Thermal-structure Finite Element Simulation System Architecture in a Cloud-Edge-end Collaborative Environment

Jialan Liu,Chi Ma,Shilong Wang

DOI: https://doi.org/10.1007/s10845-023-02269-z

IF: 8.3

2023-01-01

Journal of Intelligent Manufacturing

Abstract:In response to the urgent need for finite element simulation, a thermal-structure finite element simulation system architecture is designed to shorten the simulation cycle and improve the mechanical structure design efficiency under a cloud-edge-end collaborative environment. Then, a calculation kernel of the boundary conditions is proposed, and a thermal-structure closed-loop iterative model is established and embedded into the finite element simulation system. The interactions among the simulation results and boundary conditions are considered, and the boundary conditions are corrected by the simulation results. Finally, the thermal-structure finite element simulation system architecture is verified. The optimal configuration of the central processing unit number and memory size for different finite element models is identified, and the simulation efficiency and throughput are improved significantly. In addition, the proposed thermal-structure finite element simulation system architecture with a cloud-edge-end is applied to conduct the thermal-structure behavior simulation for the feed drive system, spindle system, precision horizontal machining center, and gantry machining center. The machine tool designer without specialized knowledge about thermal-structure simulation is able to achieve a high simulation accuracy at the design stage, and the executing performance of the proposed thermal-structure finite element simulation system with cloud-edge-end architecture is far higher than that of the thermal-structure finite element simulation systems with cloud-end and cloud architectures. With the implementation of the simulation system with the cloud-edge-end architecture, the execution time is reduced from 2557 and 2082s to 1642 s as compared with the simulation systems with the cloud-end and cloud architectures, respectively. The simulation kernel is effective in simulating thermal-structure behaviors. The average deviations between the measured and simulated temperatures for the rear and front bearings are 4.36% and 3.15%, respectively. The average deviation between the measured and simulated deformations is 8.17%.