MODELING HYDROTHERMAL POWER GENERATION SCHEDULING USING MIXED INTEGER LINEAR PROGRAMMING

Abstract

Power production and distribution in Ghana is ever more becoming erratic and expensive, both for the power producer and the consumer. It is in this regard that an investigation of hydrothermal power generation scheduling is undertaken for Volta River Authority (VRA) of Ghana. The goal of the study was to determine an optimal power production schedule that meets daily load demands at minimum cost of production and also ascertain the marginal cost of producing electricity per day and therefore the tariff rate. The problem was formulated as Mixed Integer Linear Programming (MILP) and the resulting model tested using real data obtained from VRA. The integer restrictions on the problem were relaxed initially and an optimal solution of the resulting LP problem obtained using the Simplex Algorithm. The Branch-and-Bound method due to Land and Doig which is implemented in LP Solve IDE VS.S.2.0 by Henri Gourvest was used to obtain the required integer solution. The test results which included post optimality analysis showed that daily load demands could be met at a minimum cost. Furthermore, the marginal cost of producing power obtained from the dual of the MILP model provided insight into the appropriate Tariff that is reasonable for VRA to charge consumers.