Modern energy harvesting systems push the boundaries of conventional design approaches. Numerical optimization algorithms coupled with suitable modeling tools can exploit system trade-offs and explore novel design spaces.
In particular, high-fidelity multidisciplinary design optimization (MDO) captures complex multiphysics interactions with unprecedented detail, at the cost of high computational expense and implementation effort.

This presentation showcases numerically tractable, highly scalable high-fidelity MDO studies enabled by state-of-the-art open-source frameworks that leverage efficient implementations of the adjoint method.
First, computational fluid dynamics (CFD) and structural mechanics (CSM) solvers are coupled to enable gradient-based optimization of a large wind turbine rotor. Hundreds of structural and geometrical variables are optimized simultaneously to improve aerodynamic efficiency and structural weight.

Second, an OpenFOAM-based optimization framework is used to optimize the shape of a ducted hydrokinetic turbine and maximize its power output. The design strategy was validated through an experimental campaign that measured record-breaking system efficiency.