Diagnostic models for sensor measurements in rocket engine tests

This paper presents our ongoing work in the area of using virtual reality (VR) environments for the Integrated Systems Health Management (ISHM) of rocket engine test stands. Specifically, this paper focuses on the development of an intelligent valve model that integrates into the control center at NASA Stennis Space Center. The intelligent valve model integrates diagnostic algorithms and 3D visualizations in order to diagnose and predict failures of a large linear actuator valve (LLAV). The diagnostic algorithm uses auto-associative neural networks to predict expected values of sensor data based on the current readings. The predicted values are compared with the actual values and drift is detected in order to predict failures before they occur. The data is then visualized in a VR environment using proven methods of graphical, measurement, and health visualization. The data is also integrated into the control software using an ActiveX plug-in.