Scramjet Isolator Shock-Train Leading-Edge Location Modeling

Abstract

A dynamic model of the relationship between the back pressure and the shock-train leading-edge location is useful in the development of a shock-train leading-edge automatic control algorithm. In this research, several models were created relating the back pressure within an isolator to a shock-train leading-edge location. An isolator analog was mounted in a direct-connect, cold-flow, Mach 1.8 wind tunnel. Back pressure was created within the analog isolator using an adjustable ramp mounted in the downstream portion of the test section. All the models approximate a Hammerstein model in structure and include a static nonlinear polynomial model as well as a dynamic model. While the same static polynomial is used for all models, the dynamic models considered include simple continuous models, Auto-Regression with eXogenous variables (ARX) models, and Nonlinear ARX (NARX) models. Model development was accomplished using post-processed data with three of the models implemented and tested experimentally in a real-time (1 kHz) algorithm. All three models were able to predict the location of the shock-train to within one duct height and could, potentially be used for control algorithm development. The ARX and the NARX models returned comparable results in the real-time trials.