A Time-Space Multi-Scale Parallel Method for Dynamic Behavior Analysis of Complex Power System

摘要

There are two difficulties of dynamic behavior computations, one is that the speed and memory size are often insufficient, and the other is the coupling of different scales for space and time. The objective of this paper is to propose a general time-space multi-scale parallel method for solving system dynamic problems. The first step is to partition the domain into subdomains using a dual Schur discretized formulation. The second is dividing the corresponding unknowns into a free term obtained from internal and external loads and the link term obtained from the link loads. And then the model can be uncoupled, such as the equilibrium equations, on each subdomain into a “free problem” and a “link problem”. For the different time steps, two operators must be defined in order to couple the macro timescale and the micro timescale. For the different space scale, two space scales on the fine subdomain to connect the different scales on interface of two adjacent subdomains also must be defined. Finally, there is a brief presentation of the selective catalytic reduction (SCR) system, which is going to be solved by using the method of this text in our future study.