Design of air cushion vehicles using artificial intelligence: Expert system and genetic algorithm

In the thesis, initial design of an Air Cushion Vehicle (ACV) is performed with the expert system and its skirt system is further optimized with the genetic algorithm. Both the expert system and genetic algorithm are advanced computerized design techniques of artificial intelligence. Those techniues are specifically developed for the ACVs with programming codes in this thesis. Then the main objective is to show the successful implementation of those techniques in the design of ACVs.;The thesis work is divided into two parts. In the first part, the general configuration of ACVs, including the overall dimensions, weight distribution, parametric properties, and several subsystems, is studied and designed by the expert system as an initial design phase. In the second part of the thesis, the skirt system of ACVs is further optimized. In particular, the properties of the bag and finger skirt are optimized for improved ride quality and stability by the genetic algorithm. For the validation of these two artificial intelligence techniques, the CCG (Canadian Coast Guard) 37 ton Waban-Aki and U.S. Navy's 150 ton LCAC (Landing Craft Air Cushion) are selected for the tests. The results of the tests proved that the expert system was successfully implemented and was a powerful tool for the initial design of ACVs. Furthermore, the genetic algorithm optimized the skirt system with significantly improved ride quality and stability. It was also shown that the skirt mass was an important design factor in the heave response of the bag and finger skirt. Hence, this thesis work opened the new possibility of designing ACVs with artificial intelligence techniques.