A thermal generation unit maintenance scheduling model with probability space reduction

The paper proves that unit preventive maintenance models with maintenance costs alone may only yield feasible solutions in the "window" approach because the approach treats maintenance costs as constant. Consequently, the objective function is only a constant that is a dummy in the problem formulation. It is concluded that operational costs must be included for optimality. This paper therefore proposes a maintenance scheduling model that does away with constant maintenance costs. A probability space reduction technique is also proposed to reduce computation time. A case study shows the successful application of the model.