Trade-off among Mechanical Properties and Energy Consumption in Multi-pass Friction Stir Processing of Al 7075-T651 Alloy Employing Hybrid Approach of Artificial Neural Network and Genetic Algorithm

ABSTRACT: Friction Stir Processing (FSP) is a solid state and thermomechanical processing technique that modify and improve the microstructural and mechanical properties of the material to achieve better performance in less time, using a simple and inexpensive tool and low production cost. Processed zone contains modified mechanical properties, fine grained, equaled and homogeneous microstructures. Discussions about efficient use of energy and expense have become more frequent in many sectors of industry. During manufacturing processes, power consumption of components and potential for savings can be evaluated and measures can be defined for the efficient use of energy. In the present work, FSP was applied on the Aluminum 7075-T651 alloy sheet. Application of multi-objective multivariable genetic optimization in FSP was presented. A trade-off among various mechanical properties of an aerospace alloy and energy consumed during FSP was sought out. At first, the experimental data regarding the elongation, tensile strength, hardness and the consumed electrical energy with respect to various spindle rotational speed and feed rate of FSP were measured. Then an Artificial Neural Network-based approximation approach was used to approximate the value of measured data during the Genetic Optimization. The properties like elongation, tensile strength and hardness were maximized while the cost of consumed electrical energy was minimized. Keywords: Friction Stir Processing; Al7075; Multi-objective; Genetic Algorithm; Optimization; Artificial Neural Networks. …………………………………………………………………………………………………………………………