Diakoptics-a parallel processing approach for the analysis of large-scale power systems

The structure, the quality and the dimension of the power systems are subject to change and to become more complex. For the analysis of such vast power systems the use of computers is indispensable. In the computer-aided power system analysis, the computational cost and the accuracy are two important factors. The computational cost, i.e. computational time and memory needed for the analysis, are related to the dimension of system being analyzed. The increase in the system dimension will directly increase these requirements and creates several difficulties. Some new approaches for the mathematical formulation of power systems have been considered to find a more convenient way to simulate large-scale power systems. However, a literature survey on the analyses of large-scale power systems has shown that the advantages to be gained by the use of the concept of multi-terminal component introduced in network theory have not been fully utilised. This paper deals with the application of the concept of multi-terminal components to the analysis of large-scale power systems. The new approach, which is perfectly general and can be applied to both the analysis of asymmetric and symmetric systems, aims to obtain the coefficient matrix of the system as a block-diagonal-band matrix. Instead of solving the whole matrix by using a processor, blocks representing the sub-networks can be solved simultaneously via parallel processors to speed up the process