Research On The Interval Estimation Of Aircraft Fuel Consumption Based On US-D-DD

The actual fuel consumption analysis and prediction model parameters have strong uncertainty,and in most cases,the relative differences of traditional fuel consumption due to actual payload,flight time,wind speed and other reasons are not fully considered,which affects the overall range estimation results.Therefore,it is necessary to further quantify the variation range of prediction results caused by uncertain factors on the basis of previous point prediction results,and establish the interval estimation model of aircraft fuel consumption for all itineraries data of the same aircraft type according to the data deviation and density distribution.Main research content is specifically as follows:First,a new method for the deviation of fuel consumption and undersampling of density distribution is proposed.The whole range of the same aircraft is divided by numerical values.According to the mean value and deviation degree of all fuel consumption values in each itinerary,the range data set whose deviation degree is greater than the threshold value is selected,and the fuel consumption data is analyzed by density analysis.According to the deviation and retention method,the data cluster is under sampled to remove the influence of a few itineraries on the overall results.Second,establish the model of interval estimation.Using the relevance vector machine interval estimation model to get the upper and lower boundaries of each itinerary and fit them.In the forecast results,the total forecast results of the total population are quantified to show the probability of change in fuel consumption.At the same time,the interval estimation model is divided into four evaluation indexes: Estimation Interval Coverage Percentage,Mean Width Percentage,Mean Accumulated Deviation and Estimation Interval Satisfaction Index.Third,complete the division of aircraft round-trip heading fuel consumption data.According to the monsoon wind of northern hemisphere in summer or winter,the consumption consequence will be seperated into more fuel consumption part and less fuel consumption part.The experimental results show that the model based on data deviation and density distribution under sampling method achieves better interval estimation effect,improves the overall interval results,and obtains more accurate interval range for the division of round-trip heading fuel consumption data.