Condition-Based Maintenance for Diagnosis and Prognosis in Aircraft Systems - Application in the Air Bleed System of Boeing 747

Authors

Abstract

Aircraft maintenance is an important subject matter in the aircraft field and, as more useful information is gathered and processed, improvement in this area is valuable to the aircraft industry. In this particular topic, Condition-Based Maintenance (CBM) can be useful, as it can help predict when a failure will occur based on the component's condition. Using CBM, a Prognostics and Health Management (PHM) approach can be built with the objective of identifying the degradation behavior of the aircraft equipment, and anticipating possible system failures by predicting the future degradation evolution.Therefore, this work aims to develop a PHM methodology that, based on the sensors data, is capable of diagnosing the health condition of an aircraft system, reflected by the Health Indicator (HI) value, and predict its future behavior, resulting in the computation of the Remaining Useful Life (RUL). The PHM approach is applied distinctively to three different aircraft systems, depending on the system data characteristics, different machine learning techniques are applied. An exploratory work is performed on two systems, the Turbofan engine system and the Brakes system, with the goal of testing different techniques for the diagnosis and prognosis of the systems condition.The main contribution of this Thesis results from the work applied in the Air Bleed system of Boeing 747, which contains a significant complexity embedded. In this particular system, a new data driven technique for the HI computation is proposed, which is based on the analysis of time domain features, namely the mean and standard deviation, of the raw sensors data.The scope and aim of the PHM development fits within the scope of the H2020 ReMAP project. Furthermore, some of the systems sensors data, namely the Brakes and the Air Bleed data were provided by the ReMAP project.Regarding the obtained results, particularly in the Air Bleed system, these are interesting and very promising. The formulation used for the computation of the HI was positively received by the airlines' engineers involved in ReMAP, due to its innovative content. Furthermore, it is expected that this approach can be applied to other aircraft systems and, in a near future, be integrated in the maintenance plan, as a valid contribution for the execution and planning of aircraft maintenance routines.

Keywords

Aircraft Maintenance, Condition Based Maintenance, Health Indicator, Prognostics and Health Management, Remaining Useful Life

Subject

Aircraft Condition Based Maintenance

Related Project

H2020-REMAP – Real-time Condition-based Maintenance for Adaptive Aircraft Maintenance Planning

MSc Thesis

Condition-Based Maintenance for Diagnosis and Prognosis in Aircraft Systems - Application in the Air Bleed System of Boeing 747 , September 2019