Method and device for manufacturing gear pump shell

Abstract

The embodiment of the invention discloses a gear pump shell manufacturing method and device. The method comprises the steps that a finite element model containing design space, boundary conditions and material parameters is constructed according to the structure optimization requirement of the gear pump shell; generating a data set comprising an initial density field, an intermediate iteration density field and a corresponding final optimization result through topological optimization simulation of a plurality of groups of punishment solid isotropic materials with different density configurations by an SIMP method; reconstructing a three-dimensional model of the gear pump shell by adopting a deep learning regression model and taking an initial density field and a middle iteration density field as optimization parameters and taking a final optimization result as an optimization target; the density field is a density field after current iteration updating; the density field is a density field iterated by the SIMP method; iterating the deep learning regression model to obtain a density field; and when the optimization target is achieved, the gear pump shell is manufactured according to the reconstructed three-dimensional model.

Description

Technical Field

[0001] This invention relates to the field of mechanical technology, and in particular to a method and apparatus for topology optimization of gear pump housing based on the penalized solid isotropic material with penalization (SIMP) variable density method. Background Technology

[0002] The gear pump housing contains a pair of meshing gears, which are covered by end caps at the front and rear, forming numerous sealed working chambers with the pump housing. The pump gears, as the core component of the gear pump, can be used in industrial manufacturing or transportation. For example, when the driving gear inside the gear pump housing drives the driven gear to rotate, the gear teeth in the suction chamber gradually disengage, the sealed chamber volume increases, the pressure decreases, and oil is drawn in from the oil tank. In the pressure chamber, the gear teeth gradually engage, the sealed chamber volume decreases, the oil is squeezed out, and the pressure increases, thus realizing the oil suction and pressure processes. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method and apparatus for manufacturing a gear pump housing. The technical solution of the present invention is implemented as follows: A first aspect provides a method for manufacturing a gear pump housing, the method comprising: To address the structural optimization requirements of the gear pump housing, a finite element model was constructed, including design space, boundary conditions, and material parameters. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. Through multiple sets of penalized SIMP topology optimization simulations of isotropic solid materials with different density configurations, a dataset containing an initial density field, intermediate iterative density fields, and corresponding final optimization results is generated. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP topology optimization simulation employs a density interpolation model: where... ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; Using deep learning regression models and based on The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. When the optimization goal is achieved, the gear pump housing is manufactured based on the reconstructed 3D model.

[0004] The second aspect provides a gear pump housing manufacturing apparatus, comprising: The construction module is used to build a finite element model containing design space, boundary conditions and material parameters for the structural optimization requirements of the gear pump housing. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. The optimization module is used to generate a dataset containing an initial density field, intermediate iteration density fields, and corresponding final optimization results through multiple sets of penalized SIMP method topology optimization simulations of penalized solid isotropic materials with different density configurations. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP method topology optimization simulation uses a density interpolation model: where... ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; The reconstruction module is used to employ deep learning regression models and based on... The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. A printing module is used to manufacture the gear pump housing based on the reconstructed 3D model when the optimization target is achieved.

[0005] A third aspect provides a computer-readable storage medium storing computer-executable code that can be read and executed by a processor; the executable code, after being executed by the processor, can implement the method provided in the first aspect.

[0006] The technical solution provided by the embodiments of the present invention has at least the following beneficial effects: through finite element modeling, SIMP simulation, and deep learning reconstruction, the stress concentration factor of the gear pump housing can be reduced, fatigue life can be improved, flow channel pressure loss can be reduced, fluid efficiency can be improved, and weight can be reduced. This method takes into account strength, efficiency, and lightweight, and through multiple sets of simulations and data-driven optimization, it shortens the design cycle, improves the accuracy of parameter iteration, and is suitable for the manufacture of high-performance gear pumps under complex load conditions. Attached Figure Description

[0007] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of a method for manufacturing a gear pump housing according to an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a gear pump provided in an embodiment of the present invention; Figure 3 A schematic diagram illustrating the topology change from the design space to the optimized topology, provided as an embodiment of the present invention; Figure 4 This is a schematic diagram of a process for optimizing topology using a deep learning regression model, provided as an embodiment of the present invention. Figure 5 This is a schematic diagram of a method for manufacturing a gear pump housing according to an embodiment of the present invention; Figure 6 This is a schematic diagram of a gear pump housing manufacturing device provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0008] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0009] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0010] For example, gear pumps (and their housings) are primarily used for high-precision ink delivery and pressure control, and their core functions are closely related to the housing structure design. The following explanation focuses on application scenarios, housing functions, and structural adaptability: Ink metering: During printing (such as offset, flexographic, and gravure printing), high-viscosity ink (typically 100-10000 cP) needs to be uniformly and stably delivered to the printing rollers or printhead. Gear pumps achieve metered ink delivery through volume changes caused by gear meshing, with flow control accuracy up to ±1%. Housing function: The pump chamber provides a sealed space for gear rotation. By strictly controlling the gap between the gears and the inner wall of the pump chamber (typically 0.03-0.08 mm), ink leakage is reduced, ensuring stable delivery. Ink requires a certain pressure (e.g., approximately 0.1-0.5 MPa in offset printing) to overcome pipeline resistance and adhere evenly to the substrate during transmission between printing rollers. Gear pumps establish stable pressure at the outlet by meshing and squeezing the ink, preventing color differences or uneven ink layers due to pressure fluctuations. Housing function: Reinforcing ribs and bolted connections enhance housing rigidity and prevent deformation under high pressure; sealing grooves with sealing rings (such as fluororubber) prevent ink leakage and maintain stable system pressure. Printing inks often contain pigments, resins, and other components, some of which are corrosive or prone to curing. The housing material is often made of stainless steel (such as 316L) or surface-treated aluminum alloy to prevent ink corrosion; the inner wall of the flow channel is polished (roughness Ra≤0.8μm) to reduce ink residue and cured deposits. In summary, the gear pump housing is a crucial component of printing equipment, and its structure directly affects the lifespan of the printing equipment and the print quality.

[0011] In summary, the gear pump housing is a crucial component of printing equipment. The various structures within the gear pump housing exhibit close spatial fit and functional coordination. The gear pump housing includes the pump chamber, partitions, reinforcing ribs, bolts, bearing housings, and sealing grooves, among other things. The following details their positional relationships from a three-dimensional layout and operational logic perspective: The pump chamber is located in the central area of ​​the gear pump housing, and is cylindrical or nearly cylindrical, providing space for gear rotation and fluid transport. The inner wall of the pump chamber maintains a small gap (usually 0.05-0.1mm) with the outer contour of the gear to reduce friction and ensure sealing. Both ends of the pump chamber are connected to bearing seats to support the gear shaft. The inlet and outlet flow channels are directly connected to the pump chamber, with the inlet flow channel located on the low-pressure side of the pump chamber (where the gear disengages) and the outlet flow channel located on the high-pressure side (where the gear engages).

[0012] A baffle plate is installed inside the pump chamber, typically perpendicular to the fluid flow direction, dividing the pump chamber into multiple functional areas (such as separating chambers of different pressure levels in a multi-stage gear pump). The edges of the baffle plate fit tightly against the inner wall of the pump chamber, achieving a seal through sealing grooves and seals; the gradually varying porosity lattice structure on the baffle plate works in conjunction with the inlet and outlet flow channels to regulate the fluid velocity and pressure distribution; simultaneously, the sides of the baffle plate may be connected to reinforcing ribs to enhance the overall structural stability.

[0013] Reinforcing ribs are distributed on the outer and inner walls of the shell (especially in areas of stress concentration), often in a radial, grid-like, or principal stress direction. Outer wall reinforcing ribs connect the pump cavity to the shell edge, enhancing the shell's resistance to deformation; inner wall reinforcing ribs may extend to the vicinity of partitions or bearing seats, forming a mechanical support network; around bolt holes, reinforcing ribs are distributed in a ring or radial pattern, dispersing bolt preload and working load.

[0014] Bolt holes are distributed along the edges of the mating surfaces of the upper and lower covers of the shell, typically in a circular or symmetrical pattern. These bolt holes penetrate the shell wall thickness, allowing the upper and lower covers to be securely connected via bolts. Bosses around the bolt holes are connected to reinforcing ribs, enhancing the strength of the connection. Sealing grooves are located inside the bolt holes, working in conjunction with the bolted connection structure to ensure the shell's airtightness.

[0015] The sealing groove is formed at the mating surfaces of the housing and end caps, the edges of the partitions, and other areas requiring sealing. It is either annular or a continuous strip. The sealing groove surrounds channels where fluid or lubricating oil may leak, such as the pump inlet and outlet, and bearing housing holes. When used with bolted connections, the sealing element is further compressed when the bolts are tightened, enhancing the sealing effect.

[0016] The bearing housings are located on both sides of the housing, are cylindrical, and axially aligned with the pump cavity. The inner bore of the bearing housing is used to install the rolling bearing and support the rotation of the gear shaft; the outer side of the bearing housing may be connected to reinforcing ribs to improve support rigidity; the lubricating oil grooves on the bore wall of the bearing housing are connected to the pump cavity or external lubrication system to provide lubrication for the bearing.

[0017] The gear pump casing also features a biomimetic pore structure and a variable cross-section flow channel. These structures are distributed in non-critical stress areas of the casing (such as the sidewalls far from the pump cavity and non-fluid passage areas of the baffles), achieving lightweight material distribution through topology optimization. The variable cross-section flow channel refers to the inlet and outlet channels extending from the pump cavity to the outside of the casing. The inlet section is flared on the outside of the casing, gradually narrowing to connect with the pump cavity inlet; the outlet section gradually expands outward from the high-pressure area of ​​the pump cavity, optimizing fluid flow characteristics.

[0018] like Figure 1 As shown, a method for manufacturing a gear pump housing includes: S1110: To address the structural optimization requirements of the gear pump housing, a finite element model is constructed that includes design space, boundary conditions, and material parameters. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. S1120: Through multiple sets of penalized SIMP topology optimization simulations of isotropic solid materials with different density configurations, a dataset is generated containing an initial density field, intermediate iteration density fields, and corresponding final optimization results. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP topology optimization simulation uses a density interpolation model: where, ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; S1130: Employing a deep learning regression model and based on The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. S1140: When the optimization objective is achieved, the gear pump housing is manufactured based on the reconstructed 3D model.

[0019] Figure 2The diagram shows a gear pump. The gear pump includes a gear pump housing and an internal gear pump. The internal gear pump achieves hydraulic energy conversion through the continuous action of the suction chamber and the discharge chamber to complete the work. When the gear shaft rotates counterclockwise, the right gear teeth gradually disengage, the cavity volume increases, creating a vacuum, thus drawing in oil; simultaneously, the left gear teeth engage, squeezing out the oil and generating pressure. The gear pump includes a suction chamber 1, a discharge chamber 2, a crescent plate 3, a driven internal gear 4, and a driving pinion 5. The gear pump housing not only bears external loads but also the impact caused by uneven internal flow pulsation, resulting in uneven pressure changes and impact loads. Therefore, simulation analyses were performed for internal pressures of 2MPa, 3MPa, 4MPa, and 5MPa. The high-pressure outlet is the main stress concentration point; therefore, in the simulation, an external load of 25MPa is applied to the housing to replace the impact load on the inlet and outlet.

[0020] In S1110, for a gear pump in a hydraulic system, engineers can use ANSYS software to define the design space such as the pump chamber and inlet / outlet flow channels; apply fluid pressure load (working pressure 15MPa) and dynamic pressure fluctuation load (frequency 50Hz, amplitude ±2MPa) as boundary conditions; and set the material parameters to be a mixture of aluminum alloy (shell) and steel (critical pressure-bearing parts).

[0021] In some embodiments, a finite element model is a digital model that breaks down a solid structure into tiny units for mechanical analysis, providing a data basis for optimization by simulating stress and deformation under load.

[0022] In S1120, SIMP method topology optimization with different initial densities (0.2-0.8) can be set with 10, 15, etc. Through iterative calculation using a penalty function (such as penalty factor p=3), the initial density field (preliminary material distribution), intermediate iterative density field (optimization process data) and final optimization results are generated, including structural compliance (deformation), stress distribution (maximum stress point) and fluid pressure distribution (flow channel pressure loss).

[0023] In some embodiments, the SIMP method simulates material removal or retention by adjusting the element density (between 0 and 1) to achieve structural topology optimization. For example, the SIMP method is used to adjust the density of each element within the structural space using the following formula. The design space can be divided into multiple spaces of the same or different sizes using a grid. Each space corresponds to one element. Solid elements, also called solid units, are elements that are not hollowed out. These solid elements are filled with fabrication material. Void elements can be at least partially hollowed out.

[0024] In S1130, a trained deep learning regression model (such as a neural network) is used, with the density field as the input parameter, to reconstruct the three-dimensional model of the gear pump housing, aiming to reduce stress concentration and improve fluid efficiency, thus optimizing structural details such as the pump chamber shape and baffle layout. This formula is adopted. Optimizing or reconstructing density fields can combine the advantages of both deep learning regression models and SIMP-based density field prediction or simulation, thereby achieving rapid convergence of the optimization objective.

[0025] In other embodiments, the calculation methods for the element stiffness matrix and total compliance in the SIMP method topology optimization simulation are as follows: as well as ;in, For the first Stiffness matrix of each element; Let be the stiffness matrix of the solid element; For the total compliance, For the first The compliance of each unit. In this embodiment of the disclosure, compliance is also introduced to reconstruct or optimize the parameters of the gear pump housing, thereby ensuring the deformability of the housing.

[0026] Figure 3 The diagram shows the optimized topology generated after passing through S1110 to S1130 from the design space.

[0027] In some embodiments, deep learning regression models predict optimal structural parameters by learning patterns from large amounts of data, replacing traditional empirical design. Compared to traditional designs, the optimized gear pump housing is lighter, the maximum stress is reduced, and the flow channel pressure loss is decreased, significantly improving structural strength, fluid efficiency, and weight reduction.

[0028] In some embodiments, S1110 may include: design space with the pump chamber inner wall thickness distribution, baffle layout, and inlet / outlet flow path of the gear pump housing as core parameters; constructing the boundary conditions; and setting the anisotropic parameters of the aluminum alloy and steel hybrid material.

[0029] In some embodiments, the boundary conditions may further include: bolt preload boundary conditions. Exemplarily, the following details the design space construction, boundary condition setting, and material parameter determination, and elaborates on the technical effects in conjunction with optimization logic: Taking a certain type of gear pump as an example, the thickness distribution of the pump cavity inner wall is used as a variable, with the initial thickness range set to a specified range, for example, 3-8mm. The baffle layout is set as a movable parallel plate structure, and its position is adjustable within the range of [0.2L, 0.8L] (L is the total length of the pump cavity) along the length of the pump cavity. The direction of the inlet and outlet flow channels is used as a variable, with the angle θ with the pump cavity as a variable, and the value of θ is [30°, 60°]. Using CAD software, the outer contour of the gear pump housing is extended outward by 20mm to form the design space boundary, ensuring sufficient material optimization range. The thickness of the pump cavity inner wall and the position of the baffle affect the fluid pressure distribution, and the direction of the inlet and outlet flow channels determines the fluid inlet and outlet efficiency. The three are interrelated and together constitute the core parameters of the design space.

[0030] Based on the gear pump's operating pressure of 16 MPa, a uniformly distributed pressure load is applied to the inner wall of the pump chamber. In the gear meshing region, due to pressure concentration, the local pressure is applied at 1.2 times the uniformly distributed pressure (i.e., 19.2 MPa). Considering fluid pulsation characteristics, a time-varying pressure function is set in the inlet and outlet flow channels. ,in The rated pressure is 16MPa. Dynamic pressure fluctuation load: Analyze the vibration frequency of the gear pump during operation to determine the main vibration frequency. For example, if the main vibration frequency is 50Hz, apply a dynamic pressure fluctuation load of 2MPa and 50Hz to the outer wall of the pump chamber to simulate pressure fluctuations under actual working conditions. Bolt preload: The gear pump housing is fixed with 4, 6, or 8 bolts of a specified diameter (e.g., M10). A specified preload (e.g., 1000N) is applied to each bolt. Preload boundary conditions are set around the bolt holes to simulate the effect of bolt tightening on the housing. Aluminum alloy and steel hybrid material: The critical pressure-bearing area of ​​the pump chamber inner wall uses alloy steel with an elastic modulus E1 = 210GPa and Poisson's ratio ν1 = 0.3; the non-critical parts of the housing use aluminum alloy with an elastic modulus E2 = 70GPa and Poisson's ratio ν2 = 0.33. Assuming a material transition region length of 10mm, a gradient function is used to achieve a smooth transition of properties from aluminum alloy to steel, avoiding stress concentration caused by abrupt material changes. Anisotropy parameters: For the die-cast aluminum alloy portion, considering the orientation of the material during the forming process, the anisotropy parameters are set as follows: elastic modulus along the die-casting direction Ex = 75 GPa, elastic modulus perpendicular to the die-casting direction Ey = Ez = 65 GPa. Here, Ex, Ey, and Ez are elastic moduli in different directions, and x, y, and z are the three axes of a three-dimensional coordinate system, which are perpendicular to each other. Ex = 75 GPa and Ey = Ez = 65 GPa are just examples; in reality, Ex is greater than Ey and / or Ez.

[0031] x can represent the axial direction (axis direction) of the gear pump, that is, the direction of extension of the gear rotation shaft. For example, if the inlet and outlet channels of the casing are arranged axially, then the x direction is the main direction of the channel; if the direction of bolt preload is parallel to the axis, it also corresponds to the x direction. The y direction is horizontal and perpendicular to the x-axis, pointing radially outward from the casing (such as the left-right direction of the gear meshing area). The z direction is perpendicular to the xy plane, pointing to the top or bottom of the casing (such as the bolt connection direction between the pump cover and the pump body).

[0032] In this embodiment, fluid pressure, dynamic fluctuations, and bolt preload are comprehensively considered, making the finite element model closer to the actual working condition and reducing simulation error. By combining aluminum alloy and steel hybrid materials and anisotropic parameters, materials are allocated as needed. Compared to single-material designs, the shell weight is reduced while meeting strength requirements. Structural performance is improved: The pump chamber, baffles, and flow channel parameters are rationally designed. Optimized flow channel pressure loss is reduced, gear pump volumetric efficiency is improved, and vibration amplitude is lowered, effectively enhancing overall performance and reliability.

[0033] In some embodiments, the final optimization result includes at least one of the following: a flow channel pressure loss coefficient based on fluid dynamics simulation; a stress concentration factor based on fatigue life prediction; and a structural vibration frequency response based on modal analysis. like Figure 4 As shown, S1130 includes: S1131: Using the deep learning regression model, the initial density field, intermediate iteration density field and stress concentration coefficient in the final optimization result are used as inputs to output the pump cavity inner wall thickness distribution parameters, wherein the pump cavity inner wall thickness distribution parameters make the pump cavity inner wall adopt a gradient wall thickness structure. S1132: Based on the density change rate distribution of the intermediate iteration density field, combined with the flow channel pressure loss coefficient of the final optimization result, the porosity and distribution direction of the baffle are optimized by a genetic algorithm, and the porosity and distribution direction of the baffle make the baffle have a gradient pore lattice structure. S1133: Using convolutional neural networks to analyze the boundary features of the initial density field, the local optimization results of the intermediate iteration density field, and the vibration frequency response of the final optimization result, to generate the stiffener layout and variable cross-section parameters; S1134: Based on the thickness distribution parameters of the pump cavity inner wall, the porosity and distribution direction of the baffle, and the layout and variable cross-section parameters of the reinforcing ribs, reconstruct the three-dimensional model of the gear pump housing.

[0034] In this embodiment of the disclosure, the final optimization result is a multi-dimensional index, specifically the final optimization result may include at least one of the following multi-dimensional performance indicators: Flow channel pressure loss coefficient: obtained through computational fluid dynamics simulation. This coefficient reflects the degree of energy loss when fluid flows in the inlet and outlet channels of the gear pump and is a key parameter for measuring fluid delivery efficiency. Stress concentration factor: Calculated based on fatigue life prediction model, it is used to evaluate the stress concentration in key parts of the shell under alternating load (such as pump chamber corners and near bolt holes). The smaller the value, the stronger the fatigue resistance of the structure. Structural vibration frequency response: obtained through modal analysis, it presents the vibration characteristics of the shell under different frequency excitations, which can be used to avoid resonance risks and ensure the stability of equipment operation. In some embodiments, the specific application of a deep learning regression model may include at least one of the following: Pump cavity inner wall gradient thickness optimization: Using the initial density field, intermediate iteration density field, and stress concentration coefficient as inputs, a deep learning regression model learns the correlation between density distribution and stress concentration in historical optimization data, and outputs the thickness distribution parameters of each region of the pump cavity inner wall. For example, the wall thickness is increased in the gear meshing region with high stress concentration coefficient; and appropriately thinned in low stress regions, forming a gradient wall thickness structure along the circumferential or axial direction. In some embodiments, the gradual pore lattice optimization of the baffle involves: based on the density change rate distribution of the intermediate iterative density field (reflecting the material removal and retention trend), combined with the flow channel pressure loss coefficient, using a genetic algorithm to iteratively optimize the porosity and distribution direction of the baffle. For example, near the flow channel inlet, the porosity is increased to reduce fluid resistance; in regions with drastic pressure changes, the pore direction is adjusted to align with the fluid flow direction, forming a gradual pore lattice structure. In some embodiments, the stiffener layout and cross-section optimization utilize the powerful image feature extraction capabilities of convolutional neural networks to analyze the boundary features of the initial density field (such as shell contours and mounting boundaries), the local optimization results of intermediate iterative density fields (such as material concentration areas), and vibration frequency response data to generate stiffener layout schemes (such as radial or mesh-like patterns) and cross-sectional parameters (such as thickness and width gradients). For example, in the high-amplitude regions displayed by the vibration frequency response, denser stiffeners with larger cross-sections are arranged.

[0035] In some embodiments, during the reconstruction of the three-dimensional model, the optimized pump cavity inner wall thickness distribution parameters, baffle porosity and distribution direction, stiffener layout and variable cross-section parameters are integrated and imported into the three-dimensional modeling software to accurately reconstruct the gear pump housing, forming a three-dimensional solid model that meets multiple performance requirements.

[0036] The above methods significantly improve fluid efficiency: by optimizing the gradient pore lattice structure of the baffle and the flow channel design, the flow channel pressure loss coefficient can be reduced, fluid transport efficiency can be improved, and energy consumption can be reduced. Structural reliability is enhanced: by optimizing the gradient wall thickness of the pump cavity inner wall based on the stress concentration factor, the stress concentration in key parts of the shell can be reduced. Combined with optimized stiffener layout, fatigue life can be improved, effectively reducing equipment failures caused by fatigue fracture. Vibration characteristics are improved: by optimizing the stiffener layout according to the vibration frequency response, the natural frequency of the shell avoids the operating vibration frequency, the structural vibration amplitude can be reduced, and noise and component wear caused by vibration can be decreased. By employing intelligent optimization techniques such as deep learning and genetic algorithms to replace traditional experience-based design, the optimal solution for balancing multiple objectives can be found quickly and accurately, shortening the design cycle and reducing R&D costs. Through the generation of innovative structures such as gradient wall thickness, gradually varying pore lattices, and variable cross-section reinforcing ribs, the design limitations of traditional gear pump housings are overcome, providing a new approach to manufacturing high-performance, lightweight gear pumps and enhancing the product's market competitiveness.

[0037] In some embodiments, the deep learning regression model, using the initial density field, intermediate iteration density fields, and stress concentration coefficients from the final optimization result as inputs, outputs pump cavity inner wall thickness distribution parameters, including: Using the aforementioned deep learning regression model, with the initial density field, intermediate iteration density fields, and stress concentration coefficients from the final optimization result as inputs, the output is the pump cavity inner wall thickness distribution parameters, forming a gradient wall thickness structure that satisfies the following conditions: Using the aforementioned deep learning regression model, with the initial density field, intermediate iteration density fields, and stress concentration coefficients from the final optimization result as inputs, the output is the pump cavity inner wall thickness distribution parameters, forming a gradient wall thickness structure that satisfies the following conditions: Wall thickness along the direction of fluid flow Satisfies the exponential decay function: ,in This represents the initial wall thickness at the entrance. For attenuation, Coordinates of the fluid flow path; Localized thickening zones form in areas of concentrated pressure, with the amount of thickening... , It is a nonlinear function that satisfies ∈ First predetermined range; This represents the maximum stress in the pressure concentration area. Wall thickness variation rate Ensure a smooth stress transition within the transition region, where the range is less than or equal to the second predetermined range. Circumferential wall thickness distribution and radial stress components Relationship, satisfy ,in The average wall thickness For amplitude, The harmonic order is... This is the phase angle.

[0038] In some embodiments, The initial wall thickness at the entrance is 6-8 mm. The value range can be 0.05-0.15 / mm. These are the coordinates of the fluid flow path. In some embodiments, The value of is a positive integer between 2 and 4.

[0039] In some embodiments, a localized thickening zone is formed in the pressure concentration area (such as directly below the gear meshing point), the thickening amount being... And satisfy ∈[1.5mm, 3mm].

[0040] In some embodiments, the need for local thickening of the pressure concentration region is addressed by a nonlinear function. Thickness increase must be met With maximum stress Positive correlation, and limited to a first predetermined range (assuming it is { , The following are examples of nonlinear functions that meet the requirements: Here are some examples of nonlinear functions that meet the requirements: Quadratic function type: ;in, Stress threshold (when ≤ hour, =0, To adjust the coefficients, constraints are applied. ≤ ≤ The result is determined by reverse calculation. This function allows the thickness increase to increase quadratically with the degree to which the stress exceeds the threshold, enabling a rapid response in high-stress areas.

[0041] Piecewise power function type: if , =0; if Less than ,and Less than or equal to 2 ; It can be any positive integer greater than 1; if Less than , ;in, This is the initial stress value. This is the saturation stress value. This is the proportionality coefficient. To ensure nonlinearity, the function is set to a power order. This is achieved through piecewise control when the stress exceeds [a certain threshold]. It begins to thicken, reaching... When the maximum thickness is reached .

[0042] Hyperbolic tangent function type: ;in, The coefficient controlling the slope of the function. This represents the stress threshold. The characteristics of the hyperbolic tangent function cause the thickness increase to rise rapidly after the stress exceeds the threshold, and then gradually saturate. To avoid excessive thickening.

[0043] Exponential cutoff type: if Greater than ,but ,if Less than or equal to ,but =0. Where, and For the coefficient, it must satisfy the following condition: Less than or equal to This function utilizes exponential growth to rapidly respond to high stress, but a cutoff condition must be applied to ensure the thickness increase remains within a predetermined range. These nonlinear functions, through different mathematical properties, can achieve both a sensitive response of the thickness increase to high stress regions and meet the constraints of a predetermined range, and can be selected and used according to specific engineering requirements and stress distribution characteristics.

[0044] In some embodiments, wall thickness variation rate Ensure a smooth stress transition within the transition region (≤0.3mm / mm). The amplitude is 0.5-1.5mm.

[0045] In the data preprocessing stage, the intermediate iteration density field is numerically differentiated to calculate the density change rate of each region and extract the flow channel pressure loss coefficient from the final optimization result; pressure-sensitive regions are divided according to the pressure loss coefficient.

[0046] By accurately analyzing the initial density field, intermediate iterative density field, and stress concentration coefficient using a deep learning regression model, and combining it with designs such as exponential decay function, local thickening of pressure concentration areas, and circumferential wall thickness cosine distribution, stress distribution is effectively optimized.

[0047] In regions of concentrated pressure, based on nonlinear functions By locally thickening the casing, the stress concentration factor is reduced, thereby decreasing the risk of fatigue cracks caused by stress concentration, improving the structural reliability of key parts of the gear pump casing, and extending the service life of the equipment.

[0048] The exponentially decreasing wall thickness structure along the fluid flow direction conforms to hydrodynamic characteristics, reducing resistance caused by abrupt changes in wall shape during fluid flow and lowering energy loss. Compared to traditional uniform wall thickness designs, flow channel pressure loss is reduced, fluid transport efficiency is improved, making the gear pump more energy-efficient and effective during operation. Strict control of the wall thickness variation rate is achieved. Within the transition region, a range less than or equal to the second predetermined limit is maintained to avoid stress abrupt changes caused by sudden changes in wall thickness, ensuring a smooth stress transition and reducing the risk of structural failure due to stress abrupt changes. Simultaneously, the circumferential wall thickness distribution and radial stress components... By adjusting the wall thickness using a cosine function, the circumferential forces on the shell can be further balanced, enhancing structural stability and reducing vibration and noise.

[0049] By utilizing deep learning regression models to output thickness distribution parameters, automated and parametric design is achieved, shortening the design cycle compared to traditional manual design. Furthermore, through precise parameter control, the generated gradient wall thickness structure meets the requirements of modern additive manufacturing and CNC machining processes, facilitating manufacturing implementation, reducing manufacturing difficulty and costs, and simultaneously improving the accuracy and quality of product design and manufacturing.

[0050] In some embodiments, S1120 may include: The intermediate iteration density field is numerically differentiated to calculate the density change rate of each region. And extract the flow channel pressure loss coefficient from the final optimization result. ; Based on pressure loss coefficient Delineate pressure-sensitive areas: where, when When, it is defined as a high-voltage sensitive area; when At that time, it is defined as the low-pressure sensitive area; where This represents the maximum value of the pressure loss coefficient; Smaller than the ; porosity of the partition Pore ​​distribution direction Gene segments encoded as genetic algorithms; Iterative optimization is performed using the following objective function: in and As weight, For the first The rate of change of density of each cell (also called a grid), For the first Porosity of each unit; Using selection probability, crossover probability, and mutation probability as genetic factors, the genetic algorithm outputs the optimal porosity and distribution direction parameter set. A gradient porosity lattice structure is generated based on the optimal porosity and distribution direction parameter set, satisfying the following constraints: Porosity gradient: along the fluid flow direction, porosity Satisfies a piecewise function: ,in, ; ,in, ; ,in, in and All are set values. This refers to the length of the partition. Directional distribution: In high-pressure sensitive areas, the angle between the long axis of the pores and the fluid flow direction is... ∈[15°, 30°]; in the low-pressure sensitive area, ∈[60°,90°]; Pore ​​morphology: Hexagonal pore units are used, with unit side length... Dynamically adjusted based on porosity: ; Boundary constraints: Porosity ≤0.4 within 5-8mm from the edge of the partition to ensure structural connection strength. The value range is 1.5-3mm.

[0051] For example, when When ≥, it is defined as a high-voltage sensitive area; when When this occurs, it is defined as a low-pressure sensitive area.

[0052] In the genetic algorithm optimization stage, the porosity p and pore distribution direction θ of the partition are encoded as gene segments for the genetic algorithm, assuming the initial population size is set to 50-100 individuals; iterative optimization is performed using the following objective function: ;in or After 200-300 generations of genetic operations, the optimal porosity and distribution direction parameter set is output.

[0053] Boundary constraints: Within a first range from the edge of the partition, the porosity is ≤ a second range to ensure structural connection strength. For example, boundary constraints: Within a 5-8mm range from the edge of the partition, the porosity is ≤ 0.4 to ensure structural connection strength.

[0054] In the data preprocessing and region partitioning stage, the intermediate iteration density field is numerically differentiated, and the density change rate of each three-dimensional grid region is calculated using the central difference method. Quantify the changing trend of material distribution; extract the pressure loss coefficient of the flow channel. ,based on The distribution characteristics, in terms of the proportionality coefficient , (like =0.8, =0.2) Divide the high-pressure sensitive area and the low-pressure sensitive area to identify the key areas of fluid resistance. During the genetic algorithm optimization stage, the porosity of the partition is... (Value range 0.3-0.7), Pore distribution direction (15°–90°) is encoded into gene segments to construct an initial population of 20 individuals. Objective function In this process, weights are used to balance the rate of change in density and the pressure loss. With selection probability set to 0.8, crossover probability to 0.6, and mutation probability to 0.02, the optimal porosity and distribution direction parameter set are selected after 50 iterations. During the gradual pore lattice formation stage, porosity is gradually varied along the fluid flow direction (x-axis), using the baffle length L as a reference, through a piecewise function. For example, in the inlet section ( Porosity from minimum value =0.3 increases exponentially; in the middle section ( Maintain maximum value =0.7; decreases at the outlet section to reduce fluid disturbance.

[0055] Directional distribution: The long axis of the pores in the high-pressure sensitive zone forms an angle of 15°-30° with the fluid flow direction to reduce flow resistance; the low-pressure sensitive zone is set at 60°-90° to enhance structural support. For shape and boundary constraints: regular hexagonal porous elements are used, with a side length of... With porosity Dynamic adjustment (e.g.) To ensure uniform pore distribution, the porosity is limited to ≤0.4 within 5-8mm of the partition edge to avoid weakening the structural strength. By precisely dividing the pressure-sensitive region and optimizing the pore direction, the fluid resistance in the high-pressure sensitive area is reduced, and the pressure loss coefficient of the flow channel is decreased. Overall, the volumetric efficiency of the gear pump is improved, effectively reducing energy consumption and temperature rise. Based on density change rate and genetic algorithm to optimize porosity, the weight of the partition is reduced while meeting the structural load requirements. Edge porosity constraints and directional differentiation design reduce stress concentration in key parts of the partition, avoid strength attenuation caused by excessive porosity distribution, and balance lightweight and reliability. In some embodiments, the deep learning regression model includes a fully integrated neural network (CNN); the step of using a deep learning regression model, with the initial density field and intermediate iterative density fields as optimization parameters and the final optimization result as the optimization objective, to reconstruct the three-dimensional model of the gear pump housing includes: After the SIMP method has been executed up to the preset number of iterations T, the current density field is processed as follows: Input the trained CNN, output the predicted density distribution. ; Calculate the fluid pressure gradient of the current density field. The calculation formula is: ,in The fluid pressure distribution within the gear pump housing; Calculate the density field updated in this iteration ; The density field is updated using the following function to balance global prediction and local optimization while optimizing the structural performance of the gear pump housing under fluid pressure load and dynamic pressure fluctuation load. ;or, ; in, Predict weights for the CNN, with values ​​ranging from [0.6, 0.8]. The gradient correction weights have a value range of [0.1, 0.3]. This is the updated density field.

[0056] In some embodiments, α is the CNN prediction weight, with a value range of [0.6, 0.8]; β is the gradient correction weight, with a value range of [0.1, 0.3]. The density field is updated iteratively using the current SIMP method; the fluid pressure gradient Through formula The calculation shows that γ is the fluid-structure interaction coefficient, with a value range of [1.5, 2.0].

[0057] In some embodiments, the SIMP method, or solid isotropic material interpolation method, is a commonly used numerical method in topology optimization. It introduces a density variable (with values ​​of 0-1) to simulate the presence or absence of material, transforming the continuum topology optimization problem into a mathematical programming problem. In gear pump housing optimization, it is used to iteratively update the material distribution and find the optimal structural topology.

[0058] In some embodiments, the trained CNN is used to predict the density field after the SIMP method iteration, providing optimization trends at a global level and helping to improve optimization efficiency and quality. Figure 4 The diagram shows the process of optimizing the design of the gear pump housing using CNN.

[0059] In some embodiments, density field: a function describing the density distribution of materials in space; in topology optimization, each point in the density field corresponds to a density value ( ), indicating whether the location is filled with material ( =1 represents an entity. =0 indicates an empty space. By optimizing the density field, the material layout of the gear pump housing structure can be optimized.

[0060] In some embodiments, fluid pressure gradient : Represents the rate of change of fluid pressure P(ρ) within the gear pump housing in space, calculated using the formula ∇P(ρ)=∂P(ρ) / ∂ρ. It reflects the pressure variation trend at different locations and is used in density field updates to make targeted adjustments to areas of drastic pressure changes, thereby optimizing the housing performance under fluid loads.

[0061] In some embodiments, the weight parameter and In the density field update formula middle, Used to adjust CNN prediction results ( ) and the iterative results of the SIMP method ( The weighting percentage of ) Used to control fluid pressure gradient ( The degree of influence of α and β on the update results. By adjusting α and β, global prediction, local optimization, and fluid pressure adaptability can be balanced to achieve better structural performance.

[0062] Predicting density distribution by fusing CNNs SIMP method iterative density field and fluid pressure gradient This effectively balances global structural optimization with local detail adjustments.

[0063] By using a trained CNN to predict the density field, structural optimization trends can be quickly captured. Compared with simply relying on the SIMP method for iteration, the computation time is shortened while achieving the same optimization accuracy. and The weight parameters can be flexibly adjusted to optimize the focus, for example, by increasing... Value enhancement strengthens global prediction and accelerates convergence, making it suitable for quickly obtaining preliminary optimization results. Correction term based on fluid pressure gradient ∇P(ρ) It can adaptively adjust density in areas with drastic pressure changes (such as flow channel corners and inlet / outlet junctions). Under dynamic pressure fluctuation loads, the vibration amplitude of the housing is reduced, effectively avoiding the risk of resonance and improving the operational stability and reliability of the gear pump. A hybrid update strategy avoids the limitations of a single algorithm: CNN provides a global optimization direction, SIMP ensures local detail accuracy, and pressure gradient correction enhances fluid dynamics adaptability. Under complex load or material distribution scenarios, the stability of the optimization results is improved, reducing the risk of optimization failure due to initial parameter settings and enhancing the algorithm's versatility.

[0064] In some embodiments, the functional expression of the optimization objective is as follows: ;in, For the purpose of compliance, For the maximum equivalent stress, This refers to the amount of fluid leakage. Vibration and noise index; weighting coefficient , , , All of these are weighting coefficients, and optimization must satisfy at least one of the following constraints: volume constraint, fatigue life constraint, and flow channel pressure loss constraint.

[0065] In some embodiments, the volume constraint may be 0.65 or 0.7 of the initial volume. The fatigue life constraint may be a cycle count greater than or equal to 10 to the power of 7. The flow channel pressure loss constraint may be a pressure difference less than or equal to 0.5 MPa.

[0066] In some embodiments, : : : =0.4:0.3:0.2:0.1.

[0067] The embodiments disclosed herein enhance the overall competitiveness of the product: by optimizing the multi-objective function and satisfying the constraints, the gear pump housing achieves a balance in terms of strength, sealing performance, noise control, and lightweighting, significantly improving the overall performance of the product, which helps to enhance the product's competitiveness in the market and reduce the company's R&D and production costs.

[0068] The embodiments disclosed herein achieve flexible and controllable optimization process: the setting of weight coefficients gives designers the ability to flexibly adjust the optimization focus, and can quickly adjust the priority of optimization objectives according to different application scenarios and needs, thereby efficiently obtaining optimization solutions that meet specific needs and shortening the product development cycle.

[0069] In some embodiments, the optimization objective specifically includes at least one of the following: Minimizing compliance as the compliance objective: the objective function can be expressed as follows: ,or, ;in The structural displacement vector. The stiffness matrix is ​​used as the optimization objective to reduce structural deformation. Stress constraints: The objective function may include a penalty term. ,in This represents the stress distribution corresponding to the density field. The penalty term is set as the allowable stress to achieve the optimization objective of controlling the stress to not exceed the safety threshold. Volume constraints: through Controlling material volume corresponds to the optimization objective of quantitative design. For the initial volume, This is the limiting volume.

[0070] In some embodiments, compliance Quantification is achieved through formulas, where the structural displacement vector The stiffness matrix reflects the deformation of each node under load. The material properties of the gear pump housing (such as the mixed parameters of aluminum alloy and steel) and density field And determined by geometry. During optimization, this is achieved by reducing... The numerical values ​​reduce the deformation of the shell under loads such as fluid pressure and dynamic fluctuations, ensuring structural stability. For example, in finite element analysis, the shell is discretized into multiple elements, and the density of each element is... This affects its stiffness contribution. For example, in the high-stress region where the pump chamber contacts the gear, increasing the local field density... To increase Values ​​that reduce structural deformation; in non-critical areas (such as shell edges), appropriately reduce To achieve lightweight design. When the target is constrained by stress Penalty coefficient (e.g.) =10 4 , =10 2 ), amplifying the effect of stress exceeding limits in the region.

[0071] Minimizing compliance reduces the deformation of the housing under complex loads, decreasing the risk of seal failure due to excessive deformation. Stress constraint penalties keep the maximum equivalent stress within a safe threshold, reducing stress concentration, extending fatigue life, and effectively preventing structural fracture. Volume constraints force material optimization, resulting in a lighter gear pump housing compared to traditional designs, reducing the overall load on the equipment. This is particularly suitable for weight-sensitive applications such as aerospace and automotive, while also reducing material costs. In some embodiments, the objective functions and constraints form a quantitative closed loop: the compliance formula is directly related to structural deformation, the stress penalty term dynamically adjusts material distribution, and the volume constraint limits the optimization boundary. By adjusting the parameters, different working conditions can be quickly adapted to ensure that the optimization results conform to engineering practice. In this embodiment, the multi-objective collaborative optimization capability involves three optimization objectives working together to avoid the limitations of single-objective optimization. For example, while reducing compliance, stress constraints prevent local stress overload, and volume constraints ensure lightweighting, achieving a comprehensive optimization of structural performance, safety, and cost.

[0072] In some embodiments, the reconstructed three-dimensional model includes: optimized field density, pump cavity inner wall thickness distribution parameters, and / or diaphragm porosity; The method further includes: Based on the optimized density field Automatically generate support structures in areas where the suspension angle is greater than a specified value, with the diameter of the support structure being... Related to density field: Spacing s and porosity of the support structure Related; Using a layered manufacturing strategy, the layer thickness of the pump cavity inner wall at the current location... Based on fluid pressure gradient Dynamic adjustment: ;in, The current layer fluid pressure gradient, This represents the global maximum pressure gradient. In critical areas such as the inner wall of the pump cavity, a circular scanning path is used; in linear structural areas such as reinforcing ribs, a parallel scanning path is used. Aluminum alloy and steel composite material is used for printing in high-pressure sensitive areas; wherein the steel ratio is set according to the pressure resistance required by the high-pressure sensitive areas. Gradient material printing is achieved by using pure aluminum alloy materials in low-pressure sensitive areas.

[0073] For example, the suspension angle is greater than 30 degrees or 45 degrees.

[0074] Diameter of the supporting structure The relationship with the density field is: ;in, This represents the current region density. In some cases, =0.2, =0.8, but it can also take other values, ensuring... Greater than That's all.

[0075] The supporting structure is an auxiliary structure for the 3D-printed gear pump housing.

[0076] Spacing s and porosity of the support structure The relationship is: . For any specified coefficient.

[0077] In some embodiments, the reconstructed three-dimensional model includes: optimized field density, pump cavity inner wall thickness distribution parameters, and / or baffle porosity; when the optimization target is achieved, fabricating the gear pump housing based on the reconstructed three-dimensional model includes: A support structure to assist in the fabrication of the gear pump housing is fabricated based on the optimized density field and the porosity of the partition. Using the aforementioned support structure, the gear pump housing is 3D printed based on the reconstructed 3D model; wherein, during the 3D printing, a circular scanning path is used in the pump cavity inner wall and the partition area; and a parallel scanning path is used in linear structural areas such as reinforcing ribs.

[0078] In some embodiments, the 3D printing path planning steps are as follows: A layered manufacturing strategy is adopted, with the layer thickness t determined based on the fluid pressure gradient. Dynamic adjustment: .in, The current layer fluid pressure gradient, This represents the global maximum pressure gradient.

[0079] In critical areas, a circular scanning path is used; in linear structural areas such as reinforcing ribs, a parallel scanning path is used. For example, the critical area may include the inner wall of the pump chamber and / or a partition. For example, this scanning path is also the printing path for 3D printing.

[0080] In the multi-material printing step, according to the material gradient distribution instructions, an aluminum alloy-steel composite material is used in the high-pressure sensitive area, wherein the volume percentage of steel is: ;in This represents the maximum pressure in the low-pressure sensitive area. This represents the minimum pressure value in the high-pressure sensitive area.

[0081] In subsequent steps, the printed part is placed in an alkaline solution at 50-60℃ to dissolve the support structure. The dissolution time... The dissolution time is positively correlated with the volume of the supporting structure.

[0082] Electrochemical polishing was performed on the inner wall of the pump chamber.

[0083] In some embodiments, the 3D printing employs a photopolymerization resin process, with specific parameters including: light source wavelength: 405nm ultraviolet light; exposure time per layer: dynamically adjusted according to the resin curing rate v (unit: mm / s). Exposure time is positively correlated with thickness and negatively correlated with v. Printing platform temperature: 30-40℃, maintained at a constant temperature by a PID controller.

[0084] In some embodiments, deformation is monitored and compensated in real time during the printing process, specifically including: After each layer is printed, the actual outline is obtained using a laser scanner. and with the design outline Compare and calculate the deviation. When the current printing is done If the absolute value is greater than the replenishment threshold, then compensation will be made in the next level of the printing path. A predetermined proportion (e.g., 50%). In some embodiments, the compensation formula is: . It is the design outline; It is a compensation contour.

[0085] In summary, the support diameter With density field Negative correlation ensures thicker support in low-density (easily deformable) areas; support spacing With porosity A positive correlation exists; the higher the porosity (the less material), the sparser the support.

[0086] During the dynamic adjustment phase of the printing process, the layer thickness t varies with the pressure gradient. Increase and decrease, use thinner layer thickness in high-pressure areas to improve accuracy; exposure time t is related to material density. Related measures are taken to ensure that high-density areas are fully cured.

[0087] In the stage of realizing multi-material gradient, the proportion of steel in aluminum alloy is precisely controlled by formula to achieve a smooth transition of material properties; the elastic modulus E of the supporting material is related to the volume constraint to avoid cracking of parts due to excessive support.

[0088] A deformation compensation mechanism is used, employing a closed-loop control of "layer-by-layer scanning-comparison-compensation" to ensure that the final part size error is less than the target value.

[0089] like Figure 6 As shown, this disclosure provides a gear pump housing manufacturing apparatus, comprising: Module 5110 is used to construct a finite element model containing design space, boundary conditions and material parameters for the structural optimization requirements of the gear pump housing. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. Optimization module 5120 is used to generate a dataset containing an initial density field, intermediate iteration density fields, and corresponding final optimization results through multiple sets of penalized SIMP method topology optimization simulations of penalized solid isotropic materials with different density configurations. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP method topology optimization simulation uses a density interpolation model: where... ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; Reconstruction module 5130 is used to employ deep learning regression models and based on The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. The printing module 5140 is used to manufacture the gear pump housing based on the reconstructed 3D model when the optimization target is achieved.

[0090] Combination Figure 7 As shown, this application embodiment provides an electronic device including a processor 10 and a memory 11. Optionally, the device may further include a communication interface 12 and a bus 9. The processor 10, communication interface 12, and memory 11 can communicate with each other via the bus 9. The communication interface 12 can be used for information transmission. The processor 10 can call logical instructions in the memory 11 to execute the gear pump housing manufacturing method of the above embodiment.

[0091] Furthermore, the logical instructions in the aforementioned memory 11 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0092] The memory 11, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this application. The processor 10 executes the program instructions / modules stored in the memory 11 to perform functional applications and data processing, thereby implementing the gear pump housing manufacturing method described in the above embodiments. The memory 11 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 11 may include high-speed random access memory and may also include non-volatile memory.

[0093] This electronic device can be used for any electronic device that designs a gear pump housing, such as a server or terminal device.

[0094] This application provides a computer program product, which includes a computer program stored on a storage medium. The computer program includes program instructions, which, when executed by a computer, cause the computer to perform the above-described gear pump housing manufacturing method.

[0095] The aforementioned computer-readable storage medium may be a transient computer-readable storage medium or a non-transitory computer-readable storage medium.

[0096] The technical solutions of this application embodiment can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in this application embodiment. The aforementioned storage medium can be a non-transitory storage medium, including various media capable of storing program code such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks, or it can be a transient storage medium.

[0097] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0098] The embodiments or examples disclosed in this application are not exhaustive, but merely illustrative of some embodiments or examples, and are not intended to limit the scope of protection of this disclosure. Unless contradictory, each step in a particular embodiment or example can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment or example can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment or example can be arbitrarily interchanged. Furthermore, optional methods or examples in a particular embodiment or example can be arbitrarily combined; moreover, embodiments or examples can be arbitrarily combined. For example, some or all steps of different embodiments or examples can be arbitrarily combined, and a particular embodiment or example can be arbitrarily combined with optional methods or examples of other embodiments or examples.

[0099] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0100] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0101] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0102] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0103] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for manufacturing a gear pump housing, characterized in that, include: To address the structural optimization requirements of the gear pump housing, a finite element model was constructed, including design space, boundary conditions, and material parameters. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. Through multiple sets of penalized SIMP topology optimization simulations of isotropic solid materials with different density configurations, a dataset containing an initial density field, intermediate iteration density fields, and corresponding final optimization results is generated. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP topology optimization simulation employs a density interpolation model: where... ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; Using deep learning regression models and based on The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. When the optimization goal is achieved, the gear pump housing is manufactured based on the reconstructed 3D model.

2. The method according to claim 1, characterized in that, The calculation methods for element stiffness matrix and total compliance in the SIMP method topology optimization simulation are as follows: as well as ;in, For the first Stiffness matrix of each element; Let be the stiffness matrix of the solid element; For the total compliance, For the first The flexibility of each unit.

3. The method according to claim 1, characterized in that, To address the structural optimization requirements of the gear pump housing, a finite element model was constructed, including design space, boundary conditions, and material parameters, comprising: The design space is based on the thickness distribution of the pump chamber wall, the layout of the baffles, and the direction of the inlet and outlet flow channels of the gear pump housing as core parameters; Construct the boundary conditions; Set the anisotropy parameters of the aluminum alloy and steel hybrid material.

4. The method according to any one of claims 1 to 3, characterized in that, The final optimization result includes at least one of the following: flow channel pressure loss coefficient based on fluid dynamics simulation; stress concentration factor based on fatigue life prediction; structural vibration frequency response based on modal analysis; The process employs a deep learning regression model, using the initial density field and intermediate iterative density fields as optimization parameters and the final optimization result as the optimization objective, to reconstruct the three-dimensional model of the gear pump housing, including: Using the deep learning regression model, the initial density field, intermediate iteration density field and stress concentration coefficient in the final optimization result are used as inputs to output the pump cavity inner wall thickness distribution parameters, wherein the pump cavity inner wall thickness distribution parameters enable the pump cavity inner wall to adopt a gradient wall thickness structure. Based on the density change rate distribution of the intermediate iteration density field, and combined with the flow channel pressure loss coefficient of the final optimization result, the porosity and distribution direction of the baffle are optimized by a genetic algorithm. The porosity and distribution direction of the baffle make the baffle have a gradient pore lattice structure. By using convolutional neural networks to analyze the boundary features of the initial density field, the local optimization results of the intermediate iteration density field, and the vibration frequency response of the final optimization result, stiffener layout and variable cross-section parameters are generated. Based on the thickness distribution parameters of the pump cavity inner wall, the porosity and distribution direction of the baffle, and the layout and cross-sectional parameters of the reinforcing ribs, a three-dimensional model of the gear pump housing is reconstructed.

5. The method according to claim 4, characterized in that, The deep learning regression model is used, taking the initial density field, intermediate iteration density fields, and stress concentration coefficients from the final optimization result as inputs, and outputting pump cavity inner wall thickness distribution parameters, including: Using the aforementioned deep learning regression model, with the initial density field, intermediate iteration density fields, and stress concentration coefficients from the final optimization result as inputs, the output is the pump cavity inner wall thickness distribution parameters, forming a gradient wall thickness structure that satisfies the following conditions: Wall thickness along the direction of fluid flow Satisfies the exponential decay function: ,in This represents the initial wall thickness at the entrance. For attenuation, Coordinates of the fluid flow path; Localized thickening zones form in areas of concentrated pressure, with the amount of thickening... , It is a nonlinear function that satisfies ∈ First predetermined range; This represents the maximum stress in the pressure concentration area. Wall thickness variation rate Ensure a smooth stress transition within the transition region, where the range is less than or equal to the second predetermined range. Circumferential wall thickness distribution and radial stress components Relationship, satisfy ,in The average wall thickness For amplitude, The harmonic order is... This is the phase angle.

6. The method according to claim 1, characterized in that, The deep learning regression model includes a fully integrated neural network (CNN); the process of reconstructing the three-dimensional model of the gear pump housing using a deep learning regression model, with the initial density field and intermediate iterative density fields as optimization parameters and the final optimization result as the optimization objective, includes: After the SIMP method has been executed up to the preset number of iterations T, the current density field is processed as follows: Input the trained CNN, output the predicted density distribution. ; Calculate the fluid pressure gradient of the current density field. The calculation formula is: ,in The fluid pressure distribution within the gear pump housing; Calculate the density field updated in this iteration ; use The density field is updated to balance global prediction and local optimization while simultaneously optimizing the structural performance of the gear pump housing under fluid pressure loads and dynamic pressure fluctuation loads; wherein, For the improved ; in, Predict weights for CNN; Adjust the weights for the gradient.

7. The method according to claim 1, 2, or 6, characterized in that, The functional expression of the optimization objective is as follows: ;in, For the purpose of compliance, For the maximum equivalent stress, This refers to the amount of fluid leakage. Vibration and noise index; weighting coefficient , , , All of these are weighting coefficients, and optimization must satisfy at least one of the following constraints: volume constraint, fatigue life constraint, and flow channel pressure loss constraint.

8. The method according to any one of claims 1 to 3, characterized in that, The reconstructed 3D model includes: optimized field density, pump cavity inner wall thickness distribution parameters, and / or baffle porosity; when the optimization target is achieved, the gear pump housing is fabricated based on the reconstructed 3D model, including: A support structure to assist in the fabrication of the gear pump housing is fabricated based on the optimized density field and the porosity of the partition. Using the aforementioned support structure, the gear pump housing is 3D printed based on the reconstructed 3D model; wherein, during the 3D printing, a circular scanning path is used in the pump cavity inner wall and partition area; and a parallel scanning path is used in linear structural areas such as reinforcing ribs.

9. A gear pump housing manufacturing apparatus, characterized in that, include: The construction module is used to build a finite element model containing design space, boundary conditions and material parameters for the structural optimization requirements of the gear pump housing. The boundary conditions include steady-state fluid pressure load and transient dynamic pressure fluctuation load, and the material parameters are the properties of a hybrid material of aluminum alloy and steel. The optimization module is used to generate a dataset containing an initial density field, intermediate iteration density fields, and corresponding final optimization results through multiple sets of penalized SIMP method topology optimization simulations of penalized solid isotropic materials with different density configurations. The final optimization results include compliance, stress distribution, and fluid pressure distribution. The SIMP method topology optimization simulation uses a density interpolation model: where... ; The first in the design space where the gear pump housing is located The relative density of each unit; As a penalty factor, This is the interpolated element elastic modulus. The elastic modulus of the solid element. The elastic modulus of the void element; The total number of units contained in the design space; The reconstruction module is used to employ deep learning regression models and based on... The three-dimensional model of the gear pump housing is reconstructed using the initial density field and intermediate iterative density field as optimization parameters and the final optimization result as the optimization objective. This is the density field after this iteration update; This is the density field of this SIMP method iteration; The density field is obtained by iterating the deep learning regression model; α is a preset empirical value. A printing module is used to manufacture the gear pump housing based on the reconstructed 3D model when the optimization target is achieved.

10. The apparatus according to claim 9, characterized in that, The calculation methods for element stiffness matrix and total compliance in the SIMP method topology optimization simulation are as follows: as well as ; in, For the first Stiffness matrix of each element; Let be the stiffness matrix of the solid element; For the total compliance, For the first The flexibility of each unit.

Images