Dynamic Correlation Tool (DCT) for Digital Twin Development Platform of Long-Haul Vehicles

Digital Twin (DT) is a rapidly emerging and extensively research topic across engineering disciplines, aiming to seamlessly integrate physical systems with their digital counterparts. Central to the digital twin is the digital model – a virtual representation of the physical system, which is typically developed using equation-based or hybrid physics-data approaches. These models continuously track the operational status and internal parameters of the physical system in real time. To ensure accuracy, digital models regularly compare real field data (real data) with model outputs (model data) and iteratively updating themselves for parameter tracking of physical system. Traditionally, simple cost functions like mean-square error have been employed to quantitatively compare real and model data. However, as DT development grows in complexity, more advanced correlation techniques become necessary due to challenges such as noisy real data, differences in units, time scales, and sampling frequencies, and misalignments in the time of both real data and model data. To address these challenges a ‘Dynamic Correlation Tool’ is developed. This tool features multi-source input support, comprehensive data processing and standardization, missing data handling, data synchronization algorithms, multiple correlation matrix options, and visualization capabilities. It is deployed on a web-server with API access to seamlessly integrate into digital twin development workflow. The tool is develop in the context of DT development platform for long-haul electric and hybrid vehicles. The poster will showcase its features and demonstrate them on a real-field dataset for DT development.