Search Results

A lubricating system modeling method based on flight test data is proposed in this paper. ANN model based on a large number of flight test data is trained and validated, and models of 6 lubricating system parameters in all engine operation settings and whole flight envelope are established. Model results are in good agreement with flight test results, which shows feasibility and effectiveness of the presented modeling method. The model results are packaged in dynamic link library, and the coordination between calculating model and GDAS is accomplished. Comparison of model and flight test results in real-time monitoring of flight test comes true, thus on-line trend monitoring of oil parameters is implemented and applied. Additionally, input parameters are gradually decreased as new input parameter group of ANN structure. Oil parameter model is trained and validated again with the new group of parameters, until leading to unacceptable bias between model and flight test results.