The usage of parallel calculations in state estimation algorithms

Abstract

An increasing calculation speed is one of the highest priorities in developing power system state estimation (SE) algorithms. Modern power systems are continuously growing and different types of new measurements are appearing in networks which impact on a practical implementation of SE. The expanding size and growing complexity of the SE task lead to increasing computation burden. The paper proposes parallel execution approaches of mathematical operations, which used in different stages of SE algorithms. Also, the new matrix storing format suited for parallel manipulations is suggested. The format allows adding new elements to the matrix in any sequence without reallocating memory and performing fast multiplication of two sparse matrices. Their implementation can significantly shorten computation time. Also, it allows to use resources of modern computers in a greater extent. Different approaches for SE are widely used today. Each of them has some features that must be taken into account in practical implementation. The application of parallel calculations to the SE problem in different statements is analyzed in this paper. The elapsed time for parallel executing of mathematical operations in comparison with their sequential realization is provided.