The "DOLPHINS" Project: A Low-Cost Real-Time Multivariate Process Control From Large Sensor Arrays Providing Sparse Binary Data.

The "DOLPHINS" project started in 2018 from a collaboration settled by three partners: CNH Industrial Iveco (CHNi), RADA (an informatics company), and the Chemistry Department of the University of Turin. The project's main aim concerned the need to establish a predictive maintenance method in real-time at a pilot plant (CNHi Iveco in Brescia, Italy). This project currently allows maintenance technicians to intervene on machinery preventively, avoiding any breakdowns or stops in the production process. For this purpose, several predictive maintenance models were tested starting from databases (PLCs) already available, thus taking advantage of Machine Learning techniques and without having to invest additional resources in the purchase or installation of new sensors. The instrumentation and PLCs relating to the truck sides' paneling phase were considered at the beginning of the project. The instrumentation under evaluation was equipped with sensors already connected to PLCs (only on/off switches, i.e., neither analog sensors nor continuous measurements are available, and the data are in sparse binary format), so that the data provided by PLCs were acquired in a binary way before being processed by multivariate data analysis (MDA) models. Several MDA approaches were tested (e.g., PCA, PLS-DA, SVM, XGBoost, SIMCA) and validated in the plant (in terms of repeated double cross-validation strategies). The optimal approach currently used involves a combination of PCA and SIMCA models, whose performances are continuously monitored, and the various models are updated and tested weekly. Tuning of the time-range predictions enabled the shopfloor and the maintenance operators to achieve sensitivity and specificity values higher than 90%, but the performance results constantly improve since new data are collected every day. Furthermore, the information on where to carry out intervention is provided to the maintenance technicians between 30 minutes and 3 hours before the breakdown.