Digital Twin Calibration with Big data

The advancements in numerical algorithms and computational power have elevated the role of digital twins in design and analysis of many systems. The development of digital twins often hinges on the application of physics-based first principles and necessitates the accurate designation of numerous parameters. However, certain scenarios prevent the accurate determination of these parameters via physical laws, prompting domain specialists to resort to educated guesses. These assumptions, though well-intended, can lead to significant discrepancies between a digital twin’s outputs and the actual system’s performance. To ensure the digital twins closely mimic their real-world counterparts, this research capitalizes on the power of Big Data. Concurrently, it caters to the research challenges inferred from the sheer volume and complexity of the datasets. This research lays a strong foundation for an efficient and resilient parameter calibration process by establishing a quantitative scheme by harmoniously integrating data science theories and cutting-edge tools within the data-driven optimization framework, thereby offering both practical and theoretical implications.