Thermal Buckling in CWR Tracks: Critical Aspects of Experimental Techniques for Lateral Track Resistance Evaluation

After the introduction of continuous welded rail, thermal track buckling has been recognized to be one of the unsolved problems caused by this technological railroad improvement. In general, both weak ballast strength in the lateral direction and large alignment defects are the principal causes of such phenomenon. In the UIC 720 Leaflet, which is the reference standard for the realization and maintenance of continuous welded rail tracks, two safety criteria against thermal track buckling are described: one is based on the maximum (∆Tmax) and minimum (∆Tmin) buckling temperatures, the other only on the minimum buckling temperature. In the literature, it is found that a correlation exists between ∆Tmax and the maximum (or peak, FP) lateral resistance value of the tie-ballast system, and, analogously, between ∆Tmin and the minimum (or limit, FL) lateral resistance. For this reason, railway technicians had to paid special attention in the assessment of FP and FL. Because FP concerns lateral displacements equal to 5÷10 mm, and FL occurs for lateral displacements greater than 80÷100 mm, some researchers have proposed to measure FP with a quasi-non-destructive experimental technique, the Single Tie Push Test (STPT), and, successively, to evaluate FL as a function of FP by empirical formulas, in place of the experimental evaluation of the full lateral resistance curve of the tie-ballast system. Based on these considerations, a concerning issue arises whether it is sufficient, and above all safe, to use the simpler, less destructive, and less expensive STPT technique, which requires that only one tie is detached from the rails, or if it is necessary to perform lateral resistance tests on track panels composed by 4 to 6 ties, as in the case of the Discrete Cut Panel Pull Test (DCPPT). For this purpose, in this paper the experimental results obtained in situ in full scale conditions with the two testing techniques are reported, and the differences obtained by performing tests with one, two, and four ties are analyzed with the aim of ensuring a safe evaluation of the main input parameters required for buckling temperatures calculation. It is found that the limit lateral resistance depends neither on the chosen experimental technique, nor on the compaction level of the ballast bed, whereas the peak lateral resistance appears to be dramatically altered if it is evaluated by mean of the STPT, with serious risks of an unsafe evaluation of the buckling temperatures.