Session: 01-08 Inlets, Nozzles, Mixers and Nacelles II

Paper Number: 129156

129156 - Three-Dimensional Fluid Topology Optimization for Intakes and Exhausts Flow Uniformity 

In this work, Fluid Topology Optimization (TO) is used to optimize flow uniformity for intakes and exhaust of industrial gas turbines. In compressor intakes the solver minimizes the inflow distortion to the intake, while preserving mass flow, minimizing losses and adopting a design that can be manufactured with standard processes. For turbine exhaust a simplified geometry is analysed, considering the exit swirl component. The core of the methodology is an in-house Fluid Topology Optimization solver, TOffee, extended in this work to minimize the deviation between the velocity outlet profile and a specified velocity target. The TO method relies on emulating the natural sedimentation process inside a design domain. The fluid dynamic system is numerically solved, and a scalar design variable is iteratively updated following a Lagrangian optimization approach which ensures meeting the constraints as well as the minimization of the objective functions. In this regard, the optimization is multi-objective. Pressure losses and flow profile deviation are simultaneously minimized. At the end of the process, the design solution is the set of regions with high impermeability (solid structures). The optimizer generates flow splitters that channel the flow to match the velocity target while maintaining low pressure losses. The efficiency for the intake of the compressor design is measured as relative standard deviation, which is recorded as higher than 70 percent for non-optimized test cases and lower than 14 percent for the optimized solutions. Different levels of complexity and deviation coefficient are shown accordingly with the relative weights imposed on the multi-objective optimization. The solver is able to generate several de-swirler designs to match the design conditions. Even though related theoretical modelling and similar approaches have been briefly described in the literature, this is the first time that the algorithm has been successfully applied to a realistic geometry with correct Reynolds and in its operating conditions. 

Presenting Author: Stefano Furino ToffeeAM ltd

Presenting Author Biography: Stefano obtained a Master of Science in Aerospace Engineering from “Politecnico di Torino”, and a Bachelor in Aerospace Engineering from “Universita’ di Napoli Federico II”. Before joining TOffeeAM three years ago, Stefano carried out an industrial placement at GE Avio in Turin.

Authors:

Stefano Furino ToffeeAM ltd
Alessandro Canova ToffeeAM ltd
Marco Pietropaoli ToffeeAM ltd