OPTIMUM DESIGN OF MICROWAVE TOWER USING GENETIC ALGORITHMS

Abstract

With the advances in technology and expansion in telecommunication, microwave towers are being setup at a rapid speed in all developed and developing countries. The advancement in the design and construction techniques has resulted in a new generation of structures which are tall, slender, flexible, low in damping, and light in weight. Such structures show an increased susceptibility to wind load. Reducing the weight of the structure means reducing the individual member weight and its size, which in turn reduces the wind load on the structure due to their reduced projected area. Hence, it has become necessary to develop a tool to enable the designer to select an optim urn section from the available practical sections. In this thesis, a Genetic Algorithm (GA) based approach is proposed and implemented for this purpose. Stability and strength related constraints are imposed by considering the manufacturing and erection aspects. All critical design variables such as area, base width, panel height, etc. are considered for optimization. By different combinations of design variables, the GA program gives the minimum weight of tower. A simple graphical user interface has also been developed to facilitate the optimization process with any combination of input parameters like the size of population, maximum number of generations, basic wind speed, crossover and mutation probabilities, etc. and get the optimization result. For obtaining more general solution, configuration (bracing types, number of inclined panels), size (member sizes) and topology optimization (base width of tower), are attempted simultaneously to get optimum weight of tower. Mixed type of design variables (both continuous and discrete) are accounted for in this thesis. Experiments using the implemented program show that the proposed approach is highly effective in estimating the optimum weight of the tower of a given height for a given wind zone. Rigorous experimentation has been performed by varying both the physical and genetic parameters so as to come up with an optimal set of parameters for this problem. Comparison of obtained results (minimum weight of tower) with the weight of traditionally designed towers illustrates the usefulness and contribution of the work. The studies carried out in this thesis, and the trends and databases obtained can be of great help for the designer for estimating the weight of tower.