송진호 교수

We propose a simulation and machine learning informed model (SMLIM) for the diagnosis of severe accidents. A machine learning model which consisted of one hidden Long Short Term Memory (LSTM) layer and two dense layers with variations in the number of neurons and regularization parameters and an Adams optimizer was constructed for the multi-time step ahead forecasting analysis and the regression analysis. Using feature variables of lower plenum liquid level, core liquid level, reactor vessel pre

A machine learning platform is proposed for the diagnosis of a severe accident progression in a nuclear power plant. To predict the key parameters for accident management including lost signals, a long short term memory (LSTM) network is proposed, where multiple accident scenarios are used for training. Training and test data were produced by MELCOR simulation of the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident at unit 3. Feature variables were selected among plant parameters, where th

Steam explosion experiments are performed at various modes of melt water interaction configuration using prototypic corium melt. The tests are performed to simulate both melt water interaction in a partially flooded cavity and melt water interaction in a cavity with submerged reactor. The tests are performed using zirconia and corium melts. The behavior of melt jet fragmentation during the flight in the air and fragmentation and mixing of melt jet in water is investigated by a high-speed video v

The present paper reports spontaneous steam explosions observed in fuel coolant interaction experiments using prototypic reactor materials. Pure ZrO and a mixture of UO and ZrO are used. A high temperature molten material in the form of a jet is poured into a subcooled water pool located in a pressure vessel. An induction skull melting technique is used for the melting of the reactor material. In both tests using pure ZrO and a mixture of UO and ZrO, either a quenching or a spontaneous steam exp

Investigations on the morphology and phase equilibrium characteristics of corium were performed by a series of experiments in parallel with thermodynamic phase equilibrium analyses. Melting and solidification experiments were performed using corium consists of U, Zr, ZrO2, SS, and B4C. TROI-49 and TROI-50 experiments with corium compositions representing Pressurized Water Reactor (PWR), whose compositions are similar to those of MA-3 and MA-4 of OECD MASCA (Material Scaling) Program while amount

Weaknesses of the current Severe Accident Management Guideline (SAMG) in handling the cooling of a molten core are discussed, and three improvements for the SAMG are presented. It is suggested that instrumentation to detect either a breach of the reactor vessel or a discharge of corium into the reactor cavity is essential to effectively perform the SAMG. A detailed analysis for a specific plant is necessary to make a decision as to whether preflooding or postflooding should be initiated for effe

The Fukushima accident is characterized by the fact that three reactors at the same site experienced reactor vessel failure and the accident resulted in significant radiological release to the environment, which was about 1/10 of the Chernobyl releases. The safe removal of fuel debris in the reactor vessel and Primary Containment Vessel (PCV) and treatment of huge amount of contaminated water are the major issues for the decommissioning in coming decades. Discussions on the new researches effort

A newly developed code CINEMA is exercised for simulations of severe accident progressions in OPR1000 nuclear power plant. A Large Break Loss of Coolant (LBLOCA) and Station Black Out (SBO) initiated severe accidents are selected to cover a wide range of accident progression in terms of the system pressure, amount of core damage, and fission product release. The effects of mitigation actions by the operator are also considered. In parallel, independent simulations by the MAAP5 are carried out an