In conventional power flow analysis, simultaneous linear equations numbering in thousands must be solved repeatedly and in many instances on-line. Luckily the matrix of coefficients of these equations is sparse with a few elements in each row. Therefore for a faster solution and to save computer memory, sparsity must be exploited and an optimal ordering of the equations must be sought. When considering harmonics, the number of the simultaneous equations to be solved is multiplied and therefore, considering sparsity and optimal ordering is doubly important.

In this thesis a software package for optimizing computation time and memory requirements through exploitation of sparse matrices for Power System Bus Equations was developed. Exploitation of sparsity was accomplished by utilizing a packed storage scheme and special programming.

HARMFLO package was obtained from E.P.R.I. and was used to compare the results obtained using the developed package with those obtained from HARMFLO. The equations that must be solved for a harmonic load flow analysis are detailed and an example is given. A subroutine to modify the admittance matrix in accordance with the order of harmonics was developed.