Genetic algorithm-based search heuristic for turbine balancing problem

Turbine balancing problem (TBP) can be defined as arranging blades of each stage of turbine engines in such a way that provides a minimum residual unbalance to eliminate the undesired vibration. This problem is regarded as an NP-hard problem and no exact solution has been found for such problems up to now. In each stage of the turbine, an optimal arrangement of blades around a disk by considering the difference in weight of blades to achieve a minimum residual unbalance is essential for preventing the early failure of turbine engines. In this article, a genetic algorithm (GA) has been proposed and applied to TBP for defining an optimal blade arrangement of a turbine stage. This algorithm proposes a suitable choice for operators of GA, based on permutation representation. By eliminating the shortcomings of previous researchers’ attempts with a simple form of GA, our proposed GA is capable of attaining an optimal arrangement of blades with reasonable computational efforts. Moreover, suitable values for parameters of GA including population size, crossover, and mutation rates of the proposed GA have been determined using practical sets of blades belonging to a Siemens gas turbine engine. The proposed GA with tuned parameters has been successfully employed for a variety of TBPs, and obtained results show a good convergence to an optimal value of residual unbalance with reasonable computational efforts.