This three-credit-hour graduate course introduces detailed data analysis and model identification methods derived directly from experimental data. This knowledge-base is required to fuse sensed data with the reduced order physics based models to create online adaptive models. The model adaptations will be used to perform CPM by tracking the percent changes in the model coefficients. This tracking enables system fault detection, identification and estimation. The impact of this unique approach to CPM is that the adapted models enable real-time estimation of system parameters that have engineering significance. In particular, parameters such as structural stiffness, fluid properties including viscosity and bulk modulus, and individual energy conversion of multi-domain systems will be estimated to quantify the system health at the subsystem level. The outcomes from this course include an understanding of big data analytics and its application to CPM, and reduced order modeling for performance prediction that meets best industry practices and codes. The implementation of these strategies will realized using the Industrial Internet of Things.
SUBS 6397: Guide to Engineering Data Science