A continuous compact model for cascaded hydro-power generation and preventive maintenance scheduling

• An integrated model for hydro-power generation and maintenance scheduling.
• All variables are continuous, allowing maintenance start during a time period.
• New procedure to evaluate the maintenance effect at hydro-power plants.
• The model is compact, only storage variables are explicit.
• Suitable for evolutionary algorithms, e.g. differential evolution.

This paper proposes a new nonlinear model for cascaded hydro-power generation and preventive maintenance scheduling. The reservoir level and the maintenance scheduling are optimized, simultaneously, to minimize the thermal generation complement and to maximize the future water value. To achieve better accuracy, the productivity was modeled considering the head effect. Since the model considers individualized plants, the water flow balance must be taken into account. By means of a transformation, this constraint can be modeled only with the reservoir volume. Also, by means of a simple nonlinear transformation, the maintenance scheduling was modeled as continuous variable, instead of the usual binary models. The maintenance start date is a continuous variable, so, it allows the maintenance to begin within the time period. These simplifications altogether define a compact continuous nonconvex problem. The proposed model was successfully solved by a Differential Evolution (DE) algorithm in Brazilian cascades.