CFD-based optimization of scroll compressor design and uncertainty quantification of the performance under geometrical variations

Positive displacement (PD) machines are widely used in several applications such as in vapor and power generation systems. Nevertheless, their design is still based on standard approaches mainly driven by thermodynamic analysis and theoretical correlations. Geometrical details influence on the machine performance is often neglected. Because of this, PD machines hydraulic efficiency has not really been improved. The present work shows an innovative design strategy aimed at maximizing the machine efficiency by topology optimization. The geometry of a scroll compressor has been parametrized. The parameters were optimized with a Particle Swarm Optimization (PSO) based procedure integrated with Computational Fluid Dynamics (CFD) to achieve the maximum efficiency. In order to better understand the influence of these parameters on machine performances, a Design of Experiment (DOE) approach was also used. Afterward an Uncertainty Quantification framework is applied to the compressor to identify the reliability of the optimal design subject to geometrical variations. Among all the investigated parameters, the most important seem to be the high-pressure port shape and size. Performance are highly affected also by the number of coils which defines the built-it compression ratio.