Stirling engine multi-objective optimization using a genetic algorithm

Abstract

With the growing demand of energy globally, the actual worrying state of the earth’s finite resources, namely fossil fuels, opens up the scope of energy researches to innovative and efficient solutions. Stirling engine has been an interesting subject of study since its invention, and many studies dealt with Stirling engine efficiency with attempts to optimize it in order to have a proper use of the engine in the real world, depending on the use cases. Stirling engine is an external combustion engine with a theoretical efficiency equivalent to that of Carnot. Alongside the global awareness to use efficient and less resource consuming solutions, there has been a spiking growth in the set of tools that are conceived to achieve that; specifically in the machine learning area. Among the various available algorithms, the one used in the hereby study is the non-sorted genetic algorithm II, which falls into the genetic algorithms category. This algorithm is well suited for multi-objective optimization problems; it consists of selecting the best design parameters that are contained in predefined upper and lower bounds, based on multiple objective functions.