Maintenance scheduling for different types of equipment for the first time in the literature Integrated maintenance scheduling and generation estimation Maintenance scheduling for five different types of preventive maintenance for the first time in the literature Figure A. Application steps Purpose:In this study, by using this data a 1-year maintenance schedule is obtained for the critical equipment groups that should be implemented the periodic maintenance strategies with 5 different contents and frequency via the proposed hybrid methodology. In the first step of this study, which uses the combination of Artificial Neural Network (ANN) and Integer Programming (IP), 1-year generation estimation of the power plant is realized by ANN method and operation and maintenance hours of the power plant are calculated from this forecasted data. These calculated periods are included in the proposed maintenance scheduling model which reflects all the requirements and realities of the hydroelectric power plant where the application is carried out and a feasible maintenance schedule is obtained. This study is the first in the literature in terms of handling multiple equipment groups in power plants, integrating the annual operating hours of the power plant into the maintenance scheduling problem in a consistent manner with real life, and providing an effective and applicable annual maintenance schedule that is compatible with the operation of the power plant.
Theory and Methods:Artificial Neural Network and Integer Programming methods were used in this study.
Results:As a result of the study, five different preventive maintenance of the critical equipment of the plant have been scheduled. The contributions of this study to the literature can be summarized in below: While there are often studies creating the maintenance schedules for a single equipment in the literature, in this study, more than one electrical equipment group is evaluated in a power plant and a mathematical model is presented considering the specific maintenance characteristics of each equipment group. In this study, considering the operating conditions of the power plant an applicable maintenance schedule is obtained by taking into account the 5 different periods as weekly, monthly, quarterly, 6 months and yearly. 1-year generation forecast for the power plant by using the 9-years real data of 2010-2018 in proposed ANN model, and calculating the operating and maintenance capacities (hours) for the planning period in the maintenance scheduling problem are two stages of a unique application in the literature. In addition, the integration of generation and maintenance capacities of the power plant obtained by proposed ANN model into the proposed maintenance scheduling model which is generated by using IP has enabled the combination of ANN and IP methods for the first time in the literature. It is designed as a real-life application of working with real data taken from a hydroelectric power plant in Turkey stands out as a feature which is not considered in mo...