PHM Approaches for Reliability of ECUs : Analyses of Canaries and Real-Time Data Acquisition

Abstract: Today, Scania CV AB is facing challenging demands on functionality and performance within their vehicles. The electronics are increasing rapidly and to stay competitive on the market, concerns regarding reliability of electronic systems needs to be evaluated. Prognostics and health management (PHM) is a concept where reliability of, for example, electronic control units (ECUs) are assessed. It requires customized systems for each specific environment, due to different strains and stresses. One approach is to have canaries (components with reduced soldering mass) implemented into ECUs as an indication that something is about to happen. Another essential aspect is the continuous real-time data acquisition from sensors, that can be used for different algorithms and models, which could provide forecasts on remaining useful life (RUL) of the ECUs. An appliance is the big data acquisition, where a database will collect data from vehicles, which means that ECU data need be communicated differently than today. This is why this master thesis project investigated canaries on printed circuit boards (PCBs) when they were subjected to vibrations as well as thermal cycling (TC), which are common parameters within vehicles. The PCBs consisted of both lead (Sn-Pb) and lead-free (SAC305) soldering for material comparison, since a transition to lead-free electronics is happening. Several fatigue tests were performed to collect as much information as possible. For further advances towards a PHM implementation, thermal shock (TS) tests on ECUs were performed as well. The ECUs were shocked in a temperature cabinet while being in active state. The internal temperature sensor was communicated with via controller area network (CAN). Scania’s CAN program was used and the data was logged in a computer, which in turn was compared to the data from the surrounding thermocouples, placed in specific spots. The results that were achieved clearly showed that lead-free PCBs are much more sensitive to stresses (both vibrations and TC). It was also shown that canaries failed in a much higher extent than regular resistors. The real-time data acquisition from the ECU could successfully be managed, where continuous data was logged. Also, the TS tests showed how the correlation between sensors indifferent positions was. Finally, the results from these tasks were discussed for future work. One have to keep in mind that this is just the beginning of a many-years project within Scania. The results and progress within this master thesis project will hopefully be a step in the rightdirection.