Comparison of aircraft braking performance on wet and snowy runway using BPN test, CEL numerical simulation and LSTM deep learning

The assessment of aircraft braking performance on wet and snowy runway under adverse weather condition is a critical issue. This study proposed an integrated experimental-numerical-deep learning research framework to compare and assess aircraft braking performance on wet and snowy runways. Specifically, the framework integrates British Pendulum Number (BPN) tests, Finite element method (FEM) simulation and LSTM deep learning. The BPN test conducted in a self-developed low temperature weather simulation laboratory to evaluate anti-slip performance and investigate the causes of friction degradation on wet and snowy pavement surfaces. FEM simulations employing the Coupled Eulerian-Lagrangian (CEL) approach to analyze friction coefficient and velocity decay under both steady-state and unsteady-state conditions, and furhter investigated the influence of key driving condition parameters on anti-skid performance, including velocity, water depth, snow depth, snow type, and slip ratio. Furthermore, Long Short-Term Memory (LSTM) networks was utilized to achieve highly accurate velocity decay predictions (with errors ≤0.0095 m/s) based on FEM-derived velocity data, thereby facilitating precise braking distance estimation through time-velocity integration. The research findings revealed distinct friction mechanisms between wet and snowy runways, where hydroplaning predominates on wet surfaces at high velocity, while snow compression reduces pavement roughness even with shallow snow depth (<1 mm depth), and denser snow offers reduced friction. Furthermore, the braking distance predictions showed that runway covered with dense snow require 23 % longer braking distance than wet surface at 150 km/h with 1 mm water depth, and wet conditions presented the highest overrun risk at high velocity, exceeding 3000 m at 200 km/h with 3 mm water depth. Above all, this study provided airport operators with quantitative criteria to evaluate skid resistance and braking performance on both wet and snow-covered runways through multidimensional analysis.