Study on disk and blade design based on multi-layer optimization strategy

Multi-layer optimization strategy plays a key role on solving multi-discipline optimization design (MDO) on the turbine. The main principle of MDO firstly extends to component autonomy from disciplinary autonomy. Based on component and discipline autonomy, this paper designs four kinds of MDO framework which are applied on the turbine design combining with present MDO strategies including BLISS2000, CO and ATC. Both the two-layer including double-subsystem and two-layer including triple-subsystem are built by BLISS2000 strategy as the two-layer framework for turbine design, while the three-layer including triple-subsystem strategy takes full advantage of BLISS2000 and CO strategies. Additionally, due to introduction of auxiliary variables and associated variables, the ATC optimization strategy requires huge extra optimization cost, so it is not suitable for triple-layer MDO strategy on turbine. Finally, the integrated MDO of disk and blade tests the former three kinds of MDO strategies. The preponderance of the MDO strategies including three subsystems is higher efficiency, while complex interactions between layers always lead to worse optimization accuracy and efficiency on three layers strategy comparing with two layers strategy.