The relevance of the transition from preventive maintenance of aviation equipment to predictive maintenance is substantiated.Various methods for building predictivemaintenance systems are considered:linear regression,...The relevance of the transition from preventive maintenance of aviation equipment to predictive maintenance is substantiated.Various methods for building predictivemaintenance systems are considered:linear regression,digital twins,machine learning,artificial neural networks,etc.It is shown that today there is no unified approach to the issues of predicting the technical condition of aircraft units based on data obtained during operation.An approach to the predictive analysis of failure and prefailure states of aircraft units using data obtained during their operation is proposed.An approach to the predictive analysis of failure and pre-failure states of aircraft units using data obtained during their operation is proposed.This approach is based on the joint application of such technologies and methods as a digital twin,linear regression,k-nearest neighbors,and event-based mathematical model.In the example of the analysis of a certain array of flight information obtained during the operation of the aircraft,the efficiency of the proposed approach is shown.展开更多
文摘The relevance of the transition from preventive maintenance of aviation equipment to predictive maintenance is substantiated.Various methods for building predictivemaintenance systems are considered:linear regression,digital twins,machine learning,artificial neural networks,etc.It is shown that today there is no unified approach to the issues of predicting the technical condition of aircraft units based on data obtained during operation.An approach to the predictive analysis of failure and prefailure states of aircraft units using data obtained during their operation is proposed.An approach to the predictive analysis of failure and pre-failure states of aircraft units using data obtained during their operation is proposed.This approach is based on the joint application of such technologies and methods as a digital twin,linear regression,k-nearest neighbors,and event-based mathematical model.In the example of the analysis of a certain array of flight information obtained during the operation of the aircraft,the efficiency of the proposed approach is shown.
