Abstract

Fouling and deterioration of the ballast are major problem in operation of railway tracks. Its characterization is fundamental to optimize decision making in maintenance programs. This study aims to characterize the fouled ballast by dynamic penetrometer (PANDA) and geoendoscope test (This method, hereafter PANDA and geoendoscope test, is known as Pandoscope (R) [17]). The methodology presented consists in analyzing the dynamic cone resistance (qd) of each layer and then performing a geoendoscope test to obtain borehole images. Statistical analysis of cone resistance and image parameters (texture and colorimetry) shows their dependence on the degree and of fouling material. As ballast fouling increases, the average cone resistance initially increases and then decreases. Image features also depend on the proportion of fouling but especially allow to distinguish different contents of fine particles in the fouling material. Nevertheless, each of the measurements taken separately is not enough to determine a ballast fouling amount. In this way, data fusion has been carried out and shows an improvement in the quantification of the fouling index [1].