Abstract
In this thesis, a method based on a sensitivity analysis and quadratic curve-fitting
technique for power loss reduction in a low-voltage distribution area is proposed.
Loss sensitivity based method is used to determine some potential DG sites and
least-square based curve-fitting technique is used to find out optimum DG capacity.
Results obtained from the proposed method is justified using the exhaustive search
method which is a computer program that searches for all alternatives by evaluating
each individually. For determining the optimum generation capacity of multiple
DG units, a new methodology based on an unbalanced multi-phase optimal power
flow (UM-OPF) is presented in this thesis. During the formulation of the UM-OPF,
an optimisation algorithm is developed in Matlab and the unbalanced multi-phase
power flow is solved using the Electric Power Research Institute’s (EPRI) software
OpenDSS. To ensure a global minimum loss profile, a swarm intelligence-based adaptive
weight particle swarm optimisation (AW-PSO) algorithm is used which shows
better convergence profile compared with basic PSO algorithm. Validation of the
proposed methodologies is also conducted using an exhaustive search algorithm. The
results obtained from the proposed methods show that significant loss reduction is
possible using multiple optimum sized DG units. To determine the optimum DG
capacity, with varying generation and load, a cost minimisation planning methodology
is also proposed. During the planning process, different DG technologies (such
as solar and wind) are considered. The results obtained from this methodology show
better loss reduction profile compared to the peak load planning.