CX-020790: Demonstrating Reactor Autonomous Control Framework using Graphite Exponential Pile – Massachusetts Institute of Technology

The Massachusetts Institute of Technology (MIT) proposes to develop a detection-prediction-feedback (DPF) framework for autonomous control of a nuclear system under normal as well as off-normal (accident) conditions. The tasks associated with this project are (1) Implementation of graphite exponential pile (GEP) control mechanism which includes design of the control rod actuators as well as safety evaluation and administrative approval; (2) Development of high-fidelity real-time prediction utilizing a machine-learning surrogate model that is refined and improved as measurement data is produced; (3) Development of the DPF framework including accounting for reactor kinetics for real-time prediction and system control feedback; and (4) Integrated experiments for demonstrating autonomous control. An existing GEP at MIT will be used.