OPTIMAL DESIGN OF EOLIC TURBINE WITH VERTICAL AXIS COMBINING NEURAL NETWORK AND OPENFOAM

J.L. MROGINSKI; H.G. CASTRO; J.M. PODESTÁ; R.R. PAZ

San Miguel de Tucumán

Conferencia; XII Congreso de Mecánica Computacional (MECOM 2018); 2018

AMCA

In the present work a new concept in the use of wind energy is proposed through a combination between vertical axis turbines Savonius and Darrieus whose most widespread topology is the so-called "H turbine". Combining computational fluid dynamics and neural networks, we propose the design of an intelligent turbine that combines the Savonius and Darrieus turbines according to the speed and direction of the incident wind, as well as, modify the length and/or the angle of attack of the blades in order to obtain maximum performance for different speeds and angles of wind incidence. The adopted turbine model geometry is 3 vertical blade of NACA 0018 profiles with 2 m length covering an average radius of R = 1 m. The blades have a device that allows their opening modifying its operating principle, being able to go from a Savonius turbine to a Darrieus turbine, allowing the start of the system with relatively low wind speeds. Adopting as optimization variables the angle of attack and the inclination of the blades, as well as the average radius of the turbine, an optimum performance of the turbogenerator is obtained both in the maximum power reached and in the regularity of the equipment.