Output Maximization of Afam Power Station using Artificial Neural Network Technique

Abstract

The efficiency of an electric power plant is dependent on the ability of the plant to function optimally. One of the ways of ensuring optimal operational output is by the maximization of the output of the plant. Afam power plant of the Nigerian Power System was used as a case study using ANN programming and the result was validated using non-linear computational approach. The plant has total installed capacity of 776MW out of the Nigerian installed power capacity of 11,200MW. Operational output data obtained from Afam power plant from 2010 to 1st quarter of 2020 were modelled in ANN environment and simulated using MATLAB. The results showed worst output of the plant in the years 2011, 2014 and 2016. The causes of the low generation output were traced to maintenance downtime and unavailability of transmission network to evaluate electricity from the generating plant. However, the study showed improved generation output from 2017 to the 1st quarter of 2020 due to improved availability of transmission lines, increased plant uptime and quality gas fuel availability from Okoloma Gas Plant. The results were validated with non-linear computational method. The results of the non-linear computational method are in conformity with those of the ANN. The advantages of the ANN over other techniques (Fuzzy Logic, Evolutionary Programming, Ant Colony, Particle Swarm etc) are quick convergence, reliability of results, usefulness in identifying and controlling non-linear dynamic systems, very fast computation, robustness to parameter change etc.