Un Chul Lee

Integral test and engineering analysis of coolant depletion during a large-break loss-of-coolant accident

This study concerns the development of an integrated calculation methodology with which to continually and consistently analyze the progression of an accident from the design-basis accident phase via core uncovery to the severe accident phase. The depletion rate of reactor coolant inventory was experimentally investigated after the safety injection failure during a large-break loss-of-coolant accident utilizing the Seoul National University Integral Test Facility (SNUF), which is scaled down to 1/6.4 in length and 1/178 in area from the APR1400 [Advanced Power Reactor 1400 MW(electric)]. The experimental results showed that the core coolant inventory decreased five times faster before than after the extinction of sweepout in the reactor downcomer, which is induced by the incoming steam from the intact cold legs. The sweepout occurred on top of the spillover from the downcomer region and expedited depletion of the core coolant inventory. The test result was simulated with the MAAP4 severe accident analysis code. The calculation results of the original MAAP4 deviated from the test data in terms of coolant inventory distribution in the test vessel. After the calculation algorithm of coolant level distribution was improved by including the subroutine of pseudo pressure buildup, which accounts for the differential pressure between the core and downcomer in MAAP4, the core melt progression was delayed by hundreds of seconds, and the code prediction was in reasonable agreement with the overall behavior of the SNUF experiment.

Application of Neural Networks to Analyze Load-Follow Operation in a Pressurized Water Reactor

A new analytic model based on hidden-layer neural networks is designed to analyze load-follow operation in a pressurized water reactor (PWR). The new model is mainly made up of four error backpropagation neural networks and procedures to calculate core parameters such as k∞ and xenon distributions in a transient core. The first two neural networks are designed to retrieve the power distribution, the third is for axial offset, and the fourth is for reactivity corresponding to a given core condition. The training data sets are generated by three-dimensional nodal code and the measured data of the first-day load-follow operation. The simulation results of the 5-day load-follow test in a PWR using the new analytic model show that it is an attractive tool for plant simulations in terms of accuracy, computing time, cost, and adaptability to measurements.

Comparative Study of Loss-of-Coolant Accident Using MAAP4.03 and RELAP5/MOD3.2.2

Analysis was performed for a large-break loss-of-coolant accident (LOCA) in the APR1400 (Advanced Power Reactor 1400 MWe) with the thermal-hydraulic analysis code RELAP5/ MOD3.2.2 and the severe accident analysis code MAAP4.03. The two codes predicted different sequences for essentially the same initiating condition. As for the break flow and the emergency core cooling (ECC) flow rates, MAAP4.03 predicted considerably higher values in the initial stage than RELAP5/ MOD3.2.2. It was considered that the differing break flow and ECC flow rates would cause the LOCA sequences to deviate from one another between the two codes. Hence, the break flow model in MAAP4.03 was modified with partly implementing the two-phase homogeneous critical flow model and adopting a correction term. The ECC flow model in MAAP4.03 was also varied by changing the hardwired friction factor through the sensitivity study. The modified break flow and ECC flow models yielded more consistent calculational results between RELAP5/MOD3.2.2 and MAAP4.03. It was, however, found that the resultant effect is rather limited unless more mechanistic treatments are done for the primary system in MAAP4.03.Copyright