A hybrid model proposal for maintenance scheduling in hydropower plants

Abstract

Maintenance is an important process that must be managed simultaneously with personnel and materials, production in industrial enterprises. Maintenance planning comes at the beginning of the critical stages of this important process. There are two stages required for maintenance planning. The first stage is the determination of maintenance strategies and the second stage is the generation of maintenance schedules. This study is carried out for maintenance scheduling as a continuation of the maintenance strategy selection which is the first step of maintenance planning. In the first step study realized for maintenance strategy selection, 1330 electrical equipment located in a large-scale hydroelectric power plant in Turkey is examined and the criticality level is determined for the plant. As a result of the study can be realized in accordance with a timeline, 7 critical electrical equipment groups in which preventive maintenance strategy can be applied. As a result of the study can be realized in accordance with a timeline, critical electrical 7 equipment groups determined which preventive maintenance strategy can be applied. These results are used in this study which constitutes the maintenance scheduling stage in maintenance planning and the maintenance schedule has been established for these critical electrical 7 equipment groups to which preventive maintenance strategy can be applied. In the first stage of this study for maintenance scheduling, one-year generation estimation of the plant is realized by using the Artificial Neural Network (ANN) method and operating-maintenance periods are calculated from the data obtained from this estimation. By using consistent estimation values in the 0-1 Integer Programming (IP) model, which reflects the power plant's maintenance planning realities and requirements, five different types of periodic maintenance are scheduled and an optimal maintenance plan consistent with the power plant operation and maintenance facts is achieved.