Experimental results obtained at integral test facilities are used in the validation process of thermal - hydraulic system codes for the steady - state and transient simulation of light water reactors. The expertise and guidelines derived fro m this work (nota bly the nodalis ation scheme and the specific model options) can later be applied to safety analyses of nuclear power plants. This paper describes work carried out within the STARS project at the Paul Scherrer Institut (PSI) using the U.S. NRC - sponsored sys tem code TRACE. Post - test analyses for eight selected ROSA and ROSA - 2 experiments from the OECD/NEA projects ROSA and ROSA - 2 were conducted previ ously using TRACE version 5.0 release candidate (RC ) 3 . A consistent plant nodalisation is employed across all tests, which include typical design and beyond design basis accidents, such as loss - of - coolant - accident, main steam line break, and steam generator tube rupture. The full series of cases has been consolidated and upgraded for TRACE version 5.0 Patch 4. Sin ce the newer TRACE version includes substantial changes in the area of choked flow modelling, the nodalisation of the break valves has proven to be a critical upgrade in all test cases. Selected results for the newer TRACE versi on are presented and discuss ed. The complete set of tests has been supplemented by one further post - test analysis presented herein; The OECD/NEA ROSA - 2 Test 7, a PWR 13% cold leg intermediate break los s - of - coolant accident (IBLOCA). Results show that, for this break size , the influen ce of the uncertainty in the break flow rate is substantial. T here is a fine balance between the timing of the core uncovery and loop seal clearance, the rate of coolant loss following core uncovery, and the ability of the loop seal clearance to effectivel y quench the core. This has a significant effect on the predicted peak cladding temperatures. Updated results for ROSA - 2 a nalysis results for the remaining test cases are consistent with the previous RC3 analyses and match the experimental data well . U pdat ed results for ROSA - 2 Test 4, a steam generator tube rupture scenario, show improved agreement with experimental data . KEYWORDS TRACE, ROSA/LSTF, ITF, V&V