TY - GEN
T1 - Advanced control algorithms for liquid metal reactors
AU - Coble, Jamie
AU - Upadhyaya, Belle
AU - Briere, Chris
AU - Walker, Cody
AU - O'Connor, Austin
AU - Hines, Wes
AU - Ko, Yu Chih
AU - Ding, Jun
PY - 2015
Y1 - 2015
N2 - Sodium fast reactors (SFRs) present additional control challenges compared to the existing fleet of light water reactors. SFRs have additional temperature feedback effects, and temperature-caused structural expansion of core components can have significant effects. Additionally, the large sodium pool mass and the presence of intermediate heat exchangers result in time delays among major systems in the plant. The additional feedback effects, time delays, and the unique control targets of some SFR designs (i.e., constant core inlet temperature and constant temperature change across the core) suggest that traditional proportional-integral (PI) control may not be sufficient. In this research, a simplified model of the primary system of a prototypical SFR (reactor core and IHX) was developed in MATLAB-Simulink and validated against an existing high-fidelity simulator. Two competing control algorithms were applied to the Simulink model: traditional PI control and optimal model predictive control (MPC). The two algorithms were evaluated under a variety of reactor power levels (30-100% power) with external reactivity perturbations of -1 cent. In every case studied, the MPC outperformed PI control in both maximum deviation from set points and system stabilization time.
AB - Sodium fast reactors (SFRs) present additional control challenges compared to the existing fleet of light water reactors. SFRs have additional temperature feedback effects, and temperature-caused structural expansion of core components can have significant effects. Additionally, the large sodium pool mass and the presence of intermediate heat exchangers result in time delays among major systems in the plant. The additional feedback effects, time delays, and the unique control targets of some SFR designs (i.e., constant core inlet temperature and constant temperature change across the core) suggest that traditional proportional-integral (PI) control may not be sufficient. In this research, a simplified model of the primary system of a prototypical SFR (reactor core and IHX) was developed in MATLAB-Simulink and validated against an existing high-fidelity simulator. Two competing control algorithms were applied to the Simulink model: traditional PI control and optimal model predictive control (MPC). The two algorithms were evaluated under a variety of reactor power levels (30-100% power) with external reactivity perturbations of -1 cent. In every case studied, the MPC outperformed PI control in both maximum deviation from set points and system stabilization time.
KW - Dynamic nodal model
KW - Model predictive control
KW - Proportional-integral control
KW - Sodium fast reactor
UR - https://www.scopus.com/pages/publications/84946151569
M3 - Conference contribution
AN - SCOPUS:84946151569
T3 - 9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
SP - 1157
EP - 1165
BT - 9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
PB - American Nuclear Society
T2 - 9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
Y2 - 22 February 2015 through 26 February 2015
ER -