Abstract The performance of a cyclone separator is often measured by grade-efficiency, which captures the efficiency of separating particles of a given feed size from the dusty gas stream. Accurate prediction of grade-efficiency has huge implications on design, safety, energy consumption, and environment. Based on comprehensive experimentation, this study exploits the best possible model to predict grade-efficiency and to conduct robust parameter design for minimizing the process variations. The grade-efficiency model via generalized linear mixed effects modeling (GLME) is identified to characterize grade-efficiency under uncertainties. The GLME modeling achieves 71.7% reduction compared to the ordinary linear model and 63% reduction compared to the beta regression model in prediction variability and therefore provides insights into the engineering design of the cyclone separator system through robust parameter design. Graphical abstract Based on comprehensive experimentation, the best possible model to predict grade-efficiency and to conduct robust parameter design for minimizing the process variations is proposed. The grade-efficiency model via generalized linear mixed effects modeling (GLME) is identified to characterize grade-efficiency under uncertainties. The GLME modeling achieves dramatic reduction in prediction variability and therefore provides insights into the engineering design of the cyclone separator system through robust parameter design. Unlabelled Image Highlights • The classical grade-efficiency model does not accommodate industrial uncertainties. • A new grade-efficiency model is built by generalized linear mixed effects modeling. • GLME model enables over 70% reduction in prediction error. • GLME model provides insights into robust engineering design of a cyclone separator. [ABSTRACT FROM AUTHOR]