Research On On-line Diagnosis Technology And Reliability Of Airport Blind Landing Equipment

With the continuous progress of science and technology,the new technology of civil aviation has been successfully applied in large hub airports.In order to ensure the safe approach and landing of aircraft,civil aviation departments have invested a lot of money in the construction of airport blind landing equipment system in recent years.Of course,even very precision equipment has a life cycle,especially in similar airports where the performance requirements of equipment are very high,which requires staff to have a clear idea about the operation status of equipment in order to find out the fault equipment in advance,so as to minimize the harm caused by the fault equipment.However,it is undoubtedly a time-consuming and laborious work for staff to inspect a large number of equipment blindly,so it is an urgent problem to realize the fault diagnosis of airport landing equipment.In order to centralize the monitoring and management of all the equipments,find and remove the faulty equipments in time and improve the inspection efficiency of the staff,the airport department has increased the development of the blind landing equipment monitoring system.This paper takes Chongqing Jiangbei airport blind landing equipment as the research object,takes random forest algorithm as equipment fault diagnosis model,studies the fault diagnosis method under complex circumstances,and studies the reliability of airport blind landing equipment,infers the probability distribution of its conformity,and improves the efficiency of airport staff to check blind landing equipment.The fault diagnosis method adopted in this paper is random forest algorithm.Random forest has higher classification accuracy and lower prediction error than single decision tree classifier.It is suitable for many environments,does not need pruning,is insensitive to noise data and so on.It has been widely used in many fields.But like other classifiers,the unbalanced degree of data sets will greatly interfere with the classification of classifiers.In the airport blind landing equipment monitoring system,the fault equipment data occupies a small proportion in the monitoring data,which reduces the accuracy of random forest classification.In this paper,an improved algorithm,SCSMOTE algorithm,is proposed.According to the degree of boundary discrimination between a few samples and a majority of samples,the appropriate candidate samples are obtained,and the center of the candidate samples is calculated.Finally,synthetic samples are generated on the line between candidate samples and their centers to achieve the balance of data sets.The improved new algorithm is used to synthesize the fault data in the monitoring data to achieve the balance of the monitoring data.Then the random forest is used as the method of on-line diagnosis of equipment,and a better diagnosis model is trained to realize on-line diagnosis of airport blind landing equipment.Based on the development of ground navigation equipment monitoring system of Jiangbei Airport in Chongqing,this paper obtains the fault data of blind landing equipment.In this paper,the time between failures data of blind landing equipment are taken as the observation sample.According to the frequency of observation samples of blind landing equipment,the probability density function and experience distribution function of time between failures of blind landing equipment are obtained,and the graph fitting is made and it is inferred that it may conform to Weibull distribution.In this paper,the least square method is used to transform the Weibull distribution function linearly and the distribution parameters are estimated.Finally,the Weibull distribution model of blind landing equipment's time between failures is obtained and the reliability analysis is completed by correlation test and fitting hypothesis test.The test results show that the blind landing equipment's time between failures obeys the Weibull distribution model.