A rare failure detection model for aircraft predictive maintenance using a deep hybrid learning approach

Abstract

The use of aircraft operation logs to develop a data-driven model to predict probable failures that could cause interruption poses many challenges and has yet to be fully explored. Given that aircraft is high-integrity assets, failures are exceedingly rare. Hence, the distribution of relevant log data containing prior signs will be heavily skewed towards the typical (healthy) scenario. Thus, this study presents a novel deep learning technique based on the auto-encoder and bidirectional gated recurrent unit networks to handle extremely rare failure predictions in aircraft predictive maintenance modelling. The auto-encoder is modified and trained to detect rare failures, and the result from the auto-encoder is fed into the convolutional bidirectional gated recurrent unit network to predict the next occurrence of failure. The proposed network architecture with the rescaled focal loss addresses the imbalance problem during model training. The effectiveness of the proposed method is evaluated using real-world test cases of log-based warning and failure messages obtained from the fleet database of aircraft central maintenance system records. The proposed model is compared to other similar deep learning approaches. The results indicated an 18% increase in precision, a 5% increase in recall, and a 10% increase in G-mean values. It also demonstrates reliability in anticipating rare failures within a predetermined, meaningful time frame.

Improved decision making with similarity based machine learning: applications in chemistry

“…Moreover, the implicitly defined domain of applicability of the model provides an indication for the reliability of a machine learning model for unseen queries and inherently limits the usefulness outside of this domain (extrapolation [85]). Potential applications beyond chemistry could include predictive maintenance tasks in computer aided infrastructure [86] or aerospace engineering [87,88]. Future work will address, automated experimental design, deriving ready made models in quantum chemistry or multilevel learning, improving similarity measurements, as well as exploring fast and approximate neighbour searches (cf turbo similarity fusion) [76,[89][90][91].…”

Multi-modal LSTM network for anomaly prediction in piston engine aircraft

“…On the other hand, [33] proposes RUL estimation using multiple machine learning modules including SVM, K-Nearest Neighbour (KNN), RF, and Analysis of Variance (ANOVA) statistical approaches. In [34] a combination of 1D CNN, autoencoder, and bidirectional GRU network is used on time-series data of 60 turbojet commercial aircraft in a pursue of detecting rare engine failures. This approach primarily addresses data imbalance to achieve favourable performance.…”

Machine Learning Approaches Using Phog and Smote of Image Feature Extraction Technique for Predicting Control Valve Stem Position

“…[1] Dangut et.al. [7] Proposed model is compared to other similar deep learning approaches. The results indicated an 18% increase in precision, a 5% increase in recall, and a 10% increase in G-mean values.…”

“…The positioner will amplify the air supply from actuator to precisely open the valve for the given controller output. The signals being operated are followed universally like 4mA to 20mA, 3-15 psi or (1-5) Volts and So on but often (4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) mA is used in industrial automation. In the current scenario a pneumatic control valve with positioner is used to collect the image data for building an optimum model using machine learning techniques.…”