Skip to main content
Nuclear Science and Engineering

AI/ML Power Conversion System

Model-based control provides a means to address the large nonlinearities found in some advanced power conversion cycles

A physics-based model was recently used to provide a high-precision prediction of the cold side temperature leaving the precooler of a supercritical carbon dioxide power conversion cycle. The accuracy required to enable temperature control near the critical point exceeds that achievable with a resistance temperature detector (RTD). This temperature must be precisely controlled if maximum cycle efficiency is to be obtained.


 A schematic for the Kalman filer state observer used to predict this temperature is shown.


The successful deployment of digital twin models and advanced control algorithms for the real time control of power conversion systems having significant nonlinearities and stability requirements as found, for example, in some Brayton cycle implementations.