Bayesian modeling of a wind turbine system in the preliminary design phase

Abstract

Of our days, the question of industrial competitiveness is of particular importance for the industries in all sectors of activity. The next years will be marked by persistent pressures on the cost and the quality of the systems. Of this fact, an analysis of the safety of operation constitutes a necessary approach for the improvement of the operation of the systems by acting as soon as the primary phases of the design on the settings and the factors influencing the functioning of these systems. In effect, the modeling is a critical phase in this kind of analysis and the literature presents several tools enabling this modeling. In operational safety, there is often little information available. It is therefore generally impossible to obtain the parameters describing the Act attached. That is why; modeling tools rely primarily on the judgment of experts, surveys and databases of the manufacturers to build models structured. In this perspective, the Bayesian networks are a powerful solution of modeling for complex systems to merge the knowledge of various natures within a same model. In this logic, we will present a modeling by Bayesian networks of the two concepts of turbine namely direct and indirect in the aim of constructing a decision of choice between the two concepts in the preliminary phase of design of wind turbines while presenting the method, the calculation of the probability of state and the validation criteria of choice.