Enabling Digital Twin Using MEMS and Sensor Technology with Historic Data: From Data Fusion to Knowledge Fusion
Haiyan Sally Xie, Owen Shi, Sai Ram Gandla, Mangolika Bhattacharya · 2024
Microelectromechanical systems and Sensor Technology (MEMS-ST) can be used together with historical data to enable digital twins. This paper presents a novel framework that transitions from the data fusion of MEMS-ST to the knowledge fusion necessary for estimating the strength capacity of highway bridges. This innovative approach provides critical information for bridge design, transportation planning, and the assessment of existing pavement and infrastructure. The unique framework integrates real-time data from MEMS-ST systems installed on a bridge with historical records and engineering standards, addressing a significant gap in current methodologies. Through a comprehensive literature review, this study evaluates the existing reports and articles on developing bridge digital twins to measure live loads on highways. Utilizing datasets from sensors and the National Bridge Inventory, the proposed framework processes data based on the Load and Resistance Factor Design method and continuous multivariate probability distributions. The innovative application of the Dempster-Shafer Theory within the knowledge fusion framework distinguishes this work, enabling the combination of different sources (86, 126% increase) of evidence into a cohesive global belief. The results demonstrate that this framework can automate the analysis of vast amounts of data, improve efficiency (615,000 bridges in NBI and 93.1M truck data), and provide more reliable insights compared to traditional methods. Specifically, it enhances the safety and stability of bridge structures (Credibility 1.088; Plausibility 1.089) while offering new ways to design and maintain bridges more effectively. This paper sheds light on bridge digital twin systems.