Mach Number Prediction and Analysis of Multi-Mode Wind Tunnel System

Abstract

The Mach number is one of the most important variables in the wind tunnel system, which needs to be controlled precisely. Although there are mechanical principals to present the relationship between the Mach number and other process variables, it is hard to obtain the Mach number online directly, especially when the wind tunnel system operations in multiple modes. Therefore, in this paper, a Mach number prediction algorithm based on PLS is proposed for the multi-mode wind tunnel system, as well as the analysis strategy. First, on the premise of considering the control accuracy of the Mach number as well as the process characteristics, certain modes are identified and modelled for the Mach number prediction. Second, to reflect the real-time changes of the Mach number, time-slice-based PLS regression method is adopted. And then to further enhance the prediction accuracy, the analysis of process variables influence variance is carried out, and the variables with stable regression parameter are selected in the predictive model to reconduct the prediction. The method has been adopted to the real wind tunnel system, and the outcome demonstrates the effectiveness of the method.