Research On Aircraft Fuel Quantity Calculation Method Based On Time Series Data

With the improvement of the country's comprehensive national strength,the aviation field has developed vigorously,and the degree of informatization of aviation equipment has gradually increased,which has put forward higher requirements for aircraft fuel systems.Aircraft fuel quantity calculation is an important part of aircraft fuel system,and its calculation accuracy is an important factor that affects aircraft flight safety,effective range and overall performance.However,the current aircraft fuel quantity calculation methods mainly focus on static fuel data(that is,data collected at a certain moment by aircraft sensors).This method ignores the contextual relationship of fuel data in the time dimension,which makes it difficult to learning the accurate mapping relationship between fuel data and aircraft fuel quantity,the calculation accuracy is difficult to meet the requirements.In order to solve this problem,this article analyzes the fuel data as time series data.There are two problems in this kind of fuel calculation idea:1)Due to various factors,there will be missing values in the fuel data,and the analysis of time series data needs to ensure the integrity of the data.2)The mapping relationship between fuel data and aircraft fuel quantity is relative ely complex,and it is difficult to analyze it directly,and the problem needs to be simplified.This paper mainly studies the above problems and proposes corresponding solutions.The main contents are as follows:1)Fuel data missing value filling method based on generative adversarial networkIn order to solve the problem that there are missing values in the fuel data and the existing missing value filling method requires a complete data set for training,this paper use a fuel data missing value filling method based on generative adversarial network.In addition,in order to solve the problem of information loss when the input data is too long,this paper introduces an attention mechanism in the Seq2 seq model of the generator part,which can slightly alleviate the problem.This method can realize the end-to-end filling of missing fuel data,and can achieve the best filling effect at present.2)Aircraft fuel quantity calculation method based on attitude perceptionAiming at the problem that the mapping relationship between fuel data and aircraft fuel quantity is relatively complex,this paper designs a method for calculating aircraft fuel quantity based on attitude perception.This method first classifies the fuel data according to the category information implied in the fuel data,and uses the corresponding prediction model for the fuel quantity prediction for different types of fuel data,and then the probability distribution of each category and its corresponding fuel quantity prediction result are fused to obtain the final fuel quantity calculation result.This method can learn the historical change law of fuel data,and apply the category information implied in the fuel data to the aircraft fuel quantity calculation model,which can effectively improve the fuel quantity calculation accuracy.