Early probabilistic reliability analysis of mechatronic systems

Mechatronic systems are based on the close interaction of several engineering domains and are characterized by a high complexity; the assurance of their reliability is therefore very challenging. Established reliability methods are typically conducted on detailed system designs. As a consequence, a great number of failures are first recognized very late in the product engineering process. Yet: the later failures are discovered, the more does it cost to eliminate or mitigate them. Therefore there is a need for reliability analysis methods and tools, which are used from early on, so iteration loops in the system integration are avoided. In this contribution a method for the early probabilistic analysis of the reliability of an advanced mechatronic system based on its principal solution is presented; it addresses the aforementioned need. The method supports modeling of the failure propagation within the specification of the principal solution of an advanced mechatronic system, which is the result of the engineering phase of conceptual design. For analysis purposes, the model of the failure propagation is translated into a Bayesian network, which enables sophisticated probability analyses. Using our method the critical weak points of the system under consideration are identified and respective countermeasures can be defined; the principal solution is made more reliable. The advantages of the method are shown in a case study from the field of railway technology - the innovative autonomous railway vehicle RailCab.