Evaluation of Distributed Acoustic Sensing (DAS) Systems with Polarization Diversity for Railway Infrastructure: Effects of Fiber Type and Installation Conditions on System Performance

Distributed Acoustic Sensing (DAS) systems based on ϕ-OTDR are increasingly used for real-time infrastructure monitoring, but their performance is limited by coherent and polarization fading, which create signal blind spots along the fiber. This study evaluates two coherent-detection DAS architectures—with and without polarization diversity—together with the influence of optical cable type and ground-coupling method, in an operational railway environment. Field experiments were conducted along a 40 km SNCF (Société Nationale des Chemins de fer Français) railway line using train passages as excitation sources, complemented by controlled sledgehammer tap tests on a 500 m calibration platform connecting four fiber configurations (telecom vs. sensing cables, buried vs. duct-installed). Signal amplitude, differential phase, and phase Median Absolute Deviation (MAD) were used to quantify sensing performance. Polarization diversity reduced the fraction of faded signals below −40dB from approximately ≈8.4% to ≈1% of the fiber length and improved the Phase Stability Index (PSI) from ≈6dB to ≈10dB at 20 km, significantly mitigating polarization-induced signal dropouts across the 40 km span. Among the tested cables, the tight-buffered sensing fiber directly buried in the ground consistently provided the strongest and most stable response, while telecom fiber installed in a High-Density Polyethylene (HDPE) duct showed the weakest coupling. These results demonstrate that combining polarization-diversity detection with an optimized fiber-cable coupling strategy substantially improves the robustness of DAS for railway infrastructure monitoring.