Many structural elements are exposed to load conditions that are difficult to model during the design phase, such as environmental uncertainties, random impacts, and overloading, amongst others, thus increasing unprogrammed maintenance and reducing confidence in the reliability of the structure in question. One way to deal with this problem is to monitor the structural condition of the element. This approach requires supervising several signals coming from critical locations and then performing an accurate condition estimation of the element in question based on the data collected. This study implements a method to diagnose and evaluate the reliability of the bolster beam structure of the railway vehicle during a fatigue test. The results show that multidimensional monitoring not only diagnoses the element accurately but also results in correct estimation of reliability.