The use of supersaturated experiments in turbine engine development

When designing a complex mechanical system there is always considerable emphasis on reducing the time spent determining the best design. Understanding factors that contribute to the performance of the system as rapidly as possible is very important. Supersaturated statistical designs (SSDs) offer a potentially useful way to investigate many factors with very few experiments. SSDs investigate m factors with n experiments, where m > n - 1. In this paper, we demonstrate the use of a SSD in the detailed design phase of a turbine engine. Three SSDs and three orthogonal two-level factorial experiments were used as screening experiments. The results show that SSDs are a reasonable analysis choice, since few active factors are missed.