Dr. John R. White, Justin Byard, and Ronald D. Tooker
Chemical and Nuclear Engineering Department
University of Massachusetts Lowell
June 25, 1999

Several tasks associated with this project were completed during June 1999, as follows:

1. Updated several of the computational models and techniques to refine and improve our analysis capability.
2. Made final selection of a new reference LEU configuration that retains the regulating blade in the D9 location.
3. Prepared a summary for the 1999 ANS Winter Meeting that overviews the new 3-D VENTURE and MCNP models that have been generated for the LEU core. This summary also briefly describes the development of the new proposed startup core (see Item 2).

These tasks are described in further detail in the remainder of this progress report.

As indicated in the May 1999 Progress Report, a minor error was found in the 3-D VENTURE models used thus far in this project -- a radiation basket was located in grid position G5 instead of a graphite reflector. This error has now been corrected in all the 3-D VENTURE models.

In addition, two more edit zones were added to the 3-D model (Zones 948 and 949). These zones are located in Layer 7 just above the central beam ports on both sides of the core. These zones are filled with water reflector material and the zone average fluxes in these two regions represent approximations to the environment seen by the rabbit irradiation facility.

Another minor change in the 3-D model involved replacing the 20% water composition in the beam ports with a material that represents only 5% full water density. The beam ports are normally air-filled, but diffusion theory codes have numerical difficulty with voided regions within the models. Therefore, low-density water is used to represent a near-void region. Originally, the 3-D model used 20% water density, which was a direct carryover from the 2-D XY models. The 20% value was obtained as a very rough approximation to the area covered by the beam ports when modeling a 3-D structure in 2-D geometry. However, in 3-D geometry, the area fractions are treated automatically as part of the model, and the water density in the beam ports was reduced accordingly. This very rough approximation represents a reasonable fix for a deficiency associated with the diffusion theory approximation. This change affected the material compositions associated with Materials # 28 and #30 in the 3-D VENTURE models.

2-D VENTURE Models

A minor change in the 2-D VENTURE models, involving two additional edit regions in the graphite thermal column, was made to match the 3-D VENTURE models. This modification does not affect previous results, but it now allows one to make direct comparisons of some representative thermal column fluxes between the 2-D and 3-D models.

As reported earlier, a Matlab-based plotting tool, called PLOT_VGEO.M, was generated to help visualize and debug the geometry and material maps for the VENTURE 2-D and 3-D models. The code allows one to plot a combination zone and material map for any XY, XZ, or YZ plane within the VENTURE model, and the code has really proved to be an invaluable debugging aid. However, a somewhat irritating quirk in the code was the reversal of the "y-axis" when plotting -- with the "y-direction" always increasing from bottom to top. The problem, of course, is that all our VENTURE models have the "y-axis" increasing from top to bottom. We have worked around this reversal problem by simply recognizing its existence and dealing with it accordingly -- however, it was very annoying and many of the resultant plots were misleading. This problem has now been resolved with a relatively simple fix inside the code, and a new version of PLOT_VGEO in now in use. All the zone and material maps are now fully consistent with the actual VENTURE geometry, and any maps generated in the future will use this new capability. However, all our old plots will now be inconsistent with any new ones generated after about mid-June 1999!

Last month as series of 2-D and 3-D cases was made to help identify promising configurations that retain the regulating rod in its current D9 location, and selected reactivity results from several candidate core configurations were presented in Table II of the May 1999 Progress Report. In particular, a 21-element core with 19 full fuel elements and 2 partial fuel assemblies in C5 and E5 (referred to as the leu213d/leu313d models) was chosen as a very promising candidate configuration. This choice, however, was only preliminary because full analysis of all the operational parameters of interest (blade worth distributions, shutdown margins, experimental fluxes, etc.) were not available. These analyses have now been completed and a final selection of the leu213d/leu313d configuration for the startup LEU core layout has been made. Several important operational parameters for this configuration relative to the model proposed in the FSAR Supplement¹ are tabulated in Tables I and II. These tables summarize some reactivity information and the computed fast and thermal average fluxes in several experimental locations. Note that, for convenience in naming several branch cases, the reference core layout previously referred to as the leu213d/leu313d model has now been renamed to leu214/leu314.

------------------------

Note: Another minor change to the new reference models has been made, and the latest new configuration is referred to as the leu215/leu315 configuration. Both the radiation basket and flux trap facilities have provision for inserting a bayonet that encloses the experimental sample to be irradiated. Thus, flexibility has been included in both the 2-D and 3-D VENTURE models to allow insertion of an air-filled bayonet into these regions. The leu214/leu314 models have the air-filled bayonets inserted into the baskets in the C2, D2, and E2 positions, but the flux trap is modeled with the bayonet out of the facility, with the central aluminum guide tube filled with water. We have decided that the reference model should be free of any experimental bayonets, so the homogenized radiation basket composition that includes the air-filled bayonets has been replaced with the water-filled basket material. This new model will be referred to as the leu215/leu315 configuration. This material configuration change is a relatively minor modification that only has a +0.05 base reactivity effect and negligible impact on the fluxes outside the radiation baskets. However, it should be noted that all the results given in Tables I and II are for the leu214/leu314 layout, and that minor differences in subsequent reports that use the leu215/leu315 models may occur.

------------------------

Several observations can be made from the data in Tables I and II. Focusing first on the reactivity data, we see that the initial excess reactivity and regulating blade worth for the new reference core are within acceptable limits. In addition, although more asymmetrical than the FSAR model, the blade worth distribution for the new reference core is well within operational constraints. In fact, this distribution, with Blades 1 and 2 having less worth than Blades 3 and 4, is quite consistent with the distribution observed in the current HEU core. The shutdown margin, computed with the most reactive blade in its out position and a maximum experimental worth of +0.5 , is also well above the technical specification limit of 2.7 . Because the reactivity worths computed from the 2-D and 3-D models are so close (in all cases tested so far), only part of the 3-D reactivity worth data were actually computed. Note that, because of the incomplete 3-D data, the shutdown margin for the 3-D cases was computed using the blade worth data from the 2-D models (which is expected to be a very good estimate). Note also that there is a definite bias in the 2-D versus 3-D beginning-of-life (BOL) excess reactivity. This is due to the transverse buckling approximation that is required in the 2-D models (see the May 1999 Progress Report for a discussion of this issue).

It should be emphasized that the tabulated results for the VENTURE-calculated regulating blade worths have already been reduced by an expected uncertainty of roughly +20%. This bias is the result of the modeling approximations used when the VENTURE models were first developed (i.e. spatial homogenization of the small regulating blade elements into somewhat larger spatial regions was required to maintain a reasonably consistent mesh spacing throughout the model). We have decided to report the regulating blade worths in this fashion because these estimated worths are expected to be much closer to the actual values that will be observed in the physical system. In any case, even with this estimated uncertainty applied to the calculated results, the predicted D9 blade worth for the new reference core satisfies the imposed design constraint of 0.35 .

The experimental fluxes shown in Table II come from the 3-D VENTURE models. Except for the last two rabbit zones, all the tabulated data were taken from the axial midplane in the all-blades-out configuration (i.e. Layer 9 average fluxes in the reference leu311 and leu314 models). Thus, these data represent near maximum values. Actual values in a bladed configuration may be somewhat different. The rabbit facilities are located just above the central 8-inch beam ports and this coincides with Layer 7 in the 3-D model. The fast and thermal flux data come directly from the 2-group energy model, where 1.0 eV represents the boundary between the fast and thermal groups -- multigroup calculations are required to get better energy resolution than this.

The first point to emphasize here is that these data are not directly comparable to the flux data in the FSAR Supplement¹ or from Ref. 2. Clearly the data here are from the 3-D model, whereas most of the data in Refs. 1 and 2 are from various 2-D models that assume essentially constant profiles in the untreated direction. Thus, for example, in a 2-D XY model the fluxes represent an axial average. Even without the 2-D versus 3-D difference, however, the base models are so different that direct comparison of selected tabulated results is nearly impossible. For example, the models in Refs. 1 and 2 did not include the beam ports and the radiation baskets included only a single homogeneous zone within each grid location containing a radiation basket. The current models, on the other hand, have a rough beam port model included with a 5% water composition to represent the nearly voided tubes, and the radiation baskets have a more-detailed 2-region representation. Thus, direct comparison of the computed radiation basket and beam port fluxes with previous work is not practical. Many of the results of previous models were used primarily for relative comparisons among different configurations using self-consistent models. In contrast, the current 3-D VENTURE results represent a reasonable estimate of the fluxes that are actually expected in the individual experimental locations.

Focusing now on the old versus new LEU core configurations (the leu311 case versus the leu314 model), we see that there is a definite shift in the fluxes towards the thermal column side of the core for the new leu314 configuration. This is due primarily to the additional full fuel assembly in the D8 location for the leu314 layout relative to the leu311 model. This shift is apparent in the slight decrease in the fluxes in the radiation baskets and the left beam port zone and the slight increase in the average fluxes in the right beam port. Even more dramatic and interesting is the relatively large increase in the expected thermal column fluxes. This latter observation represents a major plus for the new core configuration relative to the FSAR model, considering that all the other experimental fluxes only increased or decreased by roughly 10%. Thus, the 50% or more increase in the thermal column fluxes is an important feature of the new LEU core layout.

Table II Zone average fluxes (n/cm²-s) for the old and new LEU configurations.

One final note that is true for both LEU configurations is that the central flux trap has fluxes that are about a factor of two or more greater than those observed in the traditional radiation basket zones on the edge of the core. This feature is an added advantage of all the LEU core configurations. This enhancement was a key element of the original design studies^1-2 and it has been retained in the current designs.

The bottom line of the above discussion is that the new leu214/leu314 configuration satisfies all the design constraints imposed upon the system, including the retention of the regulating blade in the D9 grid location. The regulating blade worth has decreased somewhat, but is appears to be large enough for convenient day-to-day operation of the facility. The radiation basket fluxes have been degraded a little, but the experimental fluxes in the thermal column and the right beam ports have been enhanced. Thus, we have chosen this core as the best candidate for the startup of the LEU-fueled UMass-Lowell research reactor (UMLRR). This 21-element core layout (with 19 full assemblies and 2 partial assemblies) should be the goal of the initial core loading sequence during the startup tests beginning in Fall 1999.

The final activity to report upon for this progress report is the submission of a short summary for the upcoming 1999 ANS Winter Meeting. The paper, "Modeling and Reference Core Calculations for the LEU-Fueled UMass-Lowell Research Reactor," by J. R. White and R. D. Tooker briefly overviews the calculations and results of the design study just discussed under Task 2 in this Progress Report. In addition, it briefly discusses the current computational models used in the analyses, including a detailed MCNP model of the LEU-fueled UMLRR. The MCNP model has been developed primarily by R. D. Tooker in parallel with the VENTURE work over the last several months. The current MCNP model is still under development, but preliminary criticality and reactivity worth results are in good agreement with the 3-D VENTURE results, including our modified estimate of the VENTURE-calculated regulating blade worths. However, only a cursory review of the full MCNP model has been completed to date, and further work on the base MCNP model and on verifying its consistency with the current VENTURE capability will continue over the next few months. We expect the MCNP model to be a significant addition to our overall computational capability for the new LEU core -- since it offers state-of-the-art capability not available with the traditional deterministic codes. A variety of models within the VENTURE, DORT, and MCNP codes should give us sufficient capability to address just about any aspect of interest (the development of a consistent 2-D model within DORT is planned for next month). The availability of a variety of computational models provides added flexibility for the current and future support of the new LEU-fueled facility.

A copy of the ANS Summary is appended to this report. It contains our first preliminary comparison of the VENTURE and MCNP models. More comparisons (along with some DORT simulation results) will continue over the next few months as all the computational models evolve. A complete and consistent set of tools should be available before startup of the new core in early Fall 1999. Final verification of the models will occur once startup data are available.

1. "FSAR Supplement for Conversion to Low Enrichment Uranium (LEU) Fuel," Document submitted for review by the NRC for conversion of the UMass-Lowell Research Reactor (May 1993).
2. R. S. Freeman, "Neutronic Analysis for the Conversion of the ULR from High Enriched Uranium to Low Enriched Uranium Fuel," MS Thesis, Nuclear Engineering, University of Massachusetts Lowell (1991).

Click here to view some material maps for the new reference LEU core (leu314 model).

Back