Classification prognostics approaches in aviation

Marcia L. Baptista, Elsa M.P. Henriques, Helmut Prendinger
2021 Measurement (London)  
Traditionally, prognostics approaches to predictive maintenance have focused on estimating the remaining useful life of the equipment. However, from an industrial point of view, the goal is often not to predict the residual life but to determine the need for a maintenance action at a given time. This approach allows us to frame the data-driven prognostics problem as a binary classification task rather than a regression one. To address this problem, we propose in this paper to explore the
more » ... e strengths and limitations of a set of classifier approaches such as random forests, support vector machines, Gaussian processes, and deep learning techniques. We evaluate the models using metrics such as sensitivity, specificity, accuracy, receiver operating characteristic curve, and Fscore. This work's novelty lies in the adopted approach and the proposal and comparison of a set of classifier models. This comparison is made on two realworld datasets from the aeronautics sector. Results indicate that the proposed deep learning classifier methods are well suited for this kind of prognostics. Abstract Traditionally, prognostics approaches to predictive maintenance have focused on estimating the remaining useful life of the equipment. However, from an industrial point of view, the goal is often not to predict the residual life but to determine the need for a maintenance action at a given time window. This approach allows us to frame the data-driven prognostics problem as a binary classification task rather than a regression one. To address this problem, we propose in this paper to explore the relative strengths and limitations of a set of classifier approaches such as random forests, support vector machines, nearest neighbors, and deep learning techniques. We evaluate the models using metrics such as sensitivity, specificity, accuracy, receiver operating characteristic curve, and F-score. This work's novelty lies in adopting a modeling approach with a natural probabilistic interpretation of the prognostics exercise. The comparison of an extensive range of classifier models is performed on two real-world datasets from the aeronautics sector. Results indicate that deep learning classifier methods are well suited for this kind of prognostics and can outperform by a significant margin the traditional classification techniques. Importantly, the proposed modeling approach aims to generate an alternative prognostics representation that goes in line with the expectations of aeronautical engineers. J o u r n a l P r e -p r o o f Journal Pre-proof 165 namic model has been shown to help improve the modeling accuracy [28]. 2 J o u r n a l P r e -p r o o f Journal Pre-proof J o u r n a l P r e -p r o o f Journal Pre-proof 355 works, cut-off values are applied to multidimensional sensor data (or to a health index) to determine the end-of-life 4 J o u r n a l P r e -p r o o f Journal Pre-proof 515 the prediction window, i.e., the days in advance that the algorithm should alert the occurrence of a health event 7 J o u r n a l P r e -p r o o f Journal Pre-proof 580 iments in dataset DS-2. This choice was made to respect the window sizes usually used in the industry. 8 J o u r n a l P r e -p r o o f Journal Pre-proof J o u r n a l P r e -p r o o f Journal Pre-proof J o u r n a l P r e -p r o o f Journal Pre-proof Data-driven techniques have recently become a popular approach to prognostics. We study classification methods instead of regression methods to detect failure. We test our approach on two real-world large scale case studies from aeronautics. Deep learning classifier approaches are well-suited to address the prognostics exercise. J o u r n a l P r e -p r o o f Journal Pre-proof
doi:10.1016/j.measurement.2021.109756 fatcat:67q3owko3vdk3lqztx3dnjahme