Skip to main content
The University of New Mexico

Testing and Demonstration of Remote and Autonomous Control Architecture for the Very-Small, Long-Life, Modular (VSLLIM) Microreactor

ISNPS is collaborating with a Department of Energy Nuclear Energy University Program (NEUP) project led by Professor Stylianos Chatzidakis at Purdue University on Demonstrating Autonomous Control, Remote Operation, and Human Factors for Microreactors. The objectives of this project at UNM-ISNPS are to develop and demonstrate remote and autonomous control of the Very-Small, Long-LIfe, Modular (VSLLIM) microreactor developed at UNM-ISNPS [read more about the VSLLIM microreactor here]. As part of this effort ISNPS researchers have trained Machine Learning (ML) algorithms for autonomous control of the VSLLIM microreactor. Investigated are both Supervised Learning algorithms with Long Short-Term Memory (SL-LSTM) neural networks and Soft-Actor-Critic algorithms with Feedforward Neural Networks (SAC-FNN) for reactor control of the VSLLIM microreactor.

The research evaluates trained ML algorithms while integrated into a software Programmable Logic Controller (PLC) and compares their performance for real-time control of the VSLLIM microreactor during the same simulated startup transients. The PLC with trained ML networks is then integrated within a developed secure, remote-control testbed incorporating a digital twin model of the VSLLIM microreactor developed using MATLAB Simulink software. The human operator has a virtual console that displays the operation variables generated by the VSLLIM digital twin on a large format screen using a Graphical User Interface (GUI) application. The operator monitors the operation state of the simulated reactor and sends commands to change the reactor operation. The developed remote-control platform secures communication between the remote operator terminal and the controllers of the VSLLIM reactor using two levels of data encryption. Testing for the developed remote-control testbed with PLC incorporating the trained SAC-FNNs algorithms evaluates its performance for controlling the VSLLIM digital twin model for a range of operation transients, including startup and changes in reactor power. The testbed is then demonstrated by allowing researchers at Purdue University to control the startup of the VSLLIM digital twin at UNM-ISNPS.

VSLLIM_Remote_Control_1

 

VSLLIM_Remote_Control_2

The results of this research can be found in the following progress reports submitted to US DOE and peer reviewed journal papers:

Technical Reports

  • Mohamed S. El-Genk, Timothy M. Schriener, Testing and Demonstration of Remote and Autonomous Control Architecture for the Very-Small, Long-Life, Modular (VSLLIM) Microreactor. Technical Report UNM-ISNPS-01-2026, Institute for Space and Nuclear Power Studies (ISNPS), University of New Mexico, (January 2026). [Download Available]
  • Mohamed S. El-Genk, Timothy M. Schriener, Ahmad N. Shaheen, Testing and Demonstration of Machine Learning Algorithms for Secure Remote Control and Autonomous Operation of the Very-Small, Long-Life, Modular (VSLLIM) Microreactor. Technical Report UNM-ISNPS-02-2025, Institute for Space and Nuclear Power Studies (ISNPS), University of New Mexico, (October 2025). [Download Available]
  • Mohamed S. El-Genk, Timothy M. Schriener, Ahmad N. Shaheen, Training Machine Learning Algorithms for the Controller of VSLLIM Microreactor using Artificial Neural Networks. Technical Report UNM-ISNPS-02-2024, Institute for Space and Nuclear Power Studies (ISNPS), University of New Mexico, (October 2024). [Download Available]
  • Mohamed S. El-Genk, Timothy M. Schriener, Ahmad N. Shaheen, Training and Testing Machine Learning Algorithms for VSLLIM Reactor Controller. Technical Report UNM-ISNPS-01-2024, Institute for Space and Nuclear Power Studies (ISNPS), University of New Mexico, (April 2024). [Download Available]

Refereed Publications