Research On Multivariate Model Of Aircraft Fuel Consumption Based On Neural Network

With the rapid development of the civil aviation transportation industry, energy conservation and emission reduction of civil aviation transportation has become a hot topic. In addition, the international community increasingly concerned about environmental issues, and environmental protection has become an indicator of the social responsibility of airlines.Therefore, the establishment of fuel consumption model is one of the most important steps for energy conservation and emission reduction.Due to the complex relations among the various factors, the aircraft fuel consumption is significantly nonlinear. The difficult point of this topic is how to describe and express the complex relations among various factors, the key is the prediction accuracy of models. To solve this problem, the main work of this paper are as follows:First, this paper studies key factors of airline fuel consumption, and extracts parameters from QAR data.Second, in this article, there are two ways to establish fuel consumption models based on BP neural network. The first one is a linear model. The second one is a nonlinear model that after studied the single-factor model of each flight phase.Finally, MIV algorithm is introduced to analysis the structure of factors, the experimental results shows the directions and degree of effect. In addition, the experiment studied influencing factors of three phases of flight and quantified their effect sizes based on sensitivity analysis method. Then, the contribution caused by small changes of each factor on fuel consumption of each phases can be determined.The simulation results suggests that the models built up in this paper performing with good precision, and the experimental results based on MIV algorithms and sensitivity analysis method has reference value in actual aircraft fuel consumption estimation and structure analysis of the factors.