In order to meet the mounting social and economic demands, railway operators and manufacturers are striving for a longer availability and a better reliability of railway transportation systems. Commercial trains are being equipped with state-of-the-art onboard intelligent sensors monitoring various subsystems all over the train. These sensors provide real-time spatio-temporal data consisting of georeferenced timestamped events that tend sometimes to occur in bursts. Once ordered with respect to time, these events can be considered as long temporal sequences that can be mined for possible relationships leading to association rules. In this paper, we propose a methodology for discovering association rules in very bursty and challenging floating train data sequences with multiple constraints. This methodology is based on using null models to discover significant co-occurrences between pairs of events. Once identified and scrutinized by various metrics, these co-occurrences are then used to derive temporal association rules that can predict the imminent arrival of severe failures. Experiments performed on Alstom's TrainTracer data show encouraging results.