This Small Business Innovation Research (SBIR) Phase I project aims to address a difficult and widespread problem in analyzing repairs and maintenance data from the automotive sector: to rapidly identify the causes of system-level failures which are due to unexpected interactions among components that are part of a system rather than component-level malfunctions. The proposed approach is through the automated analysis or mining of maintenance datasets generated by technicians who perform the repairs. The proposed type of data analysis is likely to be the most effective and efficient solution to these common problems because data provides a far cheaper and centralized means to study the overall problem. Without a need to transport the physical components or having analysts travel to the repair locations, such data analyses may lead to physical root cause identification and solutions which ultimately result in engineering design modifications.
The proposed research and resulting outcome should impact not only the automotive sector, but also for several other transportation, manufacturing and service industries. This work will provide solutions wherever maintenance and repair reports are logged, thereby providing clues about problems that may arise at the level of systems rather than at their constituent components. The project will be an important case study of utilizing advanced mining algorithms, and be of immediate use in this important and extensive commercial sector. Beyond the pragmatic applications, this research may provide better insights into how components interact to form systems-level properties, and this may help with the notoriously difficult problem of modeling systems better. Also, there is also the possibility for more a efficient and cost-effective means to identify safety issues.
