The Aircraft Fuel Consumption Monitoring Method And Monitoring System Design Based On The QAR Data

Fuel cost has become the biggest cost of airlines in our country. As the oil prices rising,the fuel cost is severely eroding the profits of airlines, so it is very important to implement oil-saving work for the airlines to achieve the fine management of fuel. Before the implementation of oil-saving work, it needs to know the detailed information of the fuel consumption. So it is very important to monitor the fuel consumption.In the study of aircraft fuel consumption monitoring, a segment monitoring method and a fuel consumption model-based monitoring method were proposed. Segment monitoring method divided the fuel consumption monitoring as five parts to monitor the fuel consumption information of sliding out period, climb period, cruise period, decline period and sliding into period. A model was established based on the monitoring results, to monitor the trend of the fuel consumption in the fuel consumption model-based monitoring method. This realized mining the information of fuel consumption in-depth.A fuel consumption model based on support vector regression was established, and it only studied the fuel consumption model of climb period, cruise period and decline period.First, it used the correlation analysis method to analyze the relationship between the fuel consumption and the influence factors during the climb period, cruise period and the decline period, in order to determine the input variables of fuel consumption model. It was mainly used the correlation to determine the input variables of fuel consumption model. Secondly, to realize the parameter optimization, a grid search algorithm in combination with the cross-validation was proposed. This algorithm improved the precision of the model. Finally,the fuel consumption model of climb period, cruise period and the decline period were established based on support vector regression. And the prediction result of SVR model was compared with the prediction result of neural network model, the result show that the performance of SVR model was better.Finally, a fuel consumption monitoring system was designed based on modular design.This system realized comprehensive monitoring the fuel consumption and it can mining the information of fuel consumption in-depth. It has certain reference significance for the airlines.