虚假关系
数值积分
数值分析
数值稳定性
泰勒级数
运动方程
理论(学习稳定性)
动力系统理论
运动(物理)
数学
微分方程
计算机科学
应用数学
数学分析
经典力学
物理
人工智能
统计
机器学习
量子力学
作者
Anton Ianakiev,I.I. Esat
摘要
Numerical solution of dynamical systems with widely varying motion characteristics, such as relatively slow motion coupled with high frequency as it would be in flexible mechanisms, likely to pose problems. In this paper the mathematical model of a flexible mechanism is solved by using a mixed integration method that attempts to deal with the complexity of the coupled differential equations of the rigid-body and elastic motion. The mixed integration method consists of two integration methods (Rossenbrock-Wanner and Hilber-Hughes-Taylor methods) that have been combined in order to minimize the computational complexity required for the approximation of the real system. The a Hilber-Hughes-Taylor methods incorporates numerical damping that selectively affects only the higher modes of vibration. The improvement of the stability and the accuracy of the solution due to the numerical damping has been demonstrated via a numerical example that represents a stiff system. The example system was selected to contain a physically important low frequency and spurious highly frequency oscillations. The solution method filtered the high numerical oscillations from the response results. The Rossenbrock-Wanner integration technique was also presented. In this case it is also shown that fine adjustment of integration parameters could effect the degree of numerical damping. A mixedmore » integration method, combination of the two found to give the best performance and accuracy in the case of stiff problems.« less
科研通智能强力驱动
Strongly Powered by AbleSci AI