Abstract: Experiment study of a 5×5 rod bundle was conducted to investigate the effect of blockage on sub-channel flow field and mass flow rate. A 3×3 block was set in a 5×5 rod bundle with the largest blocking percent of about 72% in the center sub-channel. The flow field from 2D upstream of the blockage center to 5D downstream was measured by particle image velocimetry （PIV） method. Moreover, the mass flow rate of sub-channel was obtained through a numerical integration using PIV results. At the same time, numerical method was used to simulate this phenomena. A self-developed 3D numerical analysis software called KILI V1.0 was used to simulate the blocking case. This software uses the immersed surface technology with Cartesian grid which allows directly using of 3D geometry from CAD software. 3D flow field was numerically calculated using NS equations with turbulence modeled by RANS method. Widely used commercial CFD software was also used for simulation to evaluate the uncertainty of the developed software. The obtained PIV results show that the blockage has an obvious effect on the flow field of the blockage downstream. The blocking leads to decrease of velocity in the blocked sub-channel and increase in the non-blockage sub-channel. For the tested 72% percent case, the blockage has a great influence within 0.5D downstream of the blockage. The average velocity within 0.5D downstream is only about 30% of the mean velocity, and the mass flow rate is only about 25% of the theoretical value. The velocity of the unblocking sub-channel increases over 50%, and the flow rate increases 30% from the theoretical value. The experiment result also shows that the effect of blocking decreases to 40% after 1D downstream and slowly recovers to normal flow. Back flow was observed in the center blocking sub-channel closed downstream area. Normalized sub-channel flow rate index was induced to quantify the flow rate diversion from theoretical value. The index shows that the blocking leads to over 1.6 times relative diversion in mass flow rate in the center blocking sub-channel within 0.5D downstream. The index decreases to 0.9 times in 1D downstream and 0.7 times in 4D indicating that the blocking mostly affects the 0.5D area downstream where should be carefully monitored. Immersed surface technology was used to simulate the flow field, the root mean squared error (RMSE) of simulation and experiment results shows that the self-developed software obtains a good simulated result. The RMSE of KILI V1.0 is 9% larger than commercial software when grid size is lower than 500 thousands, but 15% smaller when grid size increases to about 1 000 thousands. The obtained simulation results indicate that the self-developed software has a similar uncertainty with the commercial ones and even better under some conditions. The simulation shows that the back flow caused by the blockage is the main reason for velocity reduction.