Planning model for wind turbine maintenance shutdowns and wind forecasting
Wind energy; planned maintenance; wind speed forecasting; digital twin.
Wind farms must regularly interrupt your energy production to perform the programmed maintenance. Since this pause causes losses in generation and any shortfall must be compensated with energy purchases on market, it is essential to determine the optimal time to apply the maintenance for wind turbines. Predicting wind speed through mathematical models has made it possible to define real-time production operation and equipment control. Therefore, this research has the objective in propose a model for planning wind turbine shutdowns for maintenance with use of wind forecasting and digital twin. This work was performed in three steps. The first step involved literature research about asset maintenance, wind forecasting and digital twin. The second step included the systematic literature review about the prediction models used the recent times with applications to maintenance and digital twin. The third step covered conceptual modeling for wind turbine maintenance planning and application on a wind farm. The proposed model corresponds to a combination of predictable multi-turbines spatiotemporal correlations framework method with an optimal maintenance schedule model. It consists of a probabilistic model, using predicted wind speed data and electricity spot market prices, also estimated in a stochastic program called Newave. Wind speed data is used on the multi-turbine spatiotemporal correlation algorithm; simultaneously two autoregressive moving average methods are applied for a long-term forecast of one week, while the other one is done in months. During the forecasting process, the input data is adjusted continuously for any variation. This correction is possible through the virtual simulation of the turbines using the digital twin, making the model a monitoring system with real-time feedback. The maintenance program for applying the model is centered on planned corrective, preventive and predictive methods, a decision-making is applied to choose the optimal time to interrupt production and carry out maintenance.