This paper presents the design and analysis of a data science competition on a problem of time series regression from aeronautics data. For the purpose of performing predictive maintenance, aviation companies seek to create aircraft "numerical twins", which are programs capable of accurately predicting strains at strategic positions in various body parts of the aircraft. Given a number of input parameters (sensor data) recorded in sequence during the flight, the competition participants had to predict output values (gauges), also recorded sequentially during test flights, but not recorded during regular flights. The competition data included hundreds of complete flights. It was a code submission competition with complete blind testing of algorithms. The results indicate that such a problem can be effectively solved with gradient boosted trees, after preprocessing and feature engineering. Deep learning methods did not prove as efficient.

Vibration levels that onboard equipment must be able to withstand throughout their lives for correct operation are mainly determined experimentally because predicting the dynamic behavior of a complete aircraft requires computational means and methods that are currently difficult to implement. We present a data-driven methodology that leverages flight-test accelerometer data to produce a predictive model. This model, based on an ensemble of artificial neural networks, performs a multioutput multivariate regression to estimate vibration spectra from a set of aircraft general parameters without having to characterize excitation sources. The model is compared with baseline models over two protocols, which are 1) standard training and testing as well as 2) extrapolation to high dynamic pressures, in order to assess physical consistency. Although the first protocol shows that all models can produce results accurate enough for this context, the second protocol shows that only the ensemble model is able to correctly extrapolate the energy. Using the Shapley additive explanations method, also known as SHAP, we show that these results can be explained by the ability of our model to identify the dynamic pressure as the core feature used in the extrapolation protocol. The proposed model can be used in multiple applications, such as anomaly detection and vibration flight envelope opening.