The design of an advanced heavy water reactor (AHWR) utilizing thorium is in an advanced stage. The AHWR is a boiling light water cooled, heavy water moderated pressure tube-type reactor, which derives most of its power from the thorium-uranium cycle with plutonium as the external fissile feed. The AHWR has several passive safety features, notably the negative coolant void coefficient of reactivity during both operational and transient conditions. It is of utmost importance to understand the mechanism of the coolant void reactivity (CVR), i.e., the effect of boiling on the neutron spectrum and hence the relative absorption in the different isotopes. We have performed a detailed reaction rate analysis for the isotopes in the AHWR lattice and estimated the individual components to the CVR. The AHWR fuel cluster is a heterogeneous one with both (Th,U) mixed oxide (MOX) and (Th,Pu) MOX fuel and also stainless steel as absorber in the central displacer region.

The individual contributions of the different isotopes and reactions were calculated for three major energy domains - namely, fast, resonance, and thermal - as well as for an effective energy average (one-group). The general trend of the CVR with burnup is dictated by the relative absorptions. The major contributors to the CVR were hydrogen (in the coolant), 232Th, 233U, and 239Pu: 232Th and 233U exhibit a negative contribution, whereas 239Pu and H show a positive contribution to CVR. The net absorption reaction rate in 233U becomes less negative with burnup. Since it is close to the moderator, plutonium sees a more thermal spectrum and depletes faster. The positive contribution from 239Pu decreases with burnup. At higher burnups the relative absorption upon voiding in hydrogen increases, and this is a major contributor to the CVR becoming less negative.

The results were compared for different nuclear data sets, ENDF/B-VI.8 (largely used for our design studies) and the newly available ENDF/B-VII.0. The CVR calculated with the ENDF/B-VII.0 showed significant differences at higher burnups. The ENDF/B-VII.0 data set gave lower negative value for the CVR at end of cycle. It was found that the difference in the capture cross section of the 232Th ENDF/B-VII.0 data set was largely responsible for the difference between the two data sets. All the simulations were done using the WIMSD code and the multigroup WIMS library using a 69-group structure.