Power System Dynamic Simulations Using a Parallel Two-Level Schur-Complement Decomposition

Abstract

Abstract-As the need for faster power system dynamic simulations increases, it is essential to develop new algorithms that exploit parallel computing to accelerate those simulations. This paper proposes a parallel algorithm based on a two-level, Schurcomplement-based, domain decomposition method. The twolevel partitioning provides high parallelization potential (coarseand fine-grained). In addition, due to the Schur-complement approach used to update the sub-domain interface variables, the algorithm exhibits high global convergence rate. Finally, it provides significant numerical and computational acceleration. The algorithm is implemented using the shared-memory parallel programming model, targeting inexpensive multi-core machines. Its performance is reported on a real system as well as on a large test system combining transmission and distribution networks.

Online Estimation of Power System Inertia Using Dynamic Regressor Extension and Mixing

“…2) The performance of the estimator is demonstrated on a 1013-machine ENTSO-E test system with 21382 buses and 133997 states, which is implemented in the dynamic simulation software RAMSES [22]. As with any dynamic parameter estimation method [23]- [25], also with DREM a sufficiently large system excitation is required for an accurate estimation [18].…”

“…Fig. 6: Trajectories of ∆fĤ tot av and ∆f Htot av defined in(22) for the outage of the unit 'DE912342' in Germany with S B645 = 2154 MW and P m154 = 1078.5 MW.…”

A Relaxation-Based Network Decomposition Algorithm for Parallel Transient Stability Simulation with Improved Convergence

“…In the fine-grained approach [6], [7], [11][12][13][14], the parallelism is mainly obtained through matrix/vector reduction techniques to exploit the structural properties (such as sparsity and repetition) of the linear power system network matrices and make it suitable for parallel implementation. On the other hand, the coarse-grained approach [9], [16][17][18], [20][21][22][23][24][25][26][27] is applied directly to the system of equations. Compared with fine-grained methods, coarsegrained methods have manifested various advantages, such as: the parallelism is not limited by the block structure of the system, localized information for the subproblems is kept, subproblems can be solved using locally adapted simulation settings, etc.…”

“…[10]. While coarse-grained parallelism can be further exploited from different perspectives such as spatial parallelism [9], [17][18], [23][24] , time paral-lelism and spatial-time parallelism (e.g. waveform relaxation) [16], [20][21], [25][26][27], in this paper, we mainly focus on the spatial parallelism (parallel-in-space) which incorporates a partitioning scheme to first decouple the original problem into various pieces then solve each of the partitioned subproblems in parallel at a given time instant.…”

“…In the third step, all the subproblems need to be reconciled through the interface variables to derive a coordinated, converged global solution. Based on the way the interface variables are computed, spatial DD techniques can be classified into two main approaches: Schur-complement method [9], [17], [18], [24] and Schwartz alternating method [23].…”