Wind Turbine blade optimization using genetic algorithm

Electrical energy and its resources are one of the most important topics of today's world. There are two types of energy resources, renewable and non-renewable. The world population is increasing and also technology is developing faster than before. This increases daily electrical energy needs of the world. In another words, we need efficient renewable energy systems, so that in the future we can live in a green world without fear of depletion of non-renewable energy sources. Using renewable energy is not the solution alone, but also the systems must be efficient. In this study, conventional wind turbine blade design with Blade Element Momentum (BEM) Theory and its optimization by using genetic algorithm optimization method is investigated. This study focuses on the effects of airfoil profile distribution on efficiency for the direct wind taking horizontal axis wind turbines. Blade Element Momentum Theory is combined with the genetic algorithm optimization method and tool developed for blade design and its optimization. The results, which are obtained using genetic algorithm optimization, are better than conventional designs. The best results blade profile distribution is exported to Solidwork program as points by using Solidworks Macro. After that, the points are converted 3D blade geometry. This 3D wind turbine blade geomety and its computational fluid domain, which has only the flow volume of the wind, are created in Solidworks. The fluid domain then is imported in ANSYS software and computational fluid dynamics calculations are made using FLUENT. The program results are compared with the results obtained from genetically optimized blade element momentum theory results. Blade element momentum theory is widely used for conventional wind turbine blade design. In blade element momentum theory turbine blades are divided into sections along its span. Each section is calculated individually and it is assumed that there is effect on each other. After completing sectional calculations calculated values are added up. At the begining, BEM is explained and calculations have been done for a wind turbine blade, which has only one airfoil from hub to tip. After that conventional designs are tested, in this part, first the blade is divided into three sections (root, mid and tip) and for each section profiles are defined unique for its section. The results with single profile geometry and divided blade geometry are compared with each other. The best blade geomerty of these two designs is used as a base of optimization calculations. The airfoil data (such as lift and drag coefficient, aerodynmaical informations etc.), which are used through this study, are collected from online airfoil database. The data of the best blade is exported for the optimization process. For investigating effects of foil selection in blade sections genetic algorithm optimization method is used. Genetic optimization method is a multivariable optimization algorithm which is widely used in such cases. Genetic algorithm is simply altering the airfoils attended on sections and tries to reach best efficient blade geometry. Simply genetic algorithm changes the profile distribution on each section and optimizes the wind turbine blade performance. Due to number of sections, which are used for blade element momentum calculations, and amount of airfoil profiles, which are defined in our program, makes it difficult to calculate and choose the optimum blade geometry iteratively. In order to avoid loosing plenty of time and avoiding unnecessary computational power consumption optimization tool is required. Nowadays, the cost of computational calculation is reducing and started to be used widely among such problems. They do not only require less time but also give accurate results, if the problem is defined correctly. The reduced cost and time encourage engineers to use optimization tools more often. After importing the data of the best design, it is optimized using genetic algorithm optimization tool. This tool only alters the blade profile distribution among the blade and optimizes it. The best geometry distiribution of each optimization step will be taken into account in the next steps. In this study, we investigate the effects of airfoil distribution on blade design. As a result, single profiled blade design has worst efficiency and conventional design is better than single profiled blade design, but optimized blade has the best efficiency as expected. The results of these three blade distribution solutions are given in the results section. After completing blade design using genetic algorithm profile distribution among the span is exported as cartesian coordinates. These coordinates will be used as an input for the 3D design in Solidworks. At the begining, the section coordinates are imported using Solidworks macro and from imported points B-Splines are created. The splines are converted into surfaces, by using loft command. These surfaces then converted into 3D solid geometry. In order to analyse the blade in computational fluid dynamics program, control volume is also created in Solidworks and the geometry is exported as parasolid data, so that computational fluid dynamics program can read it. The generated 3D flow domain is imported to ANSYS. After that the geometry transfered into FLUENT and all boundary conditions are defined and calculations are made. The calculations are only done for the design point of the blade and no further analyses have been done. The results are investigated in order to see force distribution among the blade and total power generated at design point is calculated. The results, which are obtained from FLUENT, are similar to the results, which are obtained from the developed tool. The optimization structure simply concentrates on the airfoil distribution along the blade and search possibilities via genetical algorithm optimization method. As a result, the developed algorithm for optimizing wind turbine blade gives us the results which are expected at the begining of the thesis. This design mechanism might be used in the future for calculating similar turbomachineries.