A dataset for thermal stratification phenomenon in lead-cooled fast reactors

This dataset focuses on the thermal stratification phenomenon in the upper chamber of a lead based fast reactor (LFR) after emergency shutdown. The system integrates experimental and numerical simulation data from 32 publicly published literature at home and abroad, including data from different reactors under typical accident conditions such as steady-state, transient, and unprotected loss of heat trap (ULOHS), protected loss of flow (PLOF), etc. The data was digitally extracted from literature curve graphs using Origin software, and its reliability was ensured through three independent extraction verifications. It was classified and stored using a structured metadata framework (including identification traceability, geometric space, physical conditions, and nine core fields of data content), with unified physical quantity units (temperature in Kelvin K, time in seconds, etc.) and standardized output in CSV format. The dataset covers key physical quantities such as vertical temperature distribution, mass flow rate, and thermal stratification interface position, solving the problems of traditional data dispersion and heterogeneous formats. It provides benchmark data for turbulence model calibration and system program optimization, while supporting machine learning based (such as LSTM, CNN-LSTM) rapid prediction research on thermal stratification, and assisting in thermal hydraulic safety analysis of lead based fast reactors.