APPLICATION OF BACK PROPAGATION NEURAL NETWORK TO THE ANALYSIS OF AIRCRAFT WITH LOADING

Abstract

Using the flying parameters to map strain on the aircraft by Neural Network is easier and more reliable than that of using the traditional method. Also, taking too many flying parameters will not only increase the complexity of the network structure, but also the difficulty of the learning scheme. Screening out the unwanted parameters is an important procedure to the strain measurement. In this paper, we use the TQEM (Taguchi Quality Engineering Method) is applied to chose the required flying parameters for better referencing. The demonstrated results show that the TQEM is a new and effective method to choose the required parameters for system identification.