Analysis, Modeling and Simulation of Reliability of Generation of Kainji Hydro Power Station, Nigeria

Abstract

Nigeria, suffers from major problems including a power supply far outstripped by demand, plagued by frequent failures and a confused policy environment. Essential information needed for the determination of robust strategies is lacking and decisions are often based on ad hoc extrapolation of the experiences of other nations without fully reflecting the Nigerian situation. This research work focuses on the hydro generation aspect of the Nigerian power system, and developed the framework and tools that can facilitate serious policy making and management. The basic framework developed here includes the nucleus of a relational database containing available data relevant to power generation in the Nigerian power system. The tools used include standard Structured Query Language (SQL), pivot tables and algorithm style macros developed using EXCEL - VBA®. Analysis of aggregated power generations indicate that almost irrespective of the amount of money spent for maintenance and refurbishing, the system’s average effective generating capacity was about 1900 MW or only about 32% of the installed capacity. Subsequently the actual average effective generating capacity increased to 2700MW. The study proved that the causes of the problems of power generation were complex, and so it was decided to study, model and simulate one of the power stations – the Kainji Hydro power station (KHPS). From the data collected, it was possible to extract information about the performance of each unit over the period from 1992 to 2008 using Microsoft EXCEL-VBA®. The hydrological data indicated that during the studied period, the reservoir emerged from the draught cycle and fully recovered by year 2000. Furthermore it was possible to estimate the aggregate conversion factor and efficiency of the turbo-alternators. The MTBF and MTTR were similarly determined for each machine and a 128-state Markov model for the seven units was constructed. The results indicate that despite their ages, the probability of having four to five operational machines was around 60%. Finally, stochastic simulations were developed for each unit and combined to obtain scenarios for the station; the result provides a useful guide on possible actions that could be taken by the management. Policy-making experimentations could also be readily carried out.