Model-based fault diagnostics in an electromechanical actuator of reusable liquid rocket engine

Electromechanical Actuators (EMAs) are gaining important roles in liquid rocket engines. A model-based fault diagnosis method for EMAs is under development at JAXA to realize affordable maintenance cost and a quick turnaround time for reusable launch vehicles. In this method, multi-physics system-level modeling and simulation for a target system are conducted by using Modelica, an equation-based, object-oriented modeling language having important modeling capability for system-level interactions that involve multi-physics phenomena to enable fault diagnostics in several different physical domains and components. In this proposed approach, Modelica-based modeling and simulation are conducted for a target system under both normal and abnormal conditions, based on an understanding of the failure mechanism to obtain prior data sets for fault diagnosis. The dynamic time warping algorithm is utilized to evaluate the dissimilarity between the prior data sets and sensor measurement data obtained during flight, and a hierarchical clustering technique is applied for categorization in failure mode based on the dissimilarity of these data. In this study, the proposed method was successfully demonstrated with fault injection experiment data.