Problem with U235 concentration with burnup of the PWR fuel cell

Hello.
We solve the problem of high burnup of a PWR fuel cell (up to 100 MWd/kgHM) with an enrichment of 9.75% in U235. The description of the problem is given in the attached dissertation (Design Strategies for Optimizing High Burnup Fuel, page 73, Design strategies for optimizing high burnup fuel in pressurized water reactors)

We simulate burnup using OpenMC version 13, and then compare the results for U235 and fission product concentrations with the results from the included benchmark (CASMO-4 program was used). For some reason we are seeing significant differences in concentrations. For example, the difference in U235 concentration is 26% for a burnup of 100 MWd/kgHM.

Similar differences in U235 concentration we have for burnup credit benchmark PHASE-1-B (https://www.oecd-nea.org/science/wpncs/Publications/BUC/NSCDOC(96)06-buc-1b.pdf)

For calculations we use the ENDF-B-VIII.0 cross section library from the openmc.org website.
As a file with decay chains we use Simplified Chain (Thermal Spectrum) - chain_casl_pwr.xml
Can anyone explain the differences between two codes.

Our python code is here:
benchmark.py (3.4 KB)

The results are here:

Nuclide OpenMC CASMO-4 Differences, %
U234 6,57E-06 6,71E-06 -2,177
U235 0,00019172 2,60E-04 -26,12
U238 0,0195818 1,97E-02 -0,457
Pu238 2,33E-05 1,97E-05 18,731
Pu239 1,44E-04 1,48E-04 -2,284
Pu240 6,97E-05 6,31E-05 10,427
Pu241 4,30E-05 4,28E-05 0,4373
Pu242 3,07E-05 2,62E-05 17,069
Am241 1,98E-06 2,35E-06 -15,82
Am243 8,10E-06 6,23E-06 30,032
Np237 3,43E-05 3,42E-05 0,1826
Mo95 1,28E-04 1,22E-04 4,5730
Tc99 1,24E-04 1,17E-04 5,8799
Ru101 1,24E-04 1,19E-04 4,0212
Rh103 4,95E-05 4,60E-05 7,5535
Ag109 8,25E-06 6,99E-06 17,957
Cs133 1,26E-04 1,15E-04 9,7489
Sm147 1,03E-05 9,57E-06 7,3881
Sm149 1,03E-07 1,25E-07 -17,24
Sm150 2,92E-05 2,68E-05 9,1166
Sm151 5,90E-07 7,68E-07 -23,21
Sm152 7,97E-06 9,39E-06 -15,20
Nd143 7,16E-05 7,42E-05 -3,550
Nd145 7,32E-05 7,11E-05 2,9357
Eu153 1,04E-05 1,18E-05 -11,73

I guess it is very strange that you use 300K for all materials INCLUDING fuel. The reference mentions 900K…

When calculating burnup steps in days, the results were good.
But when we using burnup steps in MWt*day/kg HM, the results are wasted.