Fuel Consumption Model of Aircraft in Descent Stage Based on DBN

Abstract

Due to the pressures of the current ecological environment and the rise of fuel prices, it is necessary to calculate the fuel volume of aircraft accurately. In order to calculate the fuel consumption of a flight, the key is to establish an accurate fuel consumption model. In the process of descent, because the environment around the aircraft changes dramatically, compared with other stages, the fuel consumption factors affecting the descent stage will be more. But at present, the domestic fuel consumption model for aircraft descent stage is not accurate enough. To solve this problem, a method of building fuel consumption model in aircraft descent stage based on flight data using Deep Belief Network (DBN) is proposed. Firstly, the parameters related to fuel consumption are selected from the flight data, then the correlation between each parameter and fuel consumption rate is calculated by mutual information algorithm, and finally the parameters with the highest correlation are selected for model training. Compared with the traditional BP neural network and Echo State Network (ESN), the accuracy has been greatly improved.