Two different types of perturbations of an SCWR-like fuel lattice have been investigated experimentally in the central test zone of the PROTEUS zero-power research reactor at the Paul Scherrer Institute in Switzerland. In each case, a campaign of high-resolution gamma-ray spectroscopy measurements was carried out on 34 fuel pins of the test lattice. In the first case, the test lattice was perturbed by inserting aluminum rods into the four central moderator regions, while in the second case, the perturbation was affected using steel absorber rods (instead of aluminum). The derived reaction rates are the capture rate in U-238 (C-8) and the total fission rate (F-tot), as also the reaction rate ratio C-8/F-tot. Each of these has been mapped on the lattice and compared to calculated results from whole-reactor Monte Carlo simulations with MCNPX. Excellent agreement has been obtained, for both perturbed lattices, between the calculated and experimental distributions of C-8, F-tot and C-8/F-tot. Considering that control rods in an SCWR assembly are foreseen to be inserted into the central moderator regions, these results may be considered as generic validation of Monte Carlo simulations for the two different types of lattice perturbations which inserted control rods imply, viz, moderator displacement and strong neutron absorption. In a second step, calculated C-8, F-tot and C-8/F-tot distributions for the two perturbed lattices (as well as for the unperturbed lattice) have been compared, at assembly level, between MCNPX and the deterministic LWR lattice code CASMO-4E. In the case of the unperturbed lattice, as well as for the lattice with steel rods, the agreement between the codes is found to be within similar to 1% for all pins and each reaction rate. However, for the lattice with aluminum rods, i.e. the case with mainly just moderator displacement involved, CASMO overestimates the reaction rates in the vicinity of the perturbations by up to 2-3%, when employing the standard input options. The reason for this discrepancy has been found to be the leakage treatment, which uses the fundamental-mode buckling applied in a homogenized sense across the lattice. In this way, global leakage gradients get averaged out over the entire assembly. The optional input card BZ2 for CASMO resolves this problem, and the codes then agree within 1% even for the aluminum case. (C) 2011 Elsevier Ltd. All rights reserved.