Stability of Large Scale Systems with High Order Model Reduction

尽管在模型降阶方面取得了一些成果，但用于动态分析的电力系统的仿真规模非常大。因此，需要更快的模拟方法。微分代数系统的一般最小残差法与同时隐式（SI）方法相结合是一种很有前途的方法。它保证在几乎n次迭代中收敛，并且易于并行化。我们计划证明这种方法作为SI方法中LU分解方法的替代方法的可行性。研究的第二个目标是应用轨迹灵敏度分析的丰富结果来判断系统在给定应力条件下是否接近不稳定。最后我们将比较结构稳定性、李雅普诺夫稳定性和灵敏度理论在电力系统中的应用。＊＊＊

9522547 Pai In spite of the achievements in model order reduction, the size of power systems to be simulated for dynamic analysis are very large. Hence faster methods of simulation are needed. A promising method is the general minimal residual method in conjunction with the simultaneous-implicit (SI) method for differential algebraic systems. It is guaranteed to converge in almost n iterations and easy to parallelize. We plan to demonstrate the viability of this approach as an alternative to LU decomposition method in the SI method. A second objective of the research is to apply the rich results of trajectory sensitivity analysis for judging if a system is close to instability for a given stressed condition. Finally we will compare the structural stability, Lyapunov stability and sensitivity theory as applied to power systems. ***