Design of Artificial Neuro-fuzzy Based Methodology for Six Component Force Balance

Abstract

A blunt cone model of 600 apex angle and 150 flare with internally mountable accelerometer balance has been considered for the present investigation on six component force balance. The finite element method (FEM) has been incorporated at predetermined angles of the applied force to obtain multiple accelerations in three spatial directions so as to replicate a six component force balance system. A novel intelligent soft computing technique, Artificial Neuro-Fuzzy Inference System (ANFIS) has been implemented for accurate prediction of the magnitude and trend of aerodynamic forces and moments from the transient acceleration history. The same method is also used to validate the universal approximation nature of the impulse force by accurate prediction of the hat and ramp forces from the training of the impulse forces. Furthermore, a training algorithm is deduced so as to predict the force and moment magnitude for the three and single component force balance from the six component training data. Henceforth, novelty of this study involves in deduction of a universal training method and successful implementation of it to predict short duration transient force and moment histories.