Floating Vector Processor for Power System Simulation

This paper analyzes microprogram functions of the sparse matrix solution to obtain faster power system simulations such as the power flow calculation and transient stability analysis. The hardware structure and extended instructions of a floating vector processor (FVP) which realize simulation speed up are presented. Experimental results of a sparse matrix calculation and a power flow calculation lead to the following results: (1) the sparse matrix calculation, by the FVP is completed about 1.6 times faster than by a conventional array processor; (2) the processing speed of the FVP is ten to fifteen times faster than that of the general purpose supermini- computer Hitachi V90/50 Control Computer on the conditions that the programming style of the programs executed by the FVP is microprogramming fit to the FVP architecture, and the programs executed by the V90/50 are programmed in Fortran with simple optimization; and (3) the speed of the total power flow calculation by the V90/50 with FVP is 2.6 times faster than that by the V90/50 alone.