Concurrent Design and Optimization

Vehicle design often becomes a problem of maximizing one (or more) problem-specific measures-of-merit (e.g. propulsive efficiency).  But this seemingly simple objective becomes more complex if the measure-of-merit is affected by more than one design discipline.  Fuel consumption, for example, will not be optimized either by minimizing vehicle drag or maximizing propeller efficiency.  Both disciplines have to be addressed simultaneously if a truly optimal vehicle is to result.  Sailing yacht design is even more complex.  Drag minimization must be weighed against lift generation, and aero force optimization must be weighed against hydro force optimization. 

Proper design therefore requires the accurate quantification of complex multi-disciplinary measures-of-merit.  AFT’s staff has decades of experience devising such design-critical performance predictions from RANS simulations.   In the ships’ propulsive efficiency example, we developed specialized techniques to couple hull and propeller RANS calculations so that optimization could proceed concurrently.  In the sailing yacht example we developed a Velocity Prediction Program (VPP) designed to use only RANS aero and hydro data, and then optimize over any range of input design parameters.  Our 2007 Chesapeake Sailing Yacht Symposium paper “Performance Prediction without Empiricism:  A RANS-Based VPP and Design Optimization Capability” documents this unique capability. 

But concurrent design (let only optimization) requires huge numbers of single-point design performance analyses.  If RANS cannot be made to complete hundreds (or even thousands) of simulations within a reasonable time frame, it will stand little chance contributing to the design process.   This is exactly AFT’s niche, and in fact the very reason we were founded in the first place.  We are uniquely positioned with all the software, hardware, and expertise to complete hundreds of large-scale, three-dimensional RANS simulations in any given month.