Stochastic Renewal Process Model for Maintenance in Sudan

Abstract

The electric power is considered as one of the easiest easy power resource in transmission and use and it represents the backbone to the growth and economic development although the electric power has important role in growth and position of human resource in the economic cycle. Therefore, the stationary of power should be secured regularly with high reliability degree .The machine maintenance system consider as the security gate to maintain electricity generation stations. It is as honest guard to the machine to not gat fault accidentally the machine secure stationary plays a big role to reduce operation and production cost which result positively into economic activities. This research comes as scientific and practical addition through stochastic model building for Renewal Process and the machine lifetime model which assist in precision for the future predication of the fault and setting plans to reduce it . This research aims to construct renewal process model and life time model for electricity generation machine in Sudan and to identify the probability distribution that used in lifetime test, the forecasting for the renewal period, construct data base according to the faults on yearly and monthly basis and the loss power due to the fault which result into electricity power stationary. Some of the hypothesis in the research is that: time of spare part renewal follows Weibull distribution, the generation of electricity and the process of replacing parts follow the Poisson regenerative process, there is a relationship between the time of renewal and times of renewal distribution, applications of lifetime model on machines have a positive impact on the electricity stability, the electricity generating machines have a high reliability. The data of the technical faults which belong to (stopping time, return time, failure time, time between failure and another one, loss power during the period (2011-2015) for five machine in Bahri Thermal Station The research presented the theoretical principles for renewal process model and lifetime model in chapter two and three. To apply this model on the data machines faults .The statistical package has been used which is STATGRAPHIC 17.108 for data analysis and models constricting and that is all in applied aspect for the research in chapter four. The most important findings are: the time faults for the five machines follow Weibull distribution with 2-parametrs, there is no trend exist for the time of the machine faults and the renewal process represent homogeneous Poisson process (HPP),renewal time (repair) represent Poisson process renewal process, renewal rate (repair rate) is constant for all machines, there is relationship between renewals rate and the mean time between failure(MTBF), whenever renewal(repair) rate increases the mean time between failure(MTBF) increases too. The renewal time increases in linearity way in other words whenever operation time increases the renewals increases too. Operation time of the machine increasing reduced its performance , the machines no (3,4,6) have high reliability and machines (1,5) have low reliability, hazard rate for the machines increases according to increasing of operations time. The machine that have high reliability the probability of its faults is weak and its hazard rate is weak too and vice versa. According to these findings the study recommended the following: the National electricity authority is to use the renewal process model and lifetime to predication the fault for the machines, expand the study to include hydro and gas generating machines, prepare a form to record the fault data accurately including (type of spare part, cost , renewal time) depending on stochastic model construction to forecast the faults and evaluate the quality of the machine and keeping the existing machines by conducting accidentally maintenance to insure the electricity power stability, The National electricity authority should provide Weibull++ software because it is the best in fault data analysis stochastic model construction.