CN-121835436-B - Design method and system for macro-micro double-scale heat exchanger

CN121835436BCN 121835436 BCN121835436 BCN 121835436BCN-121835436-B

Abstract

The invention belongs to the technical field of heat exchanger design, and provides a macro-micro double-scale heat exchanger design method and a system, wherein initial TPMS parameters are independently decided based on physical rules and process constraints according to design parameters, and an initial fluid calculation domain is constructed; the method comprises the steps of solving a generated fluid calculation domain, obtaining data of a speed vector field, a static pressure field and a temperature field of a full-field discrete grid point, carrying out topology analysis, generating structural feature semantics, carrying out fine adjustment on baffle layout by utilizing a macroscopic expert agent, generating a baffle modification proposal, monitoring a macroscopic optimization process, judging that a current microscopic matrix is not matched when the baffle adjustment is detected to fall into a deadlock state, generating a linear scaling lattice period or switching a TPMS topology type proposal, evaluating the generated proposal, selecting an optimal proposal, updating according to the selected proposal, and continuously iterating until the iteration requirement is met. The invention can improve the performance of the TPMS heat exchanger.

Inventors

LV LIN
DUAN XIAOWEI
YAN XIN

Assignees

山东大学

Dates

Publication Date: 20260512
Application Date: 20260306

Claims (9)

1. The design method of the macro-micro double-scale heat exchanger is characterized by comprising the following steps of: (1) Acquiring design parameters, autonomously deciding initial TPMS parameters by utilizing a microstructure intelligent agent based on physical rules and process constraints according to the design parameters, generating an initial TPMS background field according to the initial TPMS parameters, and constructing an initial fluid calculation domain; (2) Solving the generated fluid calculation domain based on a Navier-Stokes equation set and an energy conservation equation to obtain data of a speed vector field, a static pressure field and a temperature field of the full-field discrete grid points; (3) Carrying out topology analysis on data of the speed vector field, the static pressure field and the temperature field to generate structural feature semantics; (4) Constructing a macroscopic expert intelligent body, and performing fine adjustment on the layout of the baffle by utilizing the macroscopic expert intelligent body to generate a baffle modification proposal; (5) Monitoring a macroscopic optimization process, and when a baffle plate adjustment is detected to be in a deadlock state, judging that the current microscopic substrates are not matched, and generating a proposal of linearly scaling a lattice period or switching the TPMS topology type; (6) Evaluating the generated proposal, selecting an optimal proposal, updating according to the selected proposal, returning to the step (2) for the next iteration until the iteration requirement is met, and obtaining a final proposal; In the step (4), the process of constructing the macroscopic expert agent comprises the following steps: When receiving the coordinates of a short circuit path, generating an increased flow baffle proposal, wherein the position is set at the position of the short circuit path and the angle is perpendicular to the flow direction so as to forcibly cut off the straight-through flow path; and constructing a hydraulic expert intelligent body, aiming at reducing the total pressure drop, and generating a geometric modification proposal for removing the baffle, forming pressure relief holes on the surface of the baffle or reducing the area of the baffle according to the current pressure drop exceeding degree when the index of the high-resistance bottleneck baffle is received.
2. The method of claim 1, wherein the step (1) of obtaining design parameters includes parsing and extracting key design parameters according to user input requirements, including identifying fluid working fluid properties, inlet flow rate, temperature and maximum pressure drop threshold allowed by design, and identifying minimum feature size allowed by additive manufacturing as TPMS micro-element wall thickness.
3. The method for designing a macro-micro double-scale heat exchanger according to claim 1, wherein in the step (1), the process of autonomously deciding the initial TPMS parameters based on physical rules and process constraints by utilizing the microstructure intelligent body comprises the steps of establishing an association equation of pressure drop and flow channel geometric characteristics based on Darcy-Wei Siba Hz flow resistance theory, and reversely solving the minimum hydraulic diameter meeting the flow resistance requirement according to the set maximum pressure drop threshold; Converting the hydraulic diameter into a theoretical lattice period under the current topology type by utilizing an empirical analysis relation of the hydraulic diameter and the lattice period, which is established in advance; Decomposing the theoretical lattice period into a solid wall thickness and a pore diameter, checking whether the theoretical lattice period meets the extracted minimum feature size, if not, checking failure, automatically triggering topology switching, and selecting a heat exchange infinitesimal with more transparent flow passage until the final manufacturable initial lattice period is determined; and generating a TPMS background field with selected parameters by using an implicit modeling technology, generating an initial baffle group according to the length-diameter ratio of the flow channel, and constructing an initial fluid calculation domain.
4. The method of designing a macro-micro double-scale heat exchanger according to claim 1, wherein the process of the step (3) includes solving the generated initial fluid calculation domain based on a Navier-Stokes equation set and an energy conservation equation by using a computational fluid mechanics solver to obtain velocity vector field, static pressure field and temperature field data of the full-field discrete grid points; and carrying out topology analysis on the three-dimensional field data to generate structural feature semantics, and respectively extracting dead zone features, short circuit features and high resistance features.
5. The method for designing a macro-micro double-scale heat exchanger according to claim 4, wherein the process of extracting dead zone features comprises the steps of dynamically calculating an adaptive speed threshold value for adapting to current flow field features by an agent through statistical analysis of full field speed distribution, screening out a grid cell set with the full field speed lower than the threshold value, clustering low-speed cells scattered in space into a plurality of independent flow stagnation domains by using a spatial clustering algorithm based on density noise and using a grid cell center distance as a neighborhood radius, and calculating geometric centroid coordinates and volume ratio of each stagnation domain as decision input of a thermal expert agent.
6. The method of designing a macro-micro double-scale heat exchanger according to claim 4, wherein the process of extracting the dead zone features includes the steps of adopting a Lagrange particle tracking method to release trace particles from an inlet, recording residence time of each particle, screening out a fast particle group with residence time lower than a threshold value, extracting motion tracks of the fast particle group, calculating a spatial center path point of the particle group by a weighted average method, and reconstructing the discrete path points into three-dimensional skeleton line coordinates of a high-speed main flow channel.
7. The method of designing a macro-micro dual-scale heat exchanger according to claim 4, wherein the process of extracting the dead zone feature includes analyzing the distribution of the along-the-path pressure gradient, calculating the partial pressure drop before and after each baffle, identifying the baffle generating the maximum partial pressure difference as a flow resistance bottleneck, and using the baffle as a target object for the hydraulic expert agent to perform the opening pressure relief or removal operation preferentially.
8. The method for designing a macro-micro double-scale heat exchanger according to claim 1, wherein the step (6) evaluates the generated proposal and the process of selecting the optimal proposal includes preferentially adopting a macro baffle modification proposal and only approving a topology reconfiguration proposal of a micro intelligent agent when the history memory bank shows that the current strategy is in a locally optimal cycle; If the thermal expert agent and the hydraulic expert agent propose contradictory geometric modification proposals aiming at the same area, or when a single proposal possibly leads to the deterioration of another performance index, the total engineer agent calculates the comprehensive evaluation index, predicts the benefit risk ratio of each proposal and preferentially executes the proposal.
9. A macro-micro dual scale heat exchanger design system employing the method of any one of claims 1-8, comprising: the heat exchange micro-element initialization module is configured to acquire design parameters, autonomously decide initial TPMS parameters based on physical rules and process constraints by utilizing a microstructure intelligent agent according to the design parameters, generate an initial TPMS background field according to the initial TPMS parameters, and construct an initial fluid calculation domain; The characteristic extraction module is configured to solve the generated fluid calculation domain based on the Navier-Stokes equation set and the energy conservation equation to obtain data of a speed vector field, a static pressure field and a temperature field of the full-field discrete grid points; The system comprises a double-scale game strategy generation module, a macro optimization process monitoring and control module, a linear scaling lattice period generation module and a TPMS topology type switching module, wherein the double-scale game strategy generation module is configured to construct a macro expert intelligent agent, and fine-tune the baffle layout by utilizing the macro expert intelligent agent to generate a baffle modification proposal; the closed-loop decision module is configured to evaluate the generated proposal, select an optimal proposal, update according to the selected proposal, return to the calling feature extraction module, and perform the next iteration until the iteration requirement is met, so as to obtain a final scheme.

Description

Design method and system for macro-micro double-scale heat exchanger Technical Field The invention belongs to the technical field of heat exchanger design, and particularly relates to a macro-micro double-scale heat exchanger design method and system. Background The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. With the continuous improvement of the power density requirements of the heat management system in the fields of aerospace, new energy automobiles, high-performance computing servers, national defense equipment and the like, the heat exchange capability of the traditional shell-and-tube or plate-fin heat exchanger in an extremely compact space gradually reaches the bottleneck. In recent years, with the maturation of additive manufacturing technology, three-period extremely small curved surface (Triply Periodic Minimal Surfaces, TPMS) heat exchangers become research hot spots of next-generation high-efficiency compact heat exchangers due to the advantages of extremely high specific surface area, good topological connectivity, strong self-supporting capability and the like. Common TPMS structures such as Gyroid, diamond and Schwarz P can divide a calculation domain into two fluid domains which are isolated and entangled with each other naturally, and heat exchange among liquids is greatly promoted. At present, aiming at the performance optimization design of the TPMS heat exchanger, the prior art mainly shows the limitation of macro-micro fracture. On the one hand, the research on the microscopic level is mainly focused on the design of the functionally graded TPMS, i.e. the cell porosity or period is spatially continuously varied by mathematical mapping. However, the method faces a serious challenge in engineering landing, complicated gradient modeling is extremely easy to cause over-thin local wall thickness or over-small pores in the 3D printing process, so that powder is difficult to clean and even the structure collapses, the manufacturing yield is low, the guiding capability of a macroscopic flow field is very limited by only fine adjustment of microscopic pores, and the problems of common penetrating short circuit and large-area flow dead zone in a complicated flow channel are difficult to fundamentally solve, so that the effective utilization rate of the internal space of the heat exchanger is insufficient. On the other hand, in order to improve macroscopic flow field distribution, the introduction of a flow guide baffle is an effective means. However, most of the existing baffle designs rely on experience trial and error of engineers, or adopt a genetic algorithm, deep reinforcement learning and other black box intelligent algorithms to automatically optimize. These algorithms have significant drawbacks in practical applications, they typically do blind searches based on random strategies, requiring hundreds or thousands of high-fidelity computational fluid dynamics (Computational Fluid Dynamics, CFD) simulation iterations, and are not time-consuming. More seriously, the traditional algorithm only focuses on scalar results such as total pressure drop, and the like, so that massive three-dimensional unstructured field data generated by CFD simulation cannot be effectively processed and understood, and key physical information such as speed vectors, temperature gradients and the like is ignored. The end-to-end mode of the cost is that the algorithm cannot identify specific physical characteristics such as dead zones, short circuits and the like human experts, the generated scheme is often in a strange shape, and the coupling effect of the microscopic basal body and the macroscopic baffle is difficult to consider. Disclosure of Invention In order to solve the problems, the invention provides a macro-micro double-scale heat exchanger design method and a system, which fully utilize a micro-selection strategy and a targeting regulation mechanism of macro flow field characteristics under the constraint of an additive manufacturing process, realize the intelligent design of a heat exchanger structure by introducing a multi-agent framework simulating a human engineer cognitive decision process, perform cross-scale logic reasoning and collaborative decision, convert blind global search into accurate local operation, greatly reduce the dependence times on CFD simulation, ensure that the design scheme obtains the best balance among physical performance, manufacturing feasibility and structural robustness, and provide a brand-new technical path for the low-cost and high-efficiency development of the high-performance TPMS heat exchanger. According to some embodiments, the present invention employs the following technical solutions: A design method of a macro-micro double-scale heat exchanger comprises the following steps: (1) Acquiring design parameters, autonomously deciding initial TPMS parameters by utilizi