Big data analytics role in Maserati powertrain systems development. Case study: continuous improvement of IUMPR indexes.

The purpose of this document is to illustrate the internship experience I had at Maserati S.p.A. and to explain the results I obtained with my daily duties. As member of the Validation and Fleet Management team - a branch of the Powertrain Testing department of Maserati S.p.A. - I had access to the complete sets of data regarding every Validation fleet's vehicle accumulation schedule. Since those programmed activities imply the generation of huge amounts of data everyday, these ones can be gladly and fully-fledged addressed as Big Data. Pursuant of their extreme dimensions and thickness, these data are not manageable with traditional data analysis techniques, and require particular studies to exploit, give meaning and extrapolate usable indications to and from them. Big Data are surely a very precious and breaktroughing instrument, which - if properly used - can either definitely determine the success of a company with respect to competitors, or at least increase the quality and the profitability of its products in the market segments of interest. In the following Chapters, I have explained how Big Data are retrieved and stored within Maserati, and how this experimental philosophy is orienting the calibration cycle towards finer and faster results settlement. Furthermore, I have also described in detail the functionality and the usage of the developing software ETAS Moogle, which has been the main tool I have experienced and which is used as interface to manage those data and extrapolating usable, structured and synthetic indications. Eventually, I have described a practical application of how Big Data analysis can actually affect the improvement of vehicle performances, in the form of IUMPR indexes enhancements. These ones (the acronym "IUMPR" stands for In-Use Monitoring Performance Ratio, i.e. quantitative performance indicators of every specific emission-related variable that can be measured in a vehicle) are continuously-monitored quantities, whose value is real-time evaluated and stored by the engine electronic control unit for reporting purposes. Thanks to the remarkably high aggregating and didactic potential of this data analysis tool, it has been possible to identify a software miscalculation that returned unrealistic values of IUMPR - which could gave reliability problems against Authorithies inquiries - and to discern a way for its correction. The results of my study demonstrated the outstanding improvement possibilities that the analysis of Big Data implies, which could be a breaktrough attribute in nowadays automotive industry.