Skip to Main Content
The free market forces infrastructure operators to increase their operational efficiency. To achieve this, among other things personnel costs have to be reduced. This implicates the reduction of the number of train station inspectors. Since they were also responsible for checking the passing trains for fault states, the higher distances of unobserved movement of trains result in higher risks of heavy accidents. On the other hand, there were numerous developments of new sensor-systems for checking the mentioned train fault states. Unfortunately, all of these systems focus only on singular or at least on several particular deviations from normal and are completely stand-alone solutions. At higher passing speeds the accuracy and the reliability of these systems varies in a wide range. This is an important reason why today's fault-checking sensor systems are only used for alarming manned train guidance stations. Taking these trends into account, the demand for an overall train inspection system is obvious and resulted in a research project called "Checkpoint", which is funded by the Austrian government. Our approach is based on the acquisition of all relevant train properties to get knowledge of the complete train condition for better failure estimation. Thus, a concept for efficient integration of commercially available sensor systems and for universal data treatment was developed. It enables easy configuration of the data analysis and definition of fail-states by a rule based mechanism. The main technical innovation of the system comes from the direct connection to the interlocking system via the control system, which offers the ability to intervene automatically in case of a detected failure (e.g. stopping the train). To determine the reliability of new sensor systems under realistic conditions a prototype of a Checkpoint was built up and the most relevant sensor systems (hot box detector, flat wheel detector, dynamic scale and loading gauge detector) were integrated. Extensive testing would be performed to prove the functionality of the system.