High-precision manufacturing pushes the limits of what is possible with conventional manufacturing systems. The development of the new generation of high-precision manufacturing systems relies on detailed understanding of the process disturbances which cause variation in the end-product. This project aims at developing models for control of manufacturing processes which can be used for interpreting small variations in the process due to product-to-product disturbances. Such accurate and mature models can only be developed through integration of knowledge from high-fidelity physics-based models with knowledge obtained from large streams of sensor data. Integration of models and sensor data is key for the development of novel data analytic tools providing crucial causal information to the product-to-product variations which can subsequently be used by our novel data-driven control systems to pre-empt and to remove such variations in real-time.