An Ultra-High Speed Emulator Dedicated to Power System Dynamics Computation Based on a Mixed-Signal Hardware Platform

Abstract

Abstract-ThisIndex Terms-Application specific integrated circuit (ASIC), emulation, mixed analog digital integrated circuits, power system dynamics, power system simulation, power system stability.

Pipelined Numerical Integration on Reduced Accuracy Architectures for Power System Transient Simulations

“…However all of the aforementioned are in the digital domain. The authors of this work have explored the use of analog (unconventional) computing as a means of solving the linear system [4][5]. The sparse matrix Y is handled by an analog lattice computer that is briefly described in section III.…”

“…Mapping the topology of a power grid on a miniaturized electronic implementation has been proved feasible in [4]. Starting from (1) and neglecting the conductance under the assumptionԡ‫ܩ‬ԡ ‫ا‬ ԡ‫ܤ‬ԡ for the transmission system:…”

“…In order to tackle the problem in a computationally efficient way, the authors have been recently involved in the development of a transient simulator that is not based on general purpose computing units but on dedicated hardware instead [4][5]. This platform combines an analog and a digital computer part, i.e.…”

“…The analog part acts as an extremely parallelized linear system solver with very promising timing results [4][5][6]. However, imprecisions in the solution of the linear system are introduced by the imperfect reconfigurable elements (potentiometers, switches) and converters (ADCs, DACs) of the analog computer.…”

“…The pipelined architecture is an attractive integration scheme for dedicated hardware platforms as it is a tradeoff between a fully parallel and a fully sequential approach. It combines merits of both in yielding substantial resource savings against the former and significant computational benefits against the latter [4]. Digital imprecision sources also exist, and include the utilization of fixed-point arithmetic on the FPGA, as well as the inherent inaccuracy of numerical integration algorithms.…”