Improved Pareto Optimal Engine Cycle Designs Through the Use of a New Pareto Quality Indicator
Nicholas Molino, Jonathan S. Sands, Scott Duncan, Eriks Osvalds, Dimitri N. Mavris · 2012
A new quality indicator for Pareto efficient frontiers has been developed in order to map desired Pareto optimal set attributes to a single metric. This new Pareto quality metric was tested to see if it could be effectively utilized along with Response Surface Methodology (RSM) techniques to tune the parameter settings of multiobjective optimization schemes. As a working example, the parameter tuning was applied to a multiobjective genetic algorithm optimizing the engine cycle design of a geared turbofan with N+2 level technology on a 300 passenger commercial aircraft. The Environmental Design Space (EDS) tool, which analyzes aircraft performance, source noise, and exhaust emissions, was used for the cycle design. The goal of the multiobjective optimization was to simultaneously minimize total fuel burn and cumulative noise for a specified mission profile. This research has concluded that tuning the parameter settings of an optimizer can provide significant benefits in generating Pareto optimal solutions, especially when executed on an engine cycle design. The Pareto quality indicator developed has shown promise in assisting the parameter tuning effort, but must be refined through further research and development.