Search Results

A new approach for the prediction of lift, drag and moment coefficients is presented. This approach is based on the Support Vector Machines methodology, and on a optimization algorithm, the Extended Great Deluge. The novelty of this approach is the combination between the SVM and the EGD algorithm. The EGD is used to optimize the SVM parameters to allow it to predict the aerodynamic coefficients of ATR 42 model. The training and validation of this new combination method is realized using the aerodynamic coefficients of an ATR-42 wing model with Xfoil software and experimental tests using the Price-Païdoussis wind tunnel. The results obtained with our approach are compared with the XFoil results, experimental results and XFLR5 software results for different flight cases, expressed as various combinations of angles of attack and Mach numbers. The main purpose of this methodology is to rapidly predict aircraft aerodynamic coefficients.

This paper describes an optimization algorithm that provides an economical Vertical Navigation profile plan by finding the combinations of climb, cruise and descent speeds, as well as the altitudes for an aircraft to minimize flight costs. The computational algorithm profits from a space search reduction algorithm to reduce the initial number of speed and altitude combinations. Additional search space reductions were performed with the implementation of the branch and cut algorithm. A bounding function that correctly estimates the flight cost considering step climbs was developed to reduce the number of calculations. The full flight fuel burn cost was obtained using a performance database- based method. The fuel flight cost was computed using the cost index. This algorithm used a performance database instead of equations of motion to compute fuel burn. This database was developed and validated by our industrial partner using real flight experimental data.