Heat Loss Simulation and Uncertainty Analysis in Fuel Bundle CHF and Other Large Scale Thermal-Hydraulic Experiments

摘要

Heat transfer experiments play a key role in exploring the underlining mechanisms of nuclear reactor thermal hydraulic phenomena. Large scale heat transfer tests, especially for rod bundle critical heat flux (CHF) experiments, generally utilize high power to achieve necessary operation conditions including wide ranges of temperature, flow, pressure, etc.. Temperature differences between test boundary, the boarding equipment, secondary flow housing, and the surrounding environment can be significant. Besides, heavy mass components, such as flanges, housing, and adaptor, can also serve as large heat sinks or heat sources leading to potentially large heat loss/gain and measurement uncertainty. Due to the complication of the installation and coupling of thermal-hydraulic conditions between inner and outer flow paths, direct measurement of heat loss/heat gain often becomes very difficult or nearly impossible. Furthermore, if there is a secondary layer of isolation fluid with potential natural circulation under various transient and steady operations, the total heat loss could have rather complicated dependency on operating conditions. When the total heating power is measured, reliable heat loss estimation is essential for accurate experimental measurements. As for a large scale test facility, different scales, experimental configurations, and operating conditions might lead to different heat loss. This paper presents a case example of heat loss/gain for a high pressure fuel assembly CHF test facility. In this CHF test loop, the test section consists of an inner flow path with an outer surrounding chamber filled with stagnant coolant that might present a significant source of heat loss/heat gain resulting from natural circulation in the surrounding chamber. System code RELAP 5 is used to simulate the test section and the surrounding systems under typical CHF test conditions. The simulated results are further compared to experimental data to verify the model's feasibility.