Smart segmentation in tram track geometry analysis

Analysis of track geometry along 200 or 100-meter-long consecutive segments is a standard procedure for track quality evaluation on standard and wide gauge, conventional and high-speed rail systems. Previous research shows that such analysis on narrow gauge tram tracks could be misleading and that tram networks – due to the marked inhomogeneity of the factors influencing the geometry degradation – could benefit from reducing the analytical segment length. This paper investigates both the effect of shortening the analytical segment length and the introduction of “smart segmentation” in the analysis of tram track geometry. The smart segmentation considers the tram track route curvature, construction elements, exploitation intensity, and tram speed, but also the external effects from road vehicles, and the organization of future maintenance works. The investigation was carried out on a 2.2-kilometre-long tram track section in Zagreb, Croatia. Track geometry was measured using a manually operated trolley. The analysis of the results showed that the geometry degradation along the section is most pronounced in gauge deviations expressed as relative values to the regular track width of 1,000 millimetres. The section was divided into 22 100-meter-long and 44 50-meter-long consecutive segments, and 45 50 to 25-meter-long smart segments. On each segment, a mean gauge deviation value was calculated as a measure of central tendency. Standard deviation, mean absolute deviation and interquartile range were calculated as measures of gauge deviation values dispersion. The results show that shortening the segment to 50 meters reduces the dispersion of gauge deviation values along the segment by 10% and that introducing a smart segment reduces it by 25%.