Digital Tire Factory: Trends and Outliers Identification using Mahalanobis Distance

In tire manufacturing, efficient monitoring systems are needed to detect trends and anomalies in multivariate data, enhancing product consistency and operational efficiency. This paper discusses a potential way for this monitoring through a multivariate approach by comparing Mahalanobis’ distance measure towards Euclidean one in an industrial context: starting with detecting production drifts, the analysis aims to improve trend and outlier detection responsiveness. Results suggest the Mahalanobis’ distance effectively identifies trends, improves consistency, and prevents anomalies faster than current methods, promising enhanced quality control and operational efficiency in view of a for fully automated monitoring via a digital twin system.