【作者】 李浦； 袁奇； 逯代兴；

【Author】 Li Pu;Yuan Qi;Lu Daixing;Xi’an Jiaotong University;Shanghai Institute of Technology;

【机构】 西安交通大学； 上海应用技术大学；

【摘要】 协同仿真算法广泛用于多物理域耦合分析和多求解器并行计算,已成功用于汽车基础耦合、高速列车弓网耦合等分析。根据耦合变量在多求解器传递是否迭代分为显式算法和隐式算法,隐式算法数值稳定性高但迭代求解耗时严重,难以直接集成现有商业软件进行耦合分析。显式协同仿真算法由于界面误差传递所引起的数值稳定性问题制约积分步长和计算效率,已成为当下商业软件接口和第三方耦合平台发展的瓶颈问题。本文采用线性双自由度模型,定量分析线性系统数值稳定性;基于常加速度假设提出改进显式协同仿真算法,相比经典算法数值稳定性得到极大提高;基于改进协同仿真算法建立分布式多体系统控制方程,采用本文提出的显式仿真算法对两个算例进行验证分析,为大型多体系统并行求解提供高稳高效算法。更多还原

【Abstract】 Co-simulation has been widely applied in multi-physical analysis as well as parallelization of multi-solvers. It has been used for vehicle-ground interaction and Pantograph catenary dynamic simulations. According to whether coupling variables are repeated or not we can distinguish explicit and implicit algorithms. Implicit methods entail a superior stability behavior however, with larger computational costs. Therefore, implicit methods can hardly be applied to couple commercial tools due to the iterative characteristics. Explicit algorithms suffer from a severe stability problem induced by interface errors. As a result, the simulation stepsize is restricted and the computational efficiency is largely reduced that has become a major problem for the coupling software as well as the interface platform. In this paper a linear model with two degrees of freedom is introduced and the numerical stability has been analyzed using 2 D stability plots. Constant acceleration approximation is applied to obtain the coupling variable and the stability behavior is hugely enhanced. Furthermore, the governing equations of motion as well as coupling conditions for two general multibody systems have been derived. Simulations of two multibody models have been performed which provides a more efficient method for large-scaled multibody systems with better stability.更多还原