New Method for Performance Optimization of Finger Seals

Abstract

Improving simultaneously the two desirable objectives — lower leakage and lower abrasion—is very difficult for designing finger seals.Using the existing multi-objective methods to solve this problem,the shortcoming is the necessity of selecting the best design out of many possible ones.Therefore,a new method based on the Nash equilibrium game theory is presented to achieve a better design of finger seal in current study.In this method,there are two key techniques:① the BP neural network technology is used for establishing the mapping relations between optimization variables and the two objectives of leakage and abrasion respectively;② the genetic algorithm is introduced to Nash equilibrium game theory for solving the optimization process.The optimization result is compared with that of the weighted average optimization method with a numerical example.And the numerical results are validated from physical test.The results show preliminarily that:① comparing with the weighted average optimization method,the presented optimization method has only one set of optimized result,which reduces the difficulty of decision-making;② the numerical and test results all indicate that the optimized finger seal through the presented method has lower leakage and lower abrasion than other finger seals.The presented method is very significant to the design of better finger seal.