Knowledge-Based Multi-Diciplinary Optimization of Rocket Turbopump Design

The design of turbomachinery components is a complex process requiring many steps and hundreds of minute decisions in an iterative process. The design proceeds from meanline design through 3D viscous flow calculations (CFD) and finite element structural analysis (FEA). Design studies may traverse this design cycle many times, with routine tasks performed by engineers in a trial-and-error manner. The current turbomachinery design system integrates the initial meanline design considerations, the details of sophisticated blading design, the essential characteristics of modern CFD analysis, and structural and thermo-mechanical FEA. Some knowledge based engineering (KBE) aspects are applied to guide the initial design setup in directions of proven design techniques, and historical design and performance data is used to validate and calibrate the design tools. In addition, the iSIGHT optimization code is currently used to refine the specific vane shapes. In the proposed work, KBE and multi-disciplinary optimization (MDO) will be employed to improve the turbomachinery design process by capturing critical design knowledge, automating routine design tasks, and optimizing the balance between conflicting design goals and criteria. The open and extensible design system will use KBE to guide the engineer in performing design modifications, exploring trade studies, applying design criteria, and performing MDO.