Using Genetic Algorithms for the Optimal Design of Fluid Journal Bearing

Abstract

Abstract The current methods of designing journal bearings are achieved by trial and error and mathematical programming. A new method is used here to design the journal bearings, i.e., Genetic Algorithms (GA). GA is a very efficient method, providing relatively an easy way to find the optimum solution. For that purpose we need to construct the proper objective function for the design of the journal bearing. Then the rest exhaustive and repeated task could be left to the computer. GA can solve the linear or nonlinear, single parameter or multiparameter system. Genetic Algorithms are very different from the traditional optimization techniques. It is a new generation of artificial intelligence and its principles mimic the behavior of the biologic genes in the natural world. It will search the best set of parameters of the objective function, reproduce them, eliminate unsuitable sets of parameters, cross-couple the survival sets of parameters, execute mutations and then return to plug the values of the new variables into objective function to repeat these procedures. Its execution is simple and could reach the solution in a very short time. According to these characteristics, it is a very powerful method for finding the minimum dissipation of energy in the journal bearing. This approach opens new paradigm for optimization. The GA’s procedure is robust and works even if the functions involve are not continuous and have no derivatives. The resulting set of optimal parameters was compared with the results of a recent paper addressing the same problem.