This research tackles the inefficiency of thermal food processing methods like drying, cooking, and pasteurization, which consume excessive energy and contribute to high carbon emissions. A 10% efficiency improvement in these processes could cut emissions by 1.1 million tons annually, equivalent to 2.7 billion miles driven by cars.

Challenge:

1. A highly complex process with inevitable process variations such as product switchover and moisture changes

2. A product switchover or a scale-up would render existing control and model parameters irrelevant.

3. Direct feedback of product quality is unavailable in most industrial settings.

The project aims to develop a modular, cyber-physical system (CPS) for real-time optimization of thermal processes. Using physics-based models and Reduced-Order Modeling (ROM) to reduce computational costs, the system will address variability in operations caused by measurable (e.g., equipment changes) and unmeasurable (e.g., moisture fluctuations) factors. By integrating indirect feedback and advanced simulations, the CPS will enable energy-efficient, scalable, and sustainable food production. Testing at MSU will validate its application to a wide range of food processing methods.

Solution:

1. Quick assembly of pre-built model components, enabled by a physics-based approach

2. Reduced-order modeling (ROM) of complex physics for computationally-efficient high-fidelity process simulations

3. ROM-building based on a new concept of altering the digital twin of the process for improved identification

4. Soft-sensing of immeasurable product quality