[en] Post-test analyses using a multi-dimensional best-estimate analysis code, MARS, are performed for the APR1400 LBLOCA DVI (Direct Vessel Injection) performance tests. This report describes the code evaluation results for the test data of various void height tests and direct bypass tests that have been performed at MIDAS test facility. MIDAS is a scaled test facility of APR1400 with the objective of identifying multi-dimensional thermal-hydraulic phenomena in the downcomer during the reflood conditions of a large break LOCA. A modified linear scale ratio was applied in its construction and test conditions. The major thermal-hydraulic parameters such as ECC bypass fraction, steam condensation fraction, and temperature distributions in downcomer are compared and evaluated. The evaluation results of MARS code for the various test cases show that: (a) MARS code has an advanced modeling capability of well predicting major multi-dimensional thermal-hydraulic phenomena occurring in the downcomer, (b) MARS code under-predicts the steam condensation rates, which in turn causes to over-predict the ECC bypass rates. However, the trend of decrease in steam condensation rate and increase in ECC bypass rate in accordance with the increase in steam flow rate, and the calculation results of the ECC bypass rates under the EM analysis conditions generally agree with the test data