Design method and device for low-noise screw profile of twin-screw pump based on minimum flow entropy production

Abstract

This invention discloses a method and apparatus for designing low-noise screw profiles for twin-screw pumps based on minimizing flow entropy production. After receiving the operating parameters of the twin-screw pump, fluid medium properties, and noise control targets, the method performs refined simulation of the flow field during screw meshing using a flow entropy production analysis model. This identifies key regions with high entropy production and high noise, and automatically generates an optimized profile that combines low entropy production and low noise characteristics. This method integrates computational fluid dynamics, entropy production theory, and parametric profile generation technology, combined with dynamic flow field monitoring, noise spectrum analysis, and iterative profile verification mechanisms. It not only significantly reduces internal flow losses and eddy noise in the twin-screw pump but also improves the pump's operating efficiency and stability. This has significant engineering application value in fields with stringent requirements for noise and energy efficiency, such as petrochemicals and marine propulsion.

Description

Technical Field

[0001] This invention relates to the field of fluid machinery design and noise control, and in particular to a method and apparatus for designing a low-noise screw profile for a twin-screw pump based on minimizing flow entropy production. Background Technology

[0002] Twin-screw pumps are widely used in the petroleum, chemical, shipbuilding, and food industries due to their advantages such as stable delivery, wide media adaptability, and low pressure pulsation. They exhibit unique advantages in conveying high-viscosity and particulate media, especially in critical applications such as ship fuel supply and chemical raw material transportation, where their operational stability directly affects the safety and reliability of the entire system. However, the hydrodynamic noise generated during their operation is a major source of noise pollution, which not only deteriorates the working environment and harms the health of operators but may also cause equipment resonance and shorten the service life of the pump set. This has become a key bottleneck restricting their further promotion in noise-sensitive scenarios (such as ship propulsion systems, precision chemical production lines, and medical fluid delivery equipment).

[0003] This type of hydrodynamic noise mainly originates from irreversible flow processes such as periodic flow pulsations caused by screw meshing, fluid impact within the cavity, and shear mixing between gap leakage flow and the mainstream. These processes are essentially entropy increase (entropy production) processes in the flow field. Energy losses caused by fluid viscosity dissipation, flow separation, and turbulent pulsations are directly converted into noise energy. Therefore, the magnitude of entropy production and the intensity of flow noise have an inherent physical relationship. High entropy production regions often correspond to severe flow disturbances and noise radiation sources.

[0004] In existing technologies, traditional noise reduction methods mostly focus on optimizing gear parameters, improving machining accuracy, or adding soundproof covers. These are passive or post-processing methods that fail to address the root cause of the core component—the screw rotor—by optimizing the mechanism of flow noise generation. Although computational fluid dynamics (CFD) simulation has been used for pump flow field analysis, it typically uses macroscopic indicators such as efficiency and flow pulsation amplitude as optimization targets. It lacks a physical quantity that can directly and comprehensively characterize the irreversibility of flow and the potential for noise generation as an optimization target, making it difficult to fundamentally achieve a synergistic improvement in low noise and high efficiency.

[0005] Meanwhile, existing screw profile designs are mostly based on geometric modifications of classic tooth profiles such as cycloidal and circular arcs, failing to establish a direct mapping relationship between profile parameters and flow field entropy production, thus making it impossible to accurately locate flow regions with high entropy production and high noise. This results in designed profiles that often fail to meet the dual requirements of flow loss suppression and noise control: excessive pursuit of efficiency may lead to a surge in local flow field entropy production, which in turn exacerbates noise radiation; simply reducing flow pulsation may sacrifice pump volumetric efficiency. This contradiction is particularly pronounced under high-speed and high-pressure conditions, leaving considerable room for optimization in existing designs for engineering applications. Summary of the Invention

[0006] The problem to be solved by this invention is to provide a low-noise screw profile design method and device for twin-screw pumps based on minimizing flow entropy production. Starting from the second law of thermodynamics, the method uses the flow entropy production rate as the direct optimization target to design a screw profile that can suppress irreversible flow losses from the source, thereby significantly reducing hydrodynamic noise and promoting the development of twin-screw pumps towards low noise and high reliability.

[0007] This invention adopts the following technical solution: a low-noise screw profile design method for a twin-screw pump based on minimizing flow entropy production, comprising the following steps:

[0008] S1. Establish a parameterized screw profile model. Based on the cycloidal-circular arc combined tooth profile theory, define key geometric parameters and construct profile generation equations that support dynamic adjustment of key geometric parameters.

[0009] S2. Construct a CFD simulation model with flow entropy production rate as the core calculation index, and numerically simulate the three-dimensional transient flow field in the working chamber of the twin-screw pump to capture the flow state of the fluid in the entire process of meshing, compression and discharge.

[0010] S3. Establish a multi-objective optimization function with minimizing entropy production as the primary objective and taking into account efficiency and pressure fluctuations. Normalize and weight the average total entropy production rate, volumetric efficiency and the amplitude of outlet pressure fluctuations to form a comprehensive optimization objective.

[0011] S4. Utilize intelligent optimization algorithms for automatic iterative optimization. By iteratively updating key geometric parameters and driving the CFD simulation model to iteratively calculate, the screw profile parameters that optimize the multi-objective optimization function value are obtained.

[0012] As a preferred option, in step S1, an initial model of the parameterized screw profile is established:

[0013] Based on the original screw profile (such as cycloidal, involute, or a combination thereof), key control parameters are selected to establish a controllable relationship between the parameterized model and the three-dimensional solid model. These key control parameters include, but are not limited to: addendum circle radius, dedendum circle radius, meshing angle, coordinates of control points on the tooth profile curve, and helix angle.

[0014] As a preferred embodiment, in step S2, the CFD simulation model includes:

[0015] S2.1 Parameter Modification: The parametric geometric modeling kernel is called to discretize the input initial profile. The coordinates of several key control points are selected as design variables to control the tooth tip radius, tooth root curve shape and meshing area profile. When the design variables change, the parametric engine can automatically generate new geometrically perfectly matched male and female rotor 3D solid models and automatically assemble and generate a fluid domain model containing the working cavity.

[0016] S2.2 Automatic Modeling: Establish a full three-dimensional transient computational fluid dynamics (CFD) model of the working chamber of the twin-screw pump, and set accurate dynamic mesh and meshing motion;

[0017] The fluid domain model is passed to the mesh generation module. Based on the medium viscosity and rotor speed, a fine prismatic layer mesh is generated near the rotor wall using a boundary layer mesh strategy to capture high-shear flow.

[0018] The number of meshes is controlled by a polyhedral mesh inside the chamber. The mesh quality standard is preset in the mesh generation module. Unqualified meshes trigger automatic re-rendering.

[0019] S2.3 Automatic Simulation: Call the CFD solver to perform unsteady flow simulation. The settings include: selecting the Realizable k-ε turbulence model suitable for viscous flow, setting the medium properties to the lubricating oil parameters defined in the structured task model, and setting the boundary conditions according to the design conditions; the solver automatically calculates the transient flow field within the complete meshing cycle.

[0020] S2.4 Automatic extraction of target values: The flow entropy productivity calculation module is embedded in the CFD simulation model. Based on the entropy productivity theory calculation formula, the flow entropy productivity is extracted from the flow field data and integrated to obtain the time-averaged total entropy productivity, thus quantifying the irreversible losses in the flow process.

[0021] Furthermore, the flow entropy production rate consists of two parts: entropy production caused by viscous dissipation and entropy production caused by thermal conduction.

[0022] For isothermal or near-isothermal pumping processes, viscous entropy production is the dominant term. The flow entropy production rate mainly refers to the entropy production rate caused by fluid viscous dissipation. The time-averaged total entropy production rate is obtained by dividing the pump working chamber by volume, and also includes the additional entropy production component caused by flow shear and turbulent fluctuations to fully reflect the energy loss characteristics of the flow field. Its volume fraction expression is:

[0023] ;

[0024] Where μ is the fluid dynamic viscosity, T is the fluid temperature, and u, v, w are velocity components. Integrating the volume of the working chamber yields the time-averaged total entropy production rate over one cycle.

[0025] As a preferred option, in step S3, a multi-objective optimization function is established:

[0026] The multi-objective optimization function, which combines the normalized weighted value of the time-averaged total entropy productivity, volumetric efficiency, and the amplitude of export pressure fluctuations, is defined as follows:

[0027] ;

[0028] in, The objective function value, Let the initial entropy production rate be , For volumetric efficiency, This represents the amplitude of the outlet pressure pulsation; the subscript 0 represents the corresponding value of the initial profile. Let be the weighting coefficient, satisfying .

[0029] In particular, priority is given to ensuring It accounts for a relatively large proportion (generally not less than 0.5) to ensure the core design goals of low noise and low loss.

[0030] As a preferred option, in step S4, automated collaborative optimization is implemented, and the optimal profile is determined and verified:

[0031] The parametric model, CFD entropy generation simulation, and optimization objective function are integrated into an optimization platform, and intelligent optimization algorithms (such as genetic algorithms and particle swarm optimization) are used to automatically iterate and find the optimal solution. In each iteration, the algorithm generates a new set of profile parameters, automatically updates the model, runs transient CFD simulations, and calculates the objective function. The value is calculated until the convergence condition is met (e.g., ...). The value no longer decreases significantly or reaches the maximum number of iterations.

[0032] Furthermore, the output makes the objective function The optimal profile parameter combination was determined. Independent, high-precision, full-condition CFD verification simulations were performed on the optimal profile, and a prototype was fabricated for fluid noise sound pressure level testing. The noise reduction effect and performance indicators were confirmed by comparing the prototype with the initial profile.

[0033] The present invention also provides: a low-noise twin-screw pump device, designed by the aforementioned method, characterized in that it includes:

[0034] Operation ticket parsing module (corresponding profile parameterization module): It is used to receive the geometric parameter input of the initial profile, parse and extract key control variables, and generate a structured input model containing parameters such as tooth tip circle radius, meshing angle, and control point coordinates, supporting rapid iteration of multiple profiles.

[0035] Sequential control logic generation module (corresponding to flow field entropy production calculation and optimization module): Based on structured parameters and CFD simulation model, it calls the preset entropy production calculation template and intelligent optimization algorithm to dynamically generate an optimization process chain that includes profile parameter update, flow field simulation and objective function evaluation;

[0036] Interlocking verification module (corresponding to convergence and verification module): The rule engine is used to perform multi-dimensional verification of the optimization results, including entropy yield convergence judgment, volume efficiency threshold verification and pressure pulsation amplitude evaluation, and the final optimal profile scheme is formed after the simulation verification is passed.

[0037] Closed-loop control execution module (corresponding to prototype verification module): The screw rotor is manufactured according to the optimal profile scheme, and the noise, efficiency and pressure pulsation data of the prototype are collected in real time. The data are compared with the initial profile, and automatic feedback is given to trigger secondary optimization iteration to complete the closed-loop verification of the whole process.

[0038] As a preferred embodiment, the end face tooth profile of the screw rotor of the low-noise twin-screw pump device of the present invention is designed using the method described in any one of claims 1-3. This profile can be directly applied to the processing and manufacturing of screw rotors with equal pitch or variable pitch.

[0039] As a preferred embodiment, the tooth profile of the screw rotor has a smooth and continuous curvature change to reduce flow separation; the tooth profile shape in its meshing area is modified to smooth the fluid transport process; and its helix angle is finely adjusted axially to balance the pressure gradient, thereby effectively reducing the flow noise and vibration level in the pump.

[0040] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0041] 1. Root Cause Noise Reduction: This invention is the first to take the "flow entropy yield"—a thermodynamic parameter that directly characterizes the irreversibility of flow and energy dissipation (i.e., the noise generation potential)—as the core optimization target for screw profile design, thereby achieving root cause suppression of noise generation from the perspective of the noise generation mechanism.

[0042] 2. Comprehensive performance optimization: This invention uses a multi-objective function to minimize entropy production (low noise) while taking into account the pump's volumetric efficiency and pressure pulsation, thus avoiding the drawback of sacrificing other key performance aspects for noise reduction.

[0043] 3. Scientific and efficient design method: This invention combines parametric modeling, high-precision transient CFD and intelligent optimization algorithms to form an automated and quantitative modern design process, overcoming the blindness and limitations of traditional experience-based design.

[0044] 4. Significant device effect: The screw pump designed by the method of this invention can significantly reduce its hydrodynamic noise (especially in the mid-to-high frequency range), while having good comprehensive hydraulic performance, meeting the stringent requirements of high-end industrial applications for low noise. Attached Figure Description

[0045] Figure 1 This is an overall flowchart of the low-noise screw profile design method for the twin-screw pump of the present invention;

[0046] Figure 2 This is a schematic diagram of the CFD calculation domain and mesh of the working chamber of a twin-screw pump according to an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the application will be further described in detail below with reference to the accompanying drawings. The described embodiments are only a part of the embodiments involved in this invention. All non-innovative embodiments based on these embodiments by other researchers in the art are within the protection scope of this invention. Furthermore, the step numbers in the embodiments of this invention are only set for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0048] Example 1

[0049] This paper presents a method for designing a low-noise screw profile for a twin-screw pump based on minimizing flow entropy production. The paper also describes the application process, parameter settings, and verification results of this method in a specific product, aiming to fully disclose the technical details so that those skilled in the art can reproduce and implement it.

[0050] The overall flow of the method in this embodiment is as follows: Figure 1 As shown, it includes the following steps:

[0051] S1. Establish a parameterized screw profile model. Based on the cycloidal-circular arc combined tooth profile theory, define key geometric parameters such as the addendum circle, dedendum circle, and meshing point coordinates, and construct profile generation equations that support dynamic parameter adjustment.

[0052] Specifically, in some implementations, receiving initial design parameters and performance targets, analyzing pumping medium properties, design conditions, and noise constraints, and generating a structured design task model are key preliminary steps in initiating the design method of this invention. This step aims to transform fuzzy noise reduction requirements into precise quantitative indicators that can be identified and processed by the optimized system, providing clear input for subsequent parametric modeling and objective function construction.

[0053] The system first receives input from the designer, which typically includes basic pump specifications (such as center distance, screw length, and lead), design operating points (such as rated flow rate, outlet pressure, and speed), and detailed physical properties of the pumped medium (such as the density, dynamic viscosity, and specific heat capacity of the lubricating oil). Simultaneously, noise constraint targets must be clearly defined, such as "A-weighted sound pressure level at 1 meter not exceeding 85 dB" or "reducing the sharpness of fluid dynamic noise."

[0054] Then, the system standardizes this unstructured input information using a built-in media database and operating condition parser. For example, the media viscosity is interpolated based on its temperature-viscosity curve to obtain an accurate value at the design temperature; noise targets are decomposed into indirect constraints on source indicators such as pressure pulsation amplitude and flow fluctuation.

[0055] The core of this step is to construct a structured design task model containing all key variables. This model, in JSON or XML format, defines the following data structures: a set of screw geometry parameters (including the number of screw threads, tooth profile type identifier, etc.), a set of operating parameters (including flow rate, pressure, speed, and media properties), a set of optimization objectives and weights (such as entropy yield target weights, efficiency maintenance weights, and pressure pulsation weights), and a set of constraints (such as lower limits for volumetric efficiency, minimum tooth tip clearance, and lower limits for radius of curvature). The system will verify the completeness and rationality of the input data, for example, checking whether the medium viscosity is within the effective operating range of the screw pump, or whether the target pressure matches the screw strength.

[0056] This step is widely used in the noise reduction design of various positive displacement pumps. For example, when designing a low-noise screw pump for marine fuel delivery, the system needs to analyze the high viscosity characteristics of marine heavy oil and the strict space and noise constraints in the engine room, thereby generating a design task model that emphasizes low pulsation and high viscosity adaptability. As another example, in the food industry, screw pumps used to transport chocolate syrup need to focus on the high viscosity and shear sensitivity of the medium, as well as special requirements for cleanliness (such as avoiding sharp edges), while their noise targets are coupled with factory environmental noise standards.

[0057] This step achieves precise digitization and structuring of design intent, transforming complex engineering problems into explicit mathematical problems that optimization algorithms can handle. This lays a solid foundation for subsequent automated simulation-driven optimization and effectively avoids deviations in optimization direction caused by ambiguity in design objectives.

[0058] S2. Construct a CFD simulation model with flow entropy production rate as the core calculation index, and numerically simulate the three-dimensional transient flow field in the working chamber of the twin-screw pump to capture the flow state of the fluid during the entire process of meshing, compression and discharge.

[0059] A parametric geometry engine is used to construct a 3D model of the screw rotor, and a CFD simulation environment is integrated to establish an automated analysis process with flow entropy yield as the core evaluation index.

[0060] Specifically, in some implementations, based on a structured design task model, a parametric geometry engine is invoked to construct a three-dimensional model of the screw rotor, and a CFD simulation environment is integrated. An automated analysis process is established with flow entropy yield as the core evaluation index; this is the core technical aspect of this method. This step achieves a fully automated closed loop from design parameters to performance prediction. Its key lies in linking geometric modeling, mesh generation, flow field calculation, and post-processing analysis into a digital workflow that requires no manual intervention.

[0061] The system first invokes the parametric geometry modeling kernel (e.g., via Creo's API or ANSYS DesignModeler script). This embodiment uses a horizontal twin-screw pump that pumps ISO VG46 lubricating oil as an example; its initial profile is the industry-standard cycloidal-involute combination profile. To achieve flexible profile optimization, the system discretizes the screw end face profile and parametrically selects the coordinates of six key control points as design variables. These variables are sufficient to control the tooth tip radius, tooth root curve shape, and the profile of the meshing zone. When the design variables change, the parametric engine automatically generates entirely new, geometrically perfectly matched three-dimensional solid models of the male and female rotors and automatically assembles and generates a fluid domain model containing the working chamber.

[0062] The system then seamlessly transfers the fluid domain model to the mesh generation module (such as Fluent Meshing). Based on the medium viscosity and rotor speed, the system automatically applies a boundary layer mesh strategy, generating a fine prismatic layer mesh near the rotor wall to accurately capture high-shear flows; inside the chamber, a highly adaptable polyhedral mesh is used to control the number of meshes while ensuring accuracy. Mesh quality standards (such as distortion and aspect ratio) are preset in the workflow, and unqualified meshes will trigger automatic remapping.

[0063] The core of the process is the automatic invocation of a CFD solver (such as Fluent) to perform unsteady flow simulation. The settings include: selecting a Realizable k-ε turbulence model suitable for viscous flow, setting the medium properties to the lubricating oil parameters defined in the structured task model, and setting boundary conditions according to the design conditions (such as flow inlet and pressure outlet). The solver automatically calculates the transient flow field over a complete meshing cycle. Most importantly, after the simulation is complete, the system automatically executes a post-processing script to extract and integrate the time-averaged total entropy production rate from the flow field data based on the entropy production theory formula. This indicator directly quantifies the irreversible losses (including viscous dissipation and turbulent dissipation) in the flow process and is established as the core objective function for optimization.

[0064] This step is widely used in the optimization design of various positive displacement pumps, such as screw pumps, gear pumps, and piston pumps. For example, when optimizing a screw pump for conveying high-viscosity polymers, the automated process can quickly assess the impact of different profiles on polymer shear heat generation (manifested as entropy production). Similarly, in the design of medical peristaltic pumps that prioritize extreme quietness, this method can calculate entropy production by simulating the interaction between tubing deformation and the fluid, thereby optimizing the roller profile and compression.

[0065] This step, by establishing a complete chain of parameter modification, automatic modeling, automatic simulation, and automatic extraction of target values, compresses the trial-and-error process that originally required several days and relied on a lot of human experience into several hours. Furthermore, it shifts the optimization objective from indirect noise and efficiency to the more fundamental and sensitive flow irreversibility (entropy production), thus achieving a revolution in the design paradigm.

[0066] S3. Establish a multi-objective optimization function with minimizing entropy production as the primary goal and taking into account efficiency and pressure fluctuations. Normalize and weight the average total entropy production rate, volumetric efficiency, and outlet pressure fluctuation amplitude to form a comprehensive optimization objective.

[0067] Specifically, in some implementations, constructing a comprehensive objective function, integrating optimization algorithms to drive iterative updates of design variables, and automatically finding novel linear geometry parameters that minimize flow entropy production are key decision-making steps in achieving intelligent design in this invention. This step, through scientific trade-offs among multiple objectives and efficient search in a broad variable space using intelligent algorithms, ultimately finds the globally or locally optimal screw profile scheme.

[0068] Based on the automated analysis process established in S2, the system constructs a weighted comprehensive objective function F. In this embodiment, F consists of three sub-objectives: the core objective is the average total entropy yield (S), with a weighting coefficient ω1 set to 0.6, reflecting the guiding principle of suppressing noise from the source of the flow mechanism; the second objective is the outlet pressure pulsation amplitude (ΔP), with a weighting ω2 of 0.25, which is directly related to the intensity of fluid radiated noise; the third objective is the retention of volumetric efficiency (ηv), with a weighting ω3 of 0.15, to ensure that the optimized performance is not lower than the engineering acceptable range (e.g., efficiency decrease does not exceed 2%).

[0069] The optimization solver (such as the genetic algorithm module in ANSYS DesignXplorer) is integrated into the workflow. The system sets the genetic algorithm parameters to: population size 30, 50 generations. During algorithm initialization, 30 sets of control point coordinates that meet the variable boundary constraints are randomly generated (i.e., 30 different individuals). Subsequently, the optimization engine automatically substitutes these 30 sets of parameters into the automated analysis process of S2 to calculate the comprehensive objective function value F for each individual. Based on the principle of "survival of the fittest," the algorithm selects individuals with smaller F values ​​for crossover and mutation to generate a new generation of the population. This process is repeated for 50 generations, continuously improving the overall fitness (i.e., comprehensive performance) of the population.

[0070] This step can be flexibly applied to various complex fluid machinery optimization scenarios. For example, when optimizing an air compressor screw for a fuel cell system, in addition to entropy production and efficiency, the objective function should also include an increase in outlet temperature as a sub-objective to prevent overcompression from causing excessively high gas temperatures. Similarly, in optimizing a screw pump used to transport easily vaporized liquids in a chemical process, minimizing the volume of local low-pressure zones (obtained through flow field analysis) is crucial to prevent cavitation, and this should be included as an important sub-objective in the function F.

[0071] Through this step, the system no longer relies on the designer's local experience and trial and error, but can systematically explore the design space corners that are difficult to reach with traditional methods. The combination of intelligent optimization algorithms and high-fidelity CFD simulation makes it possible to find a "balanced optimal solution" that simultaneously satisfies low entropy, low pulsation, and high efficiency, realizing a leap in automation and intelligence in the design process.

[0072] S4. Utilize intelligent optimization algorithms for automatic iterative optimization. By iteratively updating the profile parameters and driving CFD simulation iterative calculations, the screw profile parameters that optimize the multi-objective optimization function value are obtained.

[0073] Specifically, in some implementations, a comprehensive performance verification and prototype validation of the optimal profile obtained through optimization is the final confirmation step in the closed-loop process of this method. This step aims to confirm the performance improvement of the new profile obtained based on entropy production optimization, especially the reduction in noise level, through multi-dimensional and higher-precision verification methods. By comparing key acoustic and hydraulic indicators, the noise reduction effect is confirmed, and a final design report is generated, forming an authoritative design output.

[0074] The system first performs detailed CFD verification simulation on the optimal profile recommended by the optimization algorithm. In this embodiment, the simulation uses a finer mesh and a smaller time step to obtain more accurate data. The results of this embodiment show that compared with the initial profile, the radius of curvature of the tooth tip transition curve of the optimal profile increases by approximately 15%, and the tooth profile near the meshing point is more "fuller". CFD verification data shows that under the design conditions, the time-averaged total entropy production rate is reduced by 22%, which indicates a significant reduction in the intensity of the flow-induced noise source; at the same time, the outlet pressure pulsation amplitude is reduced by 18%, directly corresponding to a reduction in fluid radiation noise; while the volumetric efficiency only decreases by 1.5%, which is well within the engineering allowable range, achieving the main objectives.

[0075] Subsequently, a physical prototype was fabricated based on the optimal profile drawings, and its performance and noise levels were tested on a standard pump test bench. Hydraulic performance tests verified the accuracy of the efficiency predictions. Noise tests were conducted in a semi-anechoic chamber, with the microphone positioned 1 meter from the pump surface. Test results showed that the A-weighted sound pressure level of the optimal profile prototype was 4.5 dB lower than that of the original baseline prototype. Most importantly, subjective listening evaluations indicated a significant reduction in the sharpness of the optimized pump's noise, resulting in a deeper and more even sound quality, which is crucial for improving the comfort of the operator's working environment.

[0076] This verification model is widely used for final confirmation before new product finalization. For example, after optimizing the profile of a silent lubrication pump for high-end machine tools, in addition to laboratory testing, it needs to be installed in a simulated machine tool enclosure for near-field noise and vibration testing to confirm the noise reduction effect in a real enclosed space. Similarly, when optimizing and validating a water pump for a residential water supply system, special attention must be paid to its noise performance under various partial load conditions to ensure low noise characteristics are maintained throughout the entire operating range.

[0077] This step completes the entire closed loop from "digital twin" optimization to "physical entity" verification. It not only provides quantitative data to prove the effectiveness of the method, but the detailed test reports and comparative data (CFD results, test curves, noise spectrum diagrams) generated also constitute the final design outputs and intellectual property evidence, providing solid support for the credibility and engineering value of the design method.

[0078] Example 2

[0079] This invention provides a design method for a high-pressure, low-leakage aviation fuel plunger pump slipper-swashplate pair based on flow entropy production control. It demonstrates the application of this method in the friction pair design of another key fluid power component—the aviation plunger pump. The core objective is to minimize entropy production within the lubricating film by optimizing the micro-shape of the slipper bottom cavity, thereby achieving lower leakage, better lubrication, and higher efficiency under extreme high-pressure conditions.

[0080] S1. Receive high-pressure operating conditions and sealing requirements, analyze aviation fuel characteristics and extreme operating conditions, and generate a structured optimization task model for the slipper-swashplate friction pair.

[0081] Specifically, in some implementations, for high-pressure aviation fuel plunger pumps, receiving their extreme operating conditions and performance requirements and generating a structured task model is a prerequisite for the successful application of this design method. This step requires transforming stringent engineering constraints into precise boundaries and evaluation systems that can be handled by optimization algorithms.

[0082] The system receives design inputs including: pump rated pressure (e.g., 35 MPa), speed, viscosity-temperature-pressure characteristics of aviation fuel (e.g., JP-8), maximum allowable leakage rate, and limits on the operating temperature of the friction pair (to prevent fuel coking). The system analyzes this information to construct a structured task model. The core design variables of this model are defined as the geometric parameters of the hydrostatic support cavity on the slipper bottom, such as cavity depth, width, and the coordinates of the control points of the return oil groove. The objective set focuses on the total entropy production rate of the lubrication film region (directly related to frictional power consumption and temperature rise), the fuel leakage from the high-pressure chamber to the return oil chamber in the housing, and the minimum oil film thickness on the slipper bottom (to ensure no dry friction occurs). Constraints include structural strength (e.g., maximum stress on the slipper bottom) and manufacturability (e.g., minimum groove width).

[0083] This step needs to adapt to the high reliability requirements of the aviation industry. For example, when designing a skid for a fuel pump of a certain type of UAV, the model needs to additionally consider the stability of the lubricating film under high overload maneuvering conditions, treating this as an implicit constraint. Another example is when optimizing a plunger pump for high-altitude, low-temperature environments; the mission model must integrate the changes in fuel viscosity and vaporization characteristics at low temperatures and low pressures, setting different optimization operating points.

[0084] This step transforms the design challenge of friction pairs under high pressure, high speed, and low viscosity media into a parameter optimization problem with clear boundaries, with flow irreversibility (entropy production) as the core monitoring indicator, paving the way for subsequent refined simulation optimization.

[0085] S2. Establish a joint simulation model that includes microscopic lubrication film and macroscopic leakage path, integrate parametric modeling, structured mesh generation and multiphase flow solver, and automatically extract entropy production and leakage rate.

[0086] Specifically, in some implementations, a highly integrated, automated co-simulation process needs to be established to accurately simulate the microscale flow on the bottom of the slipper. This process can simultaneously analyze the shear flow within the hydrostatic support oil film and the leakage flow from the high-pressure chamber to the low-pressure chamber, and accurately calculate the entropy production contribution of both.

[0087] The system first automatically generates a parametric 3D model based on the design variables of the slipper bottom cavity. Then, it creates a fluid domain containing the following key areas: the high-pressure plunger cavity, the micron-level lubrication gap (with optimized cavity) between the slipper bottom and the swashplate, the return oil groove, and the low-pressure zone of the casing. Meshing is extremely challenging; the system employs a partitioning strategy: in the micron-level oil film region, a refined hexahedral structured mesh with only 3-5 cells in the height direction is used to accurately resolve the velocity gradient; unstructured meshes are used in areas such as leakage channels. The mesh generation module can automatically adjust the mesh density according to changes in oil film thickness.

[0088] The CFD simulation was set to transient calculation, considering the rotational motion of the swashplate (dynamic mesh technique). The medium properties were set to aviation fuel, and its physical property model as a function of pressure and temperature was enabled. The solver was selected with a specific scheme capable of handling thin-layer flow, and the energy equation was activated to calculate the temperature field. After the simulation, the post-processing script automatically integrated the total entropy production rate from the entire computational domain (focusing on the lubrication film and leakage channels), while simultaneously calculating the transient and average fuel leakage rates through a specified cross-section.

[0089] This co-simulation method can be extended to various friction pair optimization scenarios. For example, when optimizing the distribution plate friction pair of a seawater hydraulic piston pump, the model needs to consider the low viscosity and high vaporization pressure characteristics of water, and may incorporate a cavitation model to evaluate the additional entropy production caused by cavitation. When optimizing the slipper of a high-pressure axial piston motor used in large engineering machinery, it is necessary to simulate larger overturning moments and more complex load cycles.

[0090] This step, by establishing a cross-scale automated analysis process, enables rapid and quantitative assessment of the complex fluid-thermal-solid coupling effects of friction pairs, making refined optimization with entropy production as the target possible.

[0091] S3. Construct a multi-objective function with minimizing entropy production as the core, while taking into account leakage and load capacity, and use an optimization strategy combining response surface methodology and gradient algorithm for optimization.

[0092] Specifically, in some implementations, constructing efficient multi-objective optimization strategies for the slipper optimization problem is key to balancing load-bearing capacity, leakage, and triboelectric power consumption (manifested as entropy production). Since the number of design variables is relatively small but simulation costs are high, this embodiment employs a combination of response surface methodology (RSM) and sequence optimization.

[0093] The system first selects approximately 50 sample points within the design space using the optimal Latin hypercube method. Simulation is then performed via S2's automated process to obtain entropy production rate, leakage rate, and minimum oil film thickness data for each sample point. Based on this data, a second-order response surface surrogate model is constructed for the three objectives. In terms of weighting, minimizing entropy production rate is given the highest weight (e.g., 0.5), as it directly determines frictional power consumption and thermal balance; minimizing leakage rate has the next highest weight (0.35); and maximizing minimum oil film thickness has a weight of 0.15 as a safety guarantee.

[0094] Subsequently, the optimization algorithm rapidly searches the response surface model to find a set of Pareto optimal solutions. Designers can then select the solution that best suits their engineering preferences from this set. To further improve accuracy, this optimal solution can be used as a starting point to initiate local fine-tuning based on gradient algorithms, with a small number of high-precision CFD calculations used for verification.

[0095] This optimization strategy is widely used in high-fidelity designs that are sensitive to computational resources. For example, when optimizing the sealing tooth profile of turbomachinery, a similar strategy can be adopted: first, global exploration of the response surface, and then local optimization using the gradient method, to achieve a balance between controlling leakage (entropy production reflects flow losses) and avoiding rotor dynamic instability.

[0096] Through this step, the system efficiently explored the design space with a limited number of simulations, and found a micro-shape for the bottom surface of the slipper that can significantly reduce irreversible losses (entropy production) within the lubricating film, thus providing an innovative design solution for reducing pump temperature rise and improving mechanical efficiency.

[0097] S4. Verify the performance of the optimized slipper through bench testing, and confirm the friction reduction and leakage reduction effect by comparing key indicators.

[0098] Specifically, in some implementations, rigorous bench testing of the optimized slipper assembly is the final step in verifying its performance improvement under high pressure. The tests aim to demonstrate that the design based on entropy production optimization can deliver measurable engineering benefits.

[0099] The optimized slipper assembly was assembled into a complete plunger pump core package. Testing was conducted on a dedicated high-pressure pump test bench. Comparative tests showed that, under the same 35 MPa outlet pressure and rated speed, the pump with the optimized slipper achieved an overall efficiency increase of approximately 2.1 percentage points; the external leakage rate, measured by a precision flow meter, was reduced by approximately 30%; more importantly, measurements using near-surface thermocouples embedded in the slipper showed an 8-12°C decrease in operating temperature compared to the original design, directly confirming a reduction in entropy production (frictional heat generation) within the lubricating film. Durability tests demonstrated that the optimized slipper exhibited less wear.

[0100] Such validation is crucial for high-reliability applications. For example, when validating the optimized design of a plunger pump used in a hydrazine propellant supply system for satellite attitude control, testing under vacuum and thermal cycling conditions is required to confirm its leakage and performance stability. For miniature plunger pumps used in medical devices, the focus of validation may be on their noise levels and flow stability during long-term operation.

[0101] This step completes the closed loop from theoretical optimization to practical verification, strongly demonstrating the effectiveness and versatility of the method of this invention in solving the design challenges of low leakage and high efficiency in high-pressure friction pairs.

[0102] Example 3

[0103] This invention provides a method for designing the hose profile of a peristaltic pump for low shear damage to non-Newtonian fluids based on flow entropy production optimization. This extends the method to the specific field of handling non-Newtonian fluids, focusing on peristaltic pumps used in the biopharmaceutical industry for delivering cell culture media and protein solutions. The design goal is to minimize the shear and tensile stresses experienced by the fluid during delivery (both contributing to entropy production) by optimizing the roller profile and pressure block profile of the pump head, thereby protecting shear-sensitive active ingredients.

[0104] S1. Analyze the rheological properties of non-Newtonian fluids and the requirements for protection against biological activity, and construct a structured design task with the core objective of minimizing fluid shear damage.

[0105] Specifically, in some implementations, the construction of task models for biopharmaceutical fluids requires a deep integration of rheology and bioprocessing knowledge. The core is to transform the biological goal of "reducing cell damage or protein denaturation" into the fluid dynamics metric of "minimizing entropy production".

[0106] The system receives inputs including: a detailed rheological model of the fluid (such as power-law model or Carson model parameters), the known shear sensitivity thresholds of active components (such as cells or enzymes), the required flow range, and the pump rotation speed. The system analyzes this information to construct a structured task model. Design variables are defined as the coordinates of the control points of the drive roller's profile curve and the profile of the pump trough (pressure block). The optimization objective set includes: the maximum shear rate experienced by the fluid during the pump tube's pressure-release cycle, the total entropy yield (combining the irreversible processes of shearing and stretching), and the smoothness (pulsation) of flow delivery. Constraints include the hose's fatigue life (related to compression) and the minimum compression requirement to ensure complete closure and generate flow.

[0107] This step requires adapting to complex fluid behavior. For example, when optimizing a peristaltic pump for delivering cell suspensions containing fragile microcarriers, the model needs to focus on suppressing transient high shear forces. When optimizing a pump for delivering high-viscosity, shear-thinning gels, the goal is to reduce the average shear stress level and avoid localized stagnation.

[0108] This step materializes the abstract requirement of "biocompatibility" into quantifiable and optimizable fluid dynamics design goals centered on entropy production and shear history, providing an innovative design approach for protecting high-value-added bioproducts.

[0109] S2. Establish a transient simulation process that combines fluid-structure interaction and fluid volume method to simulate large deformation of hoses and non-Newtonian flow, and automatically extract shear field and entropy production data.

[0110] Specifically, in some implementations, accurately simulating the large deformation, rolling contact, and complex flow of non-Newtonian fluids in the hoses inside a peristaltic pump requires the establishment of a highly integrated multiphysics automated simulation process.

[0111] The system automatically builds an initial model including a flexible hose and a fluid domain based on parametric roller and pump trough geometry. It employs the Arbitrary Lagrange-Euler (ALE) method or a method combined with Smoothed Particle Hydrodynamics (SPH) to handle large deformation motions of the hose wall. The mesh module can automatically re-divide or update the fluid domain mesh according to the preset roller rotation angle. The CFD solver is set to transient calculation, and the fluid properties are specified as the input non-Newtonian model. The simulation fully records the historical data of shear rate, pressure, and velocity at every point in the flow field over several cycles.

[0112] After the simulation is completed, the post-processing script not only calculates the total entropy production rate for the entire cycle, but also specifically analyzes and outputs key indicators such as the history of the maximum shear rate experienced by fluid particles and the spatiotemporal distribution of high-shear regions (where the shear rate exceeds a set threshold). These data are directly related to the degree of damage that the fluid's active components may suffer.

[0113] This simulation method can be extended to other scenarios involving flexible bodies and complex fluids. For example, when optimizing the impeller or flow channel of a cardiac assist pump (blood pump), it is necessary to simulate the flow of blood (a non-Newtonian fluid) to reduce hemolysis (blood cell damage), and the core evaluation indicators are also the shear stress level and entropy production in the flow field.

[0114] This step enables precise quantification of the complex transient flow field within the peristaltic pump, providing unprecedented detailed data insights for optimization aimed at protecting the active components of the fluid.

[0115] S3. A multi-objective genetic algorithm is adopted, with entropy production and maximum shear rate as the core objectives, to jointly optimize the roller-pump trough profile.

[0116] Specifically, in some implementations, in order to simultaneously minimize fluid damage (entropy production, maximum shear rate) and control flow pulsation, a powerful multi-objective optimization algorithm is required to jointly optimize the roller and pump trough profiles.

[0117] The system's comprehensive objective function F comprises three normalized sub-objectives: total entropy yield (weight 0.5), maximum shear rate experienced by the fluid (weight 0.3), and outlet flow fluctuation coefficient (weight 0.2). Since the design variables involve two interacting parts, resulting in a high dimensionality variable space, this embodiment employs a multi-objective genetic algorithm (such as NSGA-II). The population size is set to 60, with 80 iterations. The optimization algorithm drives the synchronous evolution of the roller and pump trough profiles, automatically evaluating the fluid damage potential and flow performance of each design scheme.

[0118] This optimization strategy is applicable to all fluid transport scenarios requiring delicate handling. For example, when optimizing a rotary pump used in the food industry to transport yogurt containing fruit pulp, the objective function needs to balance reducing pulp breakage (related to shear force) with maintaining transport efficiency. When optimizing a cosmetic cream filling pump, the focus is on avoiding excessive shear damage to the cream structure (affecting thixotropy), and entropy production is a sensitive indicator for measuring this process.

[0119] Through this step, the system automatically explores profile combinations that traditional empirical design cannot reach, and finds the optimal geometric configuration that can significantly reduce fluid shear damage.

[0120] S4. Verify the protective effect of the optimized pump head through biological process experiments, and compare the retention rate of cell activity or protein activity.

[0121] Specifically, in some implementations, the final validation must return to the performance of the bioprocess itself. Comparative experiments are used to quantify the protective effect of the novel pump head on the active ingredients.

[0122] An optimized pump head prototype was manufactured and compared with a commercial standard pump head in parallel tests. The same sensitive fluid (e.g., CHO cell culture medium) was used for circulation delivery at the same flow rate and runtime. After the test, key biometrics were measured: for cell culture medium, cell viability and cell doubling time were measured; for protein solutions, the concentration or purity of active proteins was measured. The test results confirmed that the pump head, optimized based on entropy production and shear rate, improved cell viability by approximately 15% or protein activity recovery by approximately 8%, while meeting process requirements.

[0123] This direct biological validation is the ultimate and most convincing evidence. For example, when validating a pump used to deliver a live vaccine, animal experiments are needed to assess the retention of vaccine potency. For systems delivering stem cell suspensions, it is necessary to verify whether the multi-lineage differentiation potential of stem cells is impaired during the delivery process.

[0124] This step successfully links fluid dynamics optimization with the final bioprocess output, fully demonstrating the powerful application potential and value of the method of this invention in the high-value-added and high-tech field of biopharmaceuticals.

[0125] In summary, this invention, through the three embodiments described above (low-noise screw pump, high-pressure low-leakage plunger pump, and low-shear damage peristaltic pump), fully demonstrates the universality and powerful effectiveness of the design method based on minimizing flow entropy production across different types of fluid machinery, different media properties (Newtonian / non-Newtonian fluids), and different core engineering objectives (noise reduction, efficiency improvement, and media protection). Starting from the fundamental principle of flow irreversibility, this method provides a unified and effective innovative approach and tool for the refined and high-performance design of fluid machinery.

[0126] This invention relates to a low-noise screw profile design method and device for twin-screw pumps based on minimizing flow entropy production. It can achieve multi-objective collaborative optimization with minimizing entropy production as the core, suppress flow noise at the source, significantly improve the operational stability and reliability of twin-screw pumps, and reduce operation and maintenance costs and noise pollution.

[0127] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for designing low-noise screw profiles for twin-screw pumps based on minimizing flow entropy production, characterized in that, Includes the following steps: S1. Establish a parameterized screw profile model. Based on the cycloidal-circular arc combined tooth profile theory, define key geometric parameters and construct profile generation equations that support dynamic adjustment of key geometric parameters. S2. Construct a CFD simulation model with flow entropy production rate as the core calculation index, and numerically simulate the three-dimensional transient flow field in the working chamber of the twin-screw pump to capture the flow state of the fluid in the entire process of meshing, compression and discharge. S3. Establish a multi-objective optimization function with minimizing entropy production as the primary objective and taking into account efficiency and pressure fluctuations. Normalize and weight the average total entropy production rate, volumetric efficiency and the amplitude of outlet pressure fluctuations to form a comprehensive optimization objective. S4. Utilize intelligent optimization algorithms for automatic iterative optimization. By iteratively updating key geometric parameters and driving the CFD simulation model to iteratively calculate, the screw profile parameters that optimize the multi-objective optimization function value are obtained.

2. The low-noise screw profile design method for a twin-screw pump according to claim 1, characterized in that, In step S1, key geometric parameters are selected for parameterization, and a controllable relationship is established between the parameterized model and the 3D solid model. The key geometric parameters include, but are not limited to: addendum circle radius, dedendum circle radius, meshing angle, coordinates of control points on the tooth profile curve, and helix angle.

3. The low-noise screw profile design method for a twin-screw pump according to claim 1, characterized in that, In step S2, the CFD simulation model includes: S2.1 Parameter Modification: The parametric geometric modeling kernel is called to discretize the input initial profile, and the coordinates of several key control points are selected as design variables to control the tooth tip radius, tooth root curve shape and meshing area profile. When the design variables change, the parametric engine can automatically generate new geometrically perfectly matched male and female rotor 3D solid models and automatically assemble and generate a fluid domain model containing the working cavity. S2.2 Automatic Modeling: Establish a CFD model of the working chamber of the twin-screw pump, transfer the fluid domain model to the mesh generation module, and generate a fine prism layer mesh near the rotor wall based on the medium viscosity and rotor speed, using a boundary layer mesh strategy to capture high shear flow. The number of meshes is controlled by a polyhedral mesh inside the chamber. The mesh quality standard is preset in the mesh generation module. Unqualified meshes trigger automatic re-rendering. S2.3 Automatic Simulation: Call the CFD solver to perform unsteady flow simulation. The settings include: selecting the Realizable k-ε turbulence model suitable for viscous flow, setting the medium properties to the lubricating oil parameters defined in the structured task model, and setting the boundary conditions according to the design conditions; the solver automatically calculates the transient flow field within the complete meshing cycle. S2.4 Automatic extraction of target values: The flow entropy productivity calculation module is embedded in the CFD simulation model. Based on the entropy productivity theory calculation formula, the flow entropy productivity is extracted from the flow field data and integrated to obtain the time-averaged total entropy productivity, thus quantifying the irreversible losses in the flow process.

4. The low-noise screw profile design method for a twin-screw pump according to claim 3, characterized in that, The flow entropy production rate is dominated by the entropy production rate caused by fluid viscosity dissipation. The time-averaged total entropy production rate is obtained by volume integral of the pump working chamber. It also includes additional entropy production components caused by flow shear and turbulent pulsation, which are used to fully reflect the energy loss characteristics of the flow field.

5. The low-noise screw profile design method for a twin-screw pump according to claim 4, characterized in that, The average total entropy productivity The expression is: ； in, For fluid dynamic viscosity, For fluid temperature, For velocity components, The coordinates are the rectangular coordinates of spatial points within the computational domain of the flow field in the working chamber of the twin-screw pump.

6. The low-noise screw profile design method for a twin-screw pump according to claim 5, characterized in that, The multi-objective optimization function is defined as a normalized weighted sum of the average total entropy yield, volumetric efficiency, and the amplitude of the outlet pressure fluctuation, defined as follows: ； in, The objective function value, Let the initial entropy production rate be , For volumetric efficiency, This represents the amplitude of the outlet pressure pulsation; the subscript 0 represents the corresponding value of the initial profile. Let be the weighting coefficient, satisfying ,and The largest proportion is used to ensure that minimizing entropy production becomes the primary goal.

7. A low-noise twin-screw pump device, designed according to the method described in any one of claims 1 to 6, characterized in that, include: Profile parameterization module: It is used to receive the geometric parameter input of the initial profile, parse and extract key geometric parameters, and generate a structured input model containing parameters such as tooth tip circle radius, meshing angle, and control point coordinates. It supports rapid iteration of multiple sets of key geometric parameters. Entropy production calculation and optimization module: Based on structured parameters and CFD simulation models, it calls preset entropy production calculation templates and intelligent optimization algorithms to dynamically generate an optimization process chain that includes profile parameter updates, flow field simulation, and objective function evaluation; Convergence and Verification Module: The rule engine is used to perform multi-dimensional verification of the optimization results, including entropy yield convergence judgment, volumetric efficiency threshold verification and pressure pulsation amplitude evaluation, and the final optimal profile scheme is formed after the simulation verification is passed. Prototype verification module: The screw rotor is manufactured according to the optimal profile scheme, and the noise, efficiency and pressure pulsation data of the prototype are collected in real time. The data are compared with the initial profile, and automatic feedback is provided to trigger secondary optimization iteration, thus completing the closed-loop verification of the entire process.

8. The low-noise twin-screw pump device according to claim 7, characterized in that, The end face tooth profile of the screw rotor is obtained through the flow field entropy production calculation and optimization module and the optimization process is applied directly to the processing and manufacturing of screw rotors with equal pitch or variable pitch.

9. The low-noise twin-screw pump device according to claim 7, characterized in that, The tooth profile of the screw rotor has a smooth and continuous curvature change to reduce flow separation; the tooth profile shape in its meshing area is modified to smooth the fluid transport process; and its helix angle is finely adjusted along the axial direction to balance the pressure gradient, thereby effectively reducing the flow noise and vibration level in the pump.

Images