A NEW ALTERNATIVE AIR DATA COMPUTATION METHOD BASED ON ARTIFICIAL NEURAL NETWORKS

• Main Authors: İlke Türkmen, Seda Arık
• Format: Article
• Language: English
• Published: Hezarfen Aeronautics and Space Technologies Institue 2017-01-01
• Series: Havacılık ve Uzay Teknolojileri Dergisi
• Subjects: Air data parameters; air data computer; artificial neural networks
• Online Access: http://www.jast.hho.edu.tr/JAST/index.php/JAST/article/view/24/17
• ISSN: 1304-0448; 1304-0448

Air Data Computer (ADC) is an important equipment of the aircraft and the performance of the ADC directly affects the safety and efficiency of the flight. The ADC uses sensors to get small amounts of original messages, such as dynamic pressure, static pressure, and total temperature and computes the air data parameters such as airspeed, pressure altitude, Mach number, static air temperature etc. that have fundamental importance for flight control systems and navigation systems. When ADC failure occurs; there is no alternative way to compute air data parameters in the aircraft. In order to overcome this problem, in this paper, an alternative air data computation method based on artificial neural networks (ANN) is presented. The data set used to train proposed neural model is obtained from the Digital Flight Data Acquisition Unit (DFDAU) of a commercial Boeing 737-400 type aircraft. Simulation results clearly show that the proposed neural method can be used as an alternative air data computation method when ADC failure. The proposed method also provides simple and high accuracy method for computation of the air data parameters instead of traditional nonlinear ADC equations.