Surrogate-based optimization of multiple-splitters radial compressor for solar hybrid microturbine

Abstract

In the open literature, there is no clear guidance for design procedure on the number or where best position the splitter leading-edge to improved performance in the radial compressor. A surrogate-based optimization of multiple-splitters radial compressor has been conducted for solar hybrid microturbine in the present studies. Also, the impact of the multiple-splitters leading-edge position on the rotor radial compressor is demonstrated. The radial compressor is compared using air and supercritical carbondioxide ( sCO 2 ) as the working fluid. The optimization problem is solved by applying the surrogate or response surface model using the radial basis function network (RBFN) from the opensource package in the Julia programming language environment. Starting from 65 data points from DoE based on the optimum Latin-hypercube, the optimum candidate design can be predicted with CFD and FE analysis. The results show that for air radial compressors, the optimum position of the small-splitter leading edge is 20% from the outlet rotor, and for the middle-splitter leading position at around 30% from the inlet rotor. Unlike air radial compressors, the sCO 2 radial compressor has the optimum leading-edge position for the middle splitter around 60% from the inlet rotor, and the small splitters are around 30% from the outlet rotor. Finally, compared between the non-splitter and multiple-splitter designs, the total-to-total efficiency at the same pressure ratio is improved by around 2.30% for air radial compressor and 2.25% for sCO 2 radial compressor. Meanwhile, the operating range is increased by 29.0% for air radial compressors and 18.0% for sCO 2 radial compressors.