reflects the weights of preferences for each objective, and .

[0081] In this scheme, the optimal compromise solution is selected by calculating the weighted Chebyshev distances from each solution in the final Pareto optimal solution set to the ideal point, and the solution with the smallest distance is selected. Choosing the final solution: Select the solution with the minimum weighted Chebyshev distance. , Through the above decision-making process, the algorithm ultimately outputs a unique optimal combination of parameters: , These parameters serve as the final and optimal static dimension guide for manufacturing the steam trap enhancement module.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A steam trap optimized based on the TEMOF-NSGA-II algorithm, characterized in that: It includes a steam trap body (1) and an efficiency enhancement module (2) fixedly connected to the outlet end of the steam trap body (1). The efficiency enhancement module (2) is a static integrated structure, which includes a Venturi tube section (21), a swirl flow stabilizing cavity section (22), and a sleeve-type heat recovery section (23) connected in series. The throat area A of the Venturi tube section (21) and the installation angle of the guide vanes in the swirling flow stabilizing cavity section (22) The heat transfer coefficient U of the shell-and-tube heat recovery section (23) is determined based on the following optimization design process; The execution process of the TEMOF-NSGA-II optimization algorithm includes: S1. Construct a multi-objective optimization problem with conflicting objectives of minimizing flow-induced vibration, minimizing flow pressure loss, and maximizing heat recovery efficiency, wherein the design variables are the throat area A and the installation angle. and the heat transfer coefficient U; S2. The TEMOF-NSGA-II two-stage framework is used to solve this optimization problem: In the first stage, a surrogate model is used to explore the global potential region and output the initial Pareto solution set; In the second stage, the initial Pareto solution set is used as the starting point, and the high-fidelity simulation model is switched to perform local refinement in the potential region until the convergence condition based on the hypervolume index is met, and the final Pareto optimal solution set is obtained. S3. Select the best compromise solution from the final Pareto optimal solution set, and select the corresponding throat area A and the installation angle. The heat transfer coefficient U is used as the optimal parameter combination.

2. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 1, characterized in that: In step S1, the multi-objective optimization problem is specifically defined as minimizing the objective vector. ,in, , and The throat area of ​​the Venturi tube segment (21); And θ is the installation angle of the guide vanes in the swirling flow stabilization cavity section (22); and The overall heat transfer coefficient of the shell-and-tube heat recovery section (23); The vibration index is defined based on the root mean square of the pulsating pressure on the flow channel wall. The pressure loss index is defined based on the total pressure drop of the enhancement module (2); The heat recovery efficiency index is defined based on the heat exchange performance of the shell-and-tube heat recovery section (23). This indicates minimization.

3. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 2, characterized in that: In step S2, the surrogate model is a Kriging model, which is trained by obtaining high-fidelity simulation data through Latin hypercube sampling and is used to quickly predict the approximate value of the objective function under different design variables.

4. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 1, characterized in that: In step S2, the convergence condition based on the hypervolume index is: in L consecutive generations of evolution, the hypervolume of the final Pareto solution set... The relative improvement is less than a set threshold, the supervolume Defined as the volume enclosed by the solution set and the reference point in the target space.

5. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 1, characterized in that: In step S2, the TEMOF-NSGA-II algorithm uses the cross-distribution index during the first stage of global exploration. ∈[5,15] and the variation distribution index ∈[5,15], the cross-distribution index used in the second stage of local refinement. ∈[50,100] and the variation distribution index ∈[50,100].

6. The condensate trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 5, characterized in that: In the local refinement stage of step S2, the mutation operation employed by the TEMOF-NSGA-II algorithm has a mutation distribution index. ∈[50,100], such that the probability distribution exist The mutation operation reaches its peak at 0, and follows the following probability distribution: , in, It is the variation distribution index; This represents the variation perturbation value.

7. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 1, characterized in that: In step S3, the optimal compromise solution is selected by calculating the weighted Chebyshev distance from each solution in the final Pareto optimal solution set to the ideal point, and the solution with the smallest distance is selected.

8. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 1, characterized in that: The Venturi tube section (21) is configured to perform initial throttling and kinetic energy dissipation on the fluid to suppress cavitation; the swirling flow stabilizing cavity section (22) is configured to convert the fluid into a swirling flow of optimized intensity to promote vapor-liquid separation and reduce vibration and noise; the shell-type heat recovery section (23) is configured to achieve efficient recovery of sensible heat with minimal flow resistance.

9. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 8, characterized in that: The Venturi tube section (21) has a tapering convergence section (211), a throat (212), and a diffuser section (213) determined according to the optimal throat area A. The convergence angle of the tapering convergence section (211), the length and diameter of the throat (212), and the diffusion angle of the diffuser section (213) work together with the optimal throat area A to optimize the flow field.

10. The steam trap optimized based on the TEMOF-NSGA-II algorithm as described in claim 8, characterized in that: The swirling flow stabilizing cavity section (22) is a cylindrical cavity, with its two ends having the optimal installation angle. Fixed guide vanes (221); The shell-and-tube heat recovery section (23) includes an inner tube (231) and an outer tube (232) arranged coaxially, and a shell-and-tube heat exchanger (233) located in the inner tube (231) and the outer tube (232). The inner tube (231) and the outer tube (232) form a counter-current heat exchange channel. The material, wall thickness, spacing and flow channel length of the shell-and-tube heat recovery section (23) are optimally determined by the heat transfer coefficient U.

Images