Application of the Semi-Markov Processes to Model the Enercon E82-2 Preventive Wind Turbine Maintenance System

Abstract

The share of wind energy in the energy mix is continuously increasing. However, a very important issue associated with its generation is the high failure rate of wind turbines. This situation particularly concerns large wind turbines, which are expensive and have a lower tolerance for system damage caused by various failures and faults. Vulnerable components include sensors, electronic control units, electrical systems, hydraulic systems, generators, gearboxes, rotor blades, and so on. As a result, significant emphasis is placed on improving the reliability, availability, and productivity of wind turbines. It is extremely important to detect and identify abnormalities as early as possible and predict potential failures and damages and the remaining useful life of components. One way to ensure turbine efficiency is to plan and implement preventive repairs. This work shows a semi-Markov model of a preventive maintenance system based on Enercon E82-2 wind turbines. The system’s performance quality is evaluated based on profit over time and an asymptotic availability coefficient. The developed model establishes formulas describing the efficiency functions and formulates the conditions for the existence of extremes (maxima) of these functions. Computational examples provided at the end of the paper illustrate the obtained research results. A preventive maintenance model is developed that can be applied to wind turbine hazard prevention (determining optimal times for wind turbine preventive maintenance).

The Use of Markov Chain Model for Rating Location Based Turbine Performance

“…A Markov-based Back Propagation (BP) neural network optimized by the Particle Swarm Optimization (PSO) used for wind power prediction showed a nonsignificant improvement over the traditional Markov chain [31]. A semi-Markov model was also used to evaluate the performance of wind turbine systems [26]. In order to characterize wind power more accurately forecast [2,10,15] and simulation [18,24] are two methods mainly used.…”