摘要：Abstract A diagnosis benchmark was developed for the MASCOTTE test bench of ONERA/CNES. It consists of a CARINS Simulator of the bench, model-based fault diagnosis algorithms and firing tests results. The objective of this work was to demonstrate alternatives to the classical redline monitoring strategy used on the bench. A simplified model of the cooling circuit of the test bench was built and used to apply a parameter identification detection method and a Kalman filter. The diagnosis flags are obtained through analysis of the residuals via CUSUM tests. This article shows the results obtained and proposes an example of methodology to identify and validate performances of new diagnostic strategies.
