ML23040A303

Univ of Texas - Austin, Submittal of Neutronics Report and Thermal Hydraulic Analysis of the Triga Reactor

ML23040A303
Person / Time
Site:	University of Texas at Austin
Issue date:	02/09/2023
From:	Charlton W University of Texas at Austin
To:	Office of Nuclear Reactor Regulation, Document Control Desk
References
Download: ML23040A303 (1)

Text

WALKER DEPARTMENT OF MECHANICAL ENGINEERING Nuclear Engineering Teaching Laboratory Pickle Research Campus R-9000

• Austin, Texas 78758

• 512-232-5380

• FAX 512-471-4589 nuclear. engr. utexas. edu

• wcharlton@austin. utexas. edu February 9, 2023 ATTN: Document Control Desk U. S. Nuclear Regulatory Commission 11555 Rockville Pike Rockville, Maryland, 20852 Geoffrey Wertz, P.E.

Non-Power Production and Utilization Facility Licensing Branch (UNPL)

Division of Advanced Reactors and Non-Power Production and Utilization Facilities (DANU)

Office of Nuclear Reactor Regulation

SUBJECT:

Docket No. 50-602, Facility Operating License R-129 - Submission of Neutronics Report and Thermal Hydraulic Analysis of the University of Texas (UT) TRIGA Reactor

REFERENCE:

July 23, 2020 letter: University- of Texas at Austin - Regulatory Audit re: Renewal of Facility Operating License No. -129 (EPID No. L-2017-RNM-0032), (Agencywide Documents Access and Management System (ADAMS) Accession No. ML20203M166),

Sir:

We respectfully submit the Analysis of the Neutronic Behavior of the Nuclear Engineering Teaching Laboratory Reactor at the University of Texas and the Thermal Hydraulic Analysis of the University of Texas {UT) TR/GA Reactor in support of license renewal for the University of Texas at Austin as attached. If you have any questions, please contact me at 512-232-5373 or whaley@ mail.utexas.edu.

P. M. Whaley I declare under penalty of perjury that the foregoing is true and correct.

W. S. Charlton ATT:

(1) Analysis of the Neutronic Behavior of the Nuclear Engineering Teaching Laboratory at the University of Texas (2) Thermal Hydraulic Analysis of the University of Texas (UT) TRI GA Reactor

ANALYSIS OF THE NEUTRONIC BEHAVIOR OF THE NUCLEAR ENGINEERING TEACHING LABORATORY REACTOR AT THE UNIVERSITY OF TEXAS Submitted By:

Radiation Center Oregon State University Corvallis, Oregon February 2023 NETL Neutronic Analysis 1 Feb 2023

Table of Contents

1. Introduction ......................................................................................................................... 4

2. Summary and Conclusions of Principal Safety Considerations ......................................... 4

3. Reactor Fuel ........................................................................................................................ 4

4. Reactor Core ....................................................................................................................... 6

5. Model Bias .......................................................................................................................... 7

6. Burnup Calculations.......................................................................................................... 11

7. Current Core Configuration .............................................................................................. 13

8. Limiting Core Configuration ............................................................................................ 19

9. Summary ........................................................................................................................... 24 NETL Neutronic Analysis 2 Feb 2023

List of Figures Figure 1 - TRIGA Stainless Steel Clad Fuel Element Design used in the NETL Core ................5 Figure 2 - Schematic Illustration of the NETL Showing the Current Core Configuration .............6 Figure 3 - Horizontal and Vertical Cross-sections of the NETL MCNP Model at BOL ................7 Figure 4 - Reactivity (including bias) of 80 Different BOL Critical Core Configurations .............8 Figure 5 - Reactivity (including bias) of 36 Different BOL Critical Core Configurations .............9 Figure 6 - Vertical Cross-section of Current Core Configuration MCNP Model .........................13 Figure 7 - Current Core Power-Per-Element (in kW) Distribution at 1.1 MW ............................13 Figure 8 - Current Core Configuration Prompt Temperature Coefficient, F, as a Function of Temperature .................................................................................................................16 Figure 9 - Current Core Configuration Moderator Void Coefficient ............................................17 Figure 10 - Current Core Configuration Moderator Temperature Coefficient .............................17 Figure 11 - Vertical Cross-section of Limiting Core Configuration MCNP Model .....................19 Figure 12 - Limiting Core Configuration Power-Per-Element Distribution at 1.1 MW ...............20 Figure 13 - Limiting Core Configuration Prompt Temperature Coefficient, F, as a Function of Temperature .................................................................................................................22 Figure 14 - Limiting Core Configuration Moderator Void Coefficient ........................................22 Figure 15 - Limiting Core Configuration Moderator Temperature Coefficient............................23 List of Tables Table 1 - Characteristics of Stainless Steel Clad Fuel Elements.....................................................5 Table 2 - BOL Rod Worth Calculations........................................................................................10 Table 3 - Summary of Burnup Steps .............................................................................................12 Table 4 - eff and Prompt Neutron Lifetimes for Current Core Configuration .............................14 Table 5 - Current Core Rod Worth Calculations...........................................................................15 Table 6 - Current Core Configuration Prompt Temperature Coefficient ......................................16 Table 7 - K-Effective Calculations Used to Determine Current Core Power Defect ...................18 Table 8 - eff and Prompt Neutron Lifetimes for Limiting Core Configuration............................20 Table 9 - Limiting Core Configuration Rod Worth Calculations .................................................21 Table 10 - Limiting Core Configuration Prompt Temperature Coefficient ..................................22 Table 11 - K-Effective Calculations Used to Determine Limiting Core Power Defect ................23 Table 12 - Limiting Core Hot Channel Power Summary..............................................................24 NETL Neutronic Analysis 3 Feb 2023

1. Introduction This report contains the results of investigation into the neutronic behavior of the Nuclear Engineering Teaching Laboratory reactor (NETL) at the University of Texas Austin. The objectives of this study were to: 1) create a model of the NETL to study the neutronic characteristics, and 2) demonstrate acceptable reactor performance and safety margins for the NETL core under normal conditions.

2. Summary and Conclusions of Principal Safety Considerations The conclusion of this investigation is that the MCNP model does an acceptable job of predicting behavior of the NETL core. As such, the results suggest that the current NETL core can be safely operated within the parameters set forth in the technical specifications. Discussion and specifics of the analysis are located in the following sections. The final sections of this analysis provide suggestions for a limiting core configuration.

3. Reactor Fuel The fuel utilized in the NETL is standard TRIGA fuel manufactured by General Atomics. The use of low-enriched uranium/zirconium hydride fuels in TRIGA reactors has been previously addressed in NUREG-1282 [1]. This document reviews the characteristics such as size, shape, material composition, dissociation pressure, hydrogen migration, hydrogen retention, density, thermal conductivity, volumetric specific heat, chemical reactivity, irradiation effects, prompt-temperature coefficient of reactivity and fission product retention. The conclusion of NUREG-1282 is that TRIGA fuel, including the fuel utilized in the NETL, is acceptable for use in reactors designed for such fuel.

The design of standard stainless steel clad fuel utilized in the NETL is shown in Figure 1. Stainless steel clad elements used at NETL all have fuel alloy length of 38.1 cm. The characteristics of standard fuel elements are shown in Table 1.

NETL Neutronic Analysis 4 Feb 2023

Figure 1 - TRIGA Stainless Steel Clad Fuel Element Design used in the NETL Core Table 1 - Characteristics of Stainless Steel Clad Fuel Elements Uranium content [mass %] 8.5 235 BOL U enrichment [mass % U] 19.75 Original uranium mass [gm] 37 Zirconium rod diameter [in] 0.25 Fuel meat outer diameter [in] 1.435 Cladding outer diameter [in] 1.475 Cladding material Type 304 SS Cladding thickness [in] 0.020 Fuel meat length [in] 15 Graphite slug outer diameter [in] 1.43 Upper graphite slug length [in] 2.6 Lower graphite slug length [in] 3.7 Molybdenum disc thickness [mm] 0.8 The NETL reactor initially achieved criticality in March of 1992, however all fuel (except for the fresh FFCRs) was previously used at other facilities. Most of it came from a previous reactor on campus at Taylor Hall, but there were other sources as well. This made the beginning-of-life (BOL) fuel isotopic determination difficult. UT Austin performed a SCALE analysis to burn the fuel in conjunction with the given burnup records. The SCALE outputs were used to create BOL fuel isotopics for the MCNP runs. However, the burnup records did not specify core location during previous irradiation, so these SCALE isotopics are a best guess given the previous information.

NETL Neutronic Analysis 5 Feb 2023

4. Reactor Core The NETL core is a seven-ringed hexagonal grid array (labeled A through G) with 121 positions mostly composed of stainless-steel-clad standard TRIGA fuel elements. The current core configuration contains 113 fuel elements (including three fuel-followed control rods, i.e. FFCRs).

The core also contains an air-followed transient rod in C-1, a central thimble in A-1, several non-fueled locations that allow for a larger irradiation facility (in positions E-11, F-13 and F-14), a startup source in G-32, and a pneumatic transfer (Rabbit) irradiation facility in G-34, and an empty position G-26. The reactor is controlled by three electromagnetic control rods (Shim I, located in D-6; Shim II, located in D-14; and Regulating, located in C-7) and a pneumatic air-followed control rod (Transient, located in C-1), which utilize borated graphite (B4C) as a neutron poison.

Fuel temperature is measured by an instrumented fuel element (IFE) located in B-3. The current core configuration is shown in Figure 2.

Figure 2 - Schematic Illustration of the NETL Showing the Current Core Configuration Detailed neutronic analyses of the NETL core were undertaken using MCNP6.2 [2]. MCNP6.2 is a general purpose Monte Carlo transport code which permits detailed neutronic calculations of complex 3-dimensional systems. It is well suited to explicitly handle the material and geometric heterogeneities present in the NETL core. The original input deck for the NETL model was developed at UT Austin and modified by Oregon State University. Facility drawings provided by the manufacturer at the time of construction of the facility were used to define the geometry of the core and surrounding structures. The geometry of the stainless steel clad fuel elements and control NETL Neutronic Analysis 6 Feb 2023

rods were based upon the manufacturing drawings. Representative cross-sectional views of the MCNP model (of the initial core loading) are shown in Figure 3.

Figure 3 - Horizontal and Vertical Cross-sections of the NETL MCNP Model at BOL

5. Model Bias Beginning-of-life Criticality Bias Using critical rod height data from the first few months of NETL operation, a series of MCNP analyses based upon various critical rod heights were performed to determine the criticality bias of the model. This bias represents such things as differences in material properties that are difficult to determine or unknown (i.e., exact composition of individual fuel meats and trace elements contained therein) or applicability of cross section data sets used to model the reactor (i.e.,

interpolation between temperatures). As a result, the validation of the model was based upon the ability of the code to accurately predict criticality as compared with measurements made on the reactor in early 1992.

A criticality calculation was performed using cold clean critical core configuration information from 3/23/1992, which was the first time the NETL was taken to criticality. The k-effective of this configuration was 0.99393 +/- 0.00013, or -$0.87 +/- $0.04. Eighty different critical core configurations were then analyzed to determine how they bounded around the bias of this initial critical configuration. Figure 4 shows these 80 configurations with respect to the bias run. All of these kcode calculations utilized 500,000 neutrons per cycle for 200 total cycles (175 active cycles).

NETL Neutronic Analysis 7 Feb 2023

Reactivity (including bias)

$1.00

$(1.00)

$(2.00)

$(3.00)

$(4.00)

$(5.00)

$(6.00) 0 10 20 30 40 50 60 70 80 Critical Configuration Number Figure 4 - Reactivity (including bias) of 80 Different BOL Critical Core Configurations There appears to be significant deviation in the first 40 configurations. Note that most of these configurations are at low power (less than 100 W) but some are at high power (greater than 200 kW and the point of adding heat). Note that low power with respect to modeling means that the MCNP decks utilized cross section data files at ambient room temperature, and high power means that the decks utilized cross section data files at 600K (327 ºC). Most of the configurations with significant deviation are the high power runs, which would indicate that either the model is inaccurate or there is evidence of another problem. One clue that there is a problem with the high power runs is the fuel temperature and critical rod heights. Some runs (like those on 3/23/92) listed reactor power around 400 kW, but fuel temperature only reads 20 degrees above ambient. There are also some criticality runs on 4/30/92 that produced an MCNP k-effective of approximately

$1.75, which may be indicative of a core that was not truly critical. For example, looking at the criticality state at 0944 on 4/30/92, the rods are at 0, 950, 481, and 488. At 1020, they are at 492, 950, 464, and 504. This does not make sense as the latter configuration has a rod that is 492 units removed while the other three rods are at essentially the same positions. The same occurrence can be seen at the timestamps at 1403 and 1407, where three rods are at the same heights but the transient rod is first at 396 units, then it is at 950 units.

NETL Neutronic Analysis 8 Feb 2023

If the first 44 runs are ignored (if runs after 5/1/92 are observed), the data looks more accurate (see Figure 5), with an average of -$0.23. The decision to ignore the first 44 runs was not arbitrary; the reactor was not run above 1 kW from 5/1/92 until 7/1/92. Thus, the runs between these dates were a cold clean core and it would also appear that something happened between those dates to make the MCNP model more accurate, likely an improvement in reactor power measurement, which would improve the accuracy of the critical rod heights.

Reactivity (including bias)

$0.60

$0.40

$0.20

$(0.20)

$(0.40)

$(0.60)

$(0.80) 45 50 55 60 65 70 75 80 Critical Configuration Number Figure 5 - Reactivity (including bias) of 36 Different BOL Critical Core Configurations Note that these latter 36 configurations include some full power operations (cases #70-72, 76, 78 and 80). There is only one outlier over +/-$0.60 (case #51), which would indicate that there were inconsistencies between high power operations during the first few months of operation. Other evidence, such as lower-than-expected fuel temperatures at these supposed high-power levels, would also indicate that something was inconsistent during the first few months of operation.

It is also important to note that some of these statistical outliers (e.g. cases #73 and #74) have unusual critical rod heights (0, 950, 950, 34 and 0, 855, 0, 950, respectively). It is possible that these having some control rods fully-withdrawn with others fully-inserted could be causing rod shadowing effects that are unable to be accurately simulated in MCNP.

Thus there are two aspects of the BOL criticality bias: the -$0.87 initial criticality bias which is likely due to inaccurate BOL fuel isotopics, and a -$0.23 bias that is mainly due to geometric NETL Neutronic Analysis 9 Feb 2023

differences in the critical rod height configurations. As stated earlier, the bias represents such things as differences in material properties that are difficult to determine or unknown (i.e., lack of manufacturer mass spectroscopy data on the exact composition of individual fuel meats and trace elements contained therein) or applicability of cross section data sets used to model the reactor (i.e., interpolation between temperatures). A large source of error is the uncertainty of the contents of the BOL fuel meats, as all of the fuel (except for the FFCRs) was previously irradiated. Without knowing the exact burnup and previous grid location of these elements, it is nearly impossible to accurately determine their fuel compositions.

Beginning-of-Life Rod Worth Bias NETL initially calibrated their control rods on 7/23/92. The control rod calibration procedure is listed in Appendix A. To simulate control rod calibrations in MCNP, two separate calculations were performed for each control rod. First, there are initial iterative attempts to achieve criticality by fully inserting the rod-to-be-calibrated and banking the other three rods at the same height until MCNP produces a k-effective near 1.0000. This simulates the initial state of the control rod calibration. Once this critical state is achieved, a second MCNP calculation is performed with the tested control rod fully withdrawn and all other parameters remaining the same. The difference between these two values is then simply the control rod worth. Table 2 shows a compilation of the MCNP-calculated rod worths compared to the NETL-measured control rod worths.

Table 2 - BOL Rod Worth Calculations Control Rod Full In Full Out MCNP Rod Worth Experimental Worth Difference Transient 1.00047 1.02407 $3.29 $3.26 $0.03 Regulating 1.00011 1.03067 $4.24 $4.08 $0.16 Shim 1 0.99992 1.0238 $3.33 $3.04 $0.29 Shim 2 1.00012 1.02419 $3.36 $3.17 $0.19 All Rods Out N/A 1.04248 $5.82 $6.38 -$0.56 (Core Excess)

MCNP appears to produce relatively reliable control rod worths. The control rod worth calculations are relatively close to the measured rod worths, especially the Transient and NETL Neutronic Analysis 10 Feb 2023

Regulating rods, with the model correctly predicting that the Regulating rod would have the most worth. These two rods are in the C-ring, thus closer to the center of the core. It would make sense that their predictions are more accurate as there is generally more error as one goes further from the center of the core due to the diffusive nature of neutronics, which is typically less accurate the closer you get to a boundary, as well as albedo effects near the reflector at the edge of the reactor core. It is important to note that these reactivity values do not utilize the criticality bias numbers from the previous section, as that bias would be moot, since these k-effective values are relative to one another.

MCNP under-predicts the core excess reactivity by $0.56, which coincides with the -$0.87 reactivity bias. It would appear that the BOL fuel isotopics may have been over-burned by the SCALE calculations, but as stated earlier, it is nearly impossible to determine the accuracy of the BOL fuel isotopics without knowing the fuels previous geometry during burnup.

Bias Conclusion The criticality bias and the rod worth bias calculations would appear to demonstrate that the MCNP model is relatively accurate with respect to geometry but is hampered by the inaccurate BOL fuel meat isotopics. The 44 critical core configuration bias calculations average of -$0.23 is relatively accurate considering that unknown fuel isotopics and rod shadowing effects could be causing general inaccuracies. The rod worth calculations predicted the most valuable control rod, reliably predicted the Transient Rod worth, and was still relatively accurate with regards to the Shim 1 and Shim 2 rod worths.

6. Burnup Calculations MCNP has a BURN option, which causes MCNP to invoke the CINDER90 code for depletion simulations. CINDER90 has an inventory of over 3400 nuclides and is compatible with MCNP.

This option requires the user to specify a time step (in days), a power fraction (typically 100% or 1.0), power level (in MW), and the materials that are to be depleted.

After performing the initial model bias calculations, a series of MCNP BURN calculations were performed to burn the NETL fuel to its current core configuration which was established in February 2018. This was a very detailed process as NETL is a very active facility and experienced NETL Neutronic Analysis 11 Feb 2023

many different core configurations. Using the fuel move logs, it was determined that there were 18 significant different core configurations that needed to be modeled (see Table 3).

Each burnup step involved performing the fuel burnup for the specified amount of MW-days, which is a computationally expensive task, often requiring 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to perform each step depending on cluster node usage. After the burnup calculation is completed, the output fuel isotopics were parsed, then the core model was reconfigured and the relevant fuel isotopics were pasted into the model and the next burnup step was performed.

Table 3 - Summary of Burnup Steps Burnup Total From To MW-days FEs Note Step MW-days 1 3/19/1992 10/12/1995 9.201 9.201 87 Initial Fuel Load 2 10/12/1995 1/20/1998 5.276 14.477 87 New IFE 3 1/20/1998 6/19/1998 2.789 17.266 87 Fuel Swapped Out/Add Rabbit 4 6/19/1998 3/4/1999 6.376 23.642 87 New IFE 5 3/4/1999 11/12/1999 7.671 31.313 90 Add 3 Fuel Elements 6 4/6/2000 6/29/2000 3.444 34.757 89 Core Reload 7 6/29/2000 1/29/2001 1.919 36.676 92 3L Experiment 8 1/29/2001 7/30/2001 9.138 45.814 92 3L Experiment with New IFE 9 7/30/2001 7/22/2002 21.508 67.322 95 Add 3 Fuel Elements 10 7/22/2002 11/13/2002 13.966 81.288 95 Fuel Shuffle 11 11/13/2002 4/1/2004 24.933 106.221 103 Add 8 New Fuel Elements 12 7/26/2004 7/13/2005 15.71 121.931 102 3L Experiment Core Reload 13 7/13/2005 7/11/2006 22.983 144.914 104 Add 2 Fuel Elements 14 7/11/2006 7/24/2007 41.732 186.646 104 Fuel Shuffle 15 7/24/2007 6/12/2008 18.347 204.993 108 Add 4 Fuel Elements 16 6/12/2008 6/24/2010 21.288 226.281 110 7L Experiment 17 6/24/2010 1/15/2016 73.587 299.868 114 Remove 7L Experiment 18 1/15/2016 2/22/2018 38.026 337.894 114 New IFE NETL Neutronic Analysis 12 Feb 2023

7. Current Core Configuration Once the burnup calculations were complete, the core was reconfigured to the current core configuration (as of 2/22/2018, see Figure 6). The next series of calculations were then performed to determine various neutronic characteristics of the NETL.

Figure 6 - Vertical Cross-section of Current Core Configuration MCNP Model Core Power Distribution F4 flux tallies were used to determine the power-per-element. The tallies output as a fluence per fission neutron. These units were converted to power density (W/cm3) which were then converted to power-per-element. The individual power-per-element values (in kW) are shown in Figure 7.

G26 Empty G27 5.74 G28 5.61 G29 5.54 G30 13.75 G24 5.13 F21 6.80 F22 7.61 F23 7.89 F24 7.97 F25 7.34 F26 6.61 G32 Source G23 5.49 F20 7.34 E17 8.98 E18 10.10 E19 10.40 E20 10.15 E21 9.07 F27 7.42 G33 6.31 G22 5.76 F19 7.77 E16 10.06 D13 11.66 D14 12.97 D15 12.55 D16 11.71 E22 10.28 F28 8.44 G34 Rabbit G21 5.34 F18 7.79 E15 10.16 D12 12.14 C09 13.49 C10 13.92 C11 13.71 D17 13.13 E23 10.74 F29 8.37 G35 6.38 G20 5.02 F17 7.03 E14 9.61 D11 11.95 C08 13.70 B05 15.54 B06 15.39 C12 15.12 D18 13.20 E24 10.60 F30 7.79 G36 5.74 F16 6.10 E13 8.39 D10 10.94 C07 13.72 B04 15.4 A01 CT B01 15.93 C01 Trans D01 12.70 E01 9.62 F01 7.18 G18 4.97 F15 7.23 E12 10.56 D09 12.26 C06 13.50 B03 15.82 B02 15.93 C02 14.14 D02 13.18 E02 11.38 F02 8.23 G2 5.91 G17 5.69 F14 Empty E11 Empty D08 12.16 C05 13.75 C04 14.91 C03 13.82 D03 13.01 E03 11.32 F03 8.93 G3 6.52 G16 6.53 F13 Empty E10 10.95 D07 11.27 D06 13.22 D05 12.99 D04 12.39 E04 11.05 F04 9.07 G4 6.84 G15 5.59 F12 7.44 E09 8.71 E08 10.01 E07 10.95 E06 10.90 E05 10.39 F05 8.39 G5 6.57 G14 4.96 F11 6.36 F10 7.47 F09 8.34 F08 8.53 F07 8.06 F06 7.35 G6 5.95 G12 5.20 G11 5.79 G10 6.14 G9 6.02 G8 5.71 Figure 7 - Current Core Power-Per-Element (in kW) Distribution at 1.1 MW NETL Neutronic Analysis 13 Feb 2023

The red highlighting indicates the hottest fuel element locations, which are in B-1 and B-2, with a maximum power of 15.93 kW (at a total maximum core power of 1.1 MW). B-2 is actually slightly higher than B-1 (15.931 kW vs. 15.929 kW) but both are within the 2-sigma error of 0.04 kW.

Effective Delayed Neutron Fraction and Prompt Neutron Generation Time MCNP outputs effective delayed neutron fraction (eff) and prompt neutron lifetime when using the KOPTS card. Nine different MCNP calculations (the same calculations used in the following Core Excess section) were used to determine eff and prompt neutron lifetime (see Table 4).

Table 4 - eff and Prompt Neutron Lifetimes for Current Core Configuration Prompt Neutron Case Error (s) eff Generation Time (s)

Trans fully in 47.62 7.543 0.00705 Trans fully out 46.868 7.111 0.00716 Reg fully in 48.08 7.824 0.00707 Reg fully out 46.718 6.961 0.00707 Shim I fully in 48.023 7.748 0.00702 Shim I fully out 46.777 6.974 0.00705 Shim II fully in 48.104 7.684 0.00717 Shim II fully out 46.708 7.086 0.00713 All Rods Out 45.824 6.626 0.00720 Average 47.191 7.284 0.00710 The average effective delayed neutron fraction eff was calculated to be 0.00710 +/- 0.00007. This is in reasonable agreement with values predicted in other LEU TRIGA cores (i.e., Oregon State University eff = 0.0076 [3], University of Maryland eff = 0.007 [4]) and also the value historically used for the NETL of eff = 0.007. The value eff = 0.007 will be used to express all dollar values of reactivities in this report.

The average prompt neutron generation time is 47.191 +/- 7.284 seconds.

NETL Neutronic Analysis 14 Feb 2023

Core Excess, Control Rod Worth and Shutdown Margin Nine different MCNP calculations were performed to determine core excess, control rod worth, and shutdown margin. Core excess is calculated as the reactivity of all rods withdrawn from the core. Control rod worths and shutdown margin were calculated by determining a critical state of the reactor with one rod full inserted and the other three rods banked at the same height, then fully withdrawing the previously-inserted rod. The resulting values (with comparison to values measured at NETL) are shown in Table 3.

Table 5 - Current Core Rod Worth Calculations MCNP MCNP MCNP Experimental Case k-effective k-effective Difference Rod Worth Reactivity Rod Full-In Rod Full-Out Transient 1.00035 1.02354 $3.24 $3.44 -$0.20 Regulating 0.99978 1.02214 $3.13 $3.18 -$0.05 Shim 1 1.00078 1.02248 $3.03 $3.09 -$0.06 Shim 2 1.00014 1.0211 $2.93 $2.94 -$0.01 All Rods Out

- 1.04118 $6.75 $6.06 $0.69 (Core Excess)

MCNP appears to accurately calculate the individual rod worths. The Regulating, Shim 1 and Shim 2 rods are all within the margin of error (which is approximately +/-$0.06 for each case).

These calculations show a core excess of $6.75 +/- $0.03. This is below the technical specification limit of $7.00. The core excess was measured by NETL to be $6.06 on 3/6/18. MCNP appears to have over-estimated core excess by approximately $0.70. This could be due to a variety of reasons, such as only modeling the fuel elements as one single material per element, thus some burnup resolution is lost as the fuel does not burn uniformly throughout.

The technical specification definition of shutdown margin is the minimum reactivity necessary to provide confidence that the reactor can be made subcritical by means of the control and safety systems starting from any permissible operating condition (the highest worth MOVEABLE EXPERIMENT in its most positive reactive state, each SECURED EXPERIMENT in its most reactive state), with the most reactive rod in its most reactive position, and that the reactor will remain subcritical without further operator action. The most reactive rod is the Transient rod.

NETL Neutronic Analysis 15 Feb 2023 1

Total rod worth minus the Transient rod is $9.09 +/- $0.06. NRC shutdown margin is this value minus the core excess, which would be $2.34 +/- $0.06, which is far above the technical specification limit of $0.29.

Prompt Fuel Temperature Coefficient The prompt-temperature coefficient associated with the NETL fuel, F, was calculated by varying the fuel meat temperature while leaving other core parameters fixed. The MCNP model was used to simulate the reactor with all rods out at 293, 600, 900, 1200 and 2500 K. The prompt-temperature coefficient for the fuel was calculated at the mid-point of the four temperature intervals. The results are shown in Figure 8 and tabulated in Table 5. Results from GA were added to show similarity [5]. The prompt-temperature coefficient is observed to be negative for all evaluated temperature ranges with decreasing magnitude as temperature increases. The coefficient has a value of -1.3¢/°C at 446.8 K, which is similar to the value of -$0.01/°C stated in the original SAR [6].

-$0.025 Prompt Fuel Temperature NETL GA

-$0.020

-$0.015 Coefficient ($ per K)

-$0.010

-$0.005

$0.000 200 700 1200 1700 2200 Temperature (Kelvin)

Figure 8 - Current Core Configuration Prompt Temperature Coefficient, F, as a Function of Temperature Table 6 - Current Core Configuration Prompt Temperature Coefficient Fuel Temperature [K] Prompt Temperature Coefficient [$/°C]

446.8 -$0.0130 750 -$0.0208 1050 -$0.0092 1850 -$0.0010 NETL Neutronic Analysis 16 Feb 2023

Moderator Void Coefficient The moderator void coefficient of reactivity was also determined using the MCNP model. The voiding of the core was introduced by uniformly reducing the density of the liquid moderator in the entire core. The calculation was performed from 0% to 100% voiding at 10% intervals. The void coefficient was negative for every interval and steadily decreased, as can be seen in Figure 9.

$0.00

-$0.20 Moderator Void -$0.40

-$0.60 Coefficient -$0.80

($ per % Void) -$1.00

-$1.20

-$1.40 0 20 40 60 80 100 Percent Void Figure 9 - Current Core Configuration Moderator Void Coefficient Moderator Temperature Coefficient The moderator temperature coefficient of reactivity, M, was determined by varying the moderator density with respect to temperature within the MCNP model from the expected operating temperature range of 20oC to 50oC (using Engineering Toolbox [7] to determine water density).

The results are shown in Figure 10. The moderator temperature coefficient is calculated to be slightly positive from 25C to 30 C and from 45 C to 50 C, but these changes are less than

$0.01/C and both points (with 2-sigma error) are bounded around zero. The moderator temperature coefficient appears to be negligible.

$0.020 Moderator Temperature

$0.015

$0.010

$0.005 Coefficient

$0.000

-$0.005

($ per C)

-$0.010

-$0.015

-$0.020 20 25 30 35 40 45 50 Moderator Temperature (C)

Figure 10 - Current Core Configuration Moderator Temperature Coefficient NETL Neutronic Analysis 17 Feb 2023

Power Coefficient of Reactivity The power coefficient of reactivity, otherwise known as power defect, is the amount of reactivity required to overcome the temperature feedback during the rise to power. This is modeled by analyzing two MCNP decks that are similar except for the neutron cross-sections used. Two k-effective calculations were performed with all rods out, one using cross sections at 293K (low power) and one using cross sections at 600K (full power). The results are seen in Table 6.

Table 7 - K-Effective Calculations Used to Determine Current Core Power Defect Case MCNP k-effective Standard Deviation Reactivity Error (2-sigma)

Low Power 1.04118 0.00012 $6.75 $0.03 Full Power 1.01327 0.00010 $2.94 $0.03 Power defect is simply the difference in reactivity between these two cases; thus the power defect is $3.81 +/- $0.05.

NETL Neutronic Analysis 18 Feb 2023

8. Limiting Core Configuration This section will suggest a limiting core configuration that utilizes fresh fuel to improve reactor efficiency while maintaining proper safety margins. The NETL limiting core configuration is a core that completely consists of fresh fuel.

Figure 11 shows the suggested limiting core configuration. For this analysis, it is suggested that the core is loaded with 84 fresh fuel elements (including FFCRs), which will provide just under the license limit of $7.00 core excess ($6.93 +/- $0.07). This is comparable to the original 1992 BOL core configuration, which was measured to have a $6.38 core excess on a core of 87 lightly-irradiated fuel elements. This configuration will provide maximum flux to the beam port facilities while maintaining safety margins.

Figure 11 - Vertical Cross-section of Limiting Core Configuration MCNP Model NETL Neutronic Analysis 19 Feb 2023

Core Power Distribution Figure 12 shows the power-per-element (in kW) in the suggested limiting core configuration.

G26 Empty G27 6.21 G28 5.93 G29 5.90 G30 18.01 G24 Empty F21 Empty F22 10.37 F23 10.17 F24 9.88 F25 9.49 F26 Empty G32 Source G23 6.74 F20 Empty E17 13.18 E18 13.64 E19 14.05 E20 12.77 E21 11.52 F27 Empty G33 Empty G22 5.76 F19 10.65 E16 14.12 D13 16.05 D14 14.99 D15 16.87 D16 14.61 E22 12.74 F28 Empty G34 Rabbit G21 5.76 F18 9.70 E15 13.76 D12 17.09 C09 19.22 C10 19.62 C11 18.13 D17 15.33 E23 13.08 F29 Empty G35 Empty G20 Empty F17 9.36 E14 12.68 D11 16.75 C08 19.75 B05 22.14 B06 21.59 C12 18.75 D18 15.70 E24 11.85 F30 8.55 G36 4.72 F16 Empty E13 11.45 D10 14.75 C07 15.98 B04 22 A01 CT B01 21.48 C01 Trans D01 14.41 E01 9.72 F01 6.43 G18 Empty F15 Empty E12 12.69 D09 15.69 C06 18.87 B03 21.58 B02 21.22 C02 18.52 D02 15.56 E02 11.81 F02 8.53 G2 4.67 G17 Empty F14 Empty E11 Empty D08 15.35 C05 18.01 C04 18.79 C03 17.52 D03 14.90 E03 12.93 F03 Empty G3 Empty G16 Empty F13 Empty E10 12.46 D07 14.34 D06 13.87 D05 15.90 D04 14.04 E04 12.43 F04 Empty G4 Empty G15 Empty F12 Empty E09 11.06 E08 12.17 E07 13.02 E06 12.02 E05 11.02 F05 Empty G5 Empty G14 Empty F11 Empty F10 8.90 F09 9.20 F08 9.20 F07 8.90 F06 Empty G6 Empty G12 Empty G11 5.45 G10 5.45 G9 5.46 G8 Empty Figure 12 - Limiting Core Configuration Power-Per-Element Distribution at 1.1 MW The hottest fuel element in now in location B-5. This makes sense as the core is more shifted to the northwest, which would better centralize the location of the maximum power production around B-5. Also, the hottest power-per-element is now 22.14 +/- 0.06 kW, which is higher than the current core hot channel, due to a lower fuel loading concentrating more power at the center of the core. According to the SAR analysis, CHFR values agree well and remain much greater than 2 at power levels up to 22.5 kW per unit cell. Thus 22.14 is acceptable maximum hot channel power.

Effective Delayed Neutron Fraction and Prompt Neutron Generation Time Once again using the KOPTS card and running nine cases, the effective delayed neutron fraction eff and prompt neutron generation times were calculated Table 8 - eff and Prompt Neutron Lifetimes for Limiting Core Configuration Case Prompt Neutron Generation Time (s) Error (s) eff Trans fully in 42.828 5.531 0.00743 Trans fully out 42.721 5.024 0.00725 Reg fully in 43.764 5.502 0.00732 Reg fully out 41.951 4.985 0.00742 Shim I fully in 43.546 5.616 0.00737 Shim I fully out 42.407 5.104 0.00737 Shim II fully in 43.614 5.458 0.00733 Shim II fully out 42.261 5.200 0.00728 All Rods Out 42.024 4.965 0.00742 Average 42.791 5.265 0.00735 NETL Neutronic Analysis 20 Feb 2023

The average eff was calculated to be 0.00735 +/- 0.00007. There is a slight increase in eff compared to the current core configuration, but for consistency, 0.007 will continue to be used to express all dollar values of reactivities in this report.

The average prompt neutron generation time is 42.791 +/- 5.265 seconds.

Core Excess, Control Rod Worth, and Shutdown Margin The same nine MCNP rod worth calculations were performed again for the limiting core configuration: Core excess, shutdown margin, and individual rod worths were calculated from these outputs and the reactivity values (with the bias taken into account) of each of these calculations are shown in Table 7.

Table 9 - Limiting Core Configuration Rod Worth Calculations MCNP k-effective MCNP k-effective MCNP Rod Case Rod Full-In Rod Full-Out Worth Transient 0.99886 1.02191 $3.22 Regulating 1.00024 1.03222 $4.43 Shim 1 1.00003 1.02431 $3.39 Shim 2 1.0003 1.02857 $3.93 All Rods Out (Core Excess) - 1.04257 $6.93 These calculations show a core excess of $6.93 +/- $0.07. This is below the technical specification limit of $7.00.

Now the most reactive rod is the Regulating, due to having more fuel near its vicinity and the power shifted to the northwest side of the core. Total rod worth minus the Regulating Rod is $10.53

+/- $0.16. NRC shutdown margin is this value minus the core excess, which would be $3.60 +/- $0.16, which is still far above the technical specification limit of $0.29.

NETL Neutronic Analysis 21 Feb 2023

Prompt Fuel Temperature Coefficient The results of the limiting core configuration prompt fuel temperature coefficient calculations are shown in Figure 13 and tabulated in Table 9.

-$0.020 Prompt Fuel Temperature

-$0.018 NETL GA

-$0.016

-$0.014

-$0.012

-$0.010 Coefficient ($ per K)

-$0.008

-$0.006

-$0.004

-$0.002

$0.000 200 700 1200 1700 2200 Temperature (Kelvin)

Figure 13 - Limiting Core Configuration Prompt Temperature Coefficient, F, as a Function of Temperature Table 10 - Limiting Core Configuration Prompt Temperature Coefficient Fuel Temperature [K] Prompt Temperature Coefficient [$/°C]

446.8 -$0.01014 750 -$0.01858 1050 -$0.00860 1850 -$0.000989 These values are slightly higher than the original BOL coefficients, likely due to the fresh fuel.

Moderator Void Coefficient Figure 14 shows the moderator void coefficient in the suggested limiting core configuration.

$0.00 Moderator Void Coefficient

-$0.20

-$0.40

-$0.60

($ per % Void)

-$0.80

-$1.00

-$1.20

-$1.40 0 20 40 60 80 100 Percent Void Figure 14 - Limiting Core Configuration Moderator Void Coefficient NETL Neutronic Analysis 22 Feb 2023

The void coefficient was negative for every interval and steadily decreased, similar to the current core configuration. The void coefficient is slightly more negative in the limiting core configuration, likely due to having more moderator in the core configuration.

Moderator Temperature Coefficient Figure 15 shows the moderator temperature coefficient in the suggested limiting core configuration.

$0.015

$0.010 Moderator Temperature

$0.005 Coefficient

$0.000

-$0.005

($ per C)

-$0.010

-$0.015

-$0.020 20 25 30 35 40 45 50 Moderator Temperature (C)

Figure 15 - Limiting Core Configuration Moderator Temperature Coefficient Once again the moderator temperature coefficient appears to be negligible as it bounds around

$0.00 at nearly all observed temperature ranges.

Power Coefficient of Reactivity The power coefficient of reactivity results are seen in Table 10.

Table 11 - K-Effective Calculations Used to Determine Limiting Core Power Defect Case MCNP k-effective Standard Deviation Reactivity Error (2-sigma)

Low Power 1.04231 0.00015 $6.90 $0.04 Full Power 1.01921 0.00010 $3.79 $0.03 NETL Neutronic Analysis 23 Feb 2023

Thus the power defect is $3.11 +/- $0.05. This is lower than the current core configurations power defect, likely due to less resistance at the point-of-adding-heat due to the lower amount of zirconium-hydride in the core.

Hot Channel Power Summary The hot channel in the limiting core configuration was determined to be B-5. An fmesh calculation was performed to analyze a 20 by 20 mesh array to determine axial and radial power distributions.

Table 11 summarizes the results of this calculation.

Table 12 - Limiting Core Hot Channel Power Summary Hot Rod Hot Rod Hot Rod Hot Rod Core Hot Rod Axial Peak Radial Peak Effective Thermal Peak Factor Configuration Location Factor Factor Peak Factor Power [kW] [Pmax/Pavg]

[Pmax/Pavg] [Pmax/Pavg]

Limiting Core B6 22.14 1.691 1.296 1.017 2.229

9. Summary MCNP6.2 was used to calculate fundamental and operational parameters for the Nuclear Engineering Teaching Laboratory Reactor to demonstrate the reactors adherence to safety margins in the technical specifications. Values of fundamental parameters agree well with theoretical values. Values of operational parameters agree well with measured values, giving confidence in the models ability to predict the viability of future core configurations. The results of this study indicate that the NETL can be operated safely within the Technical Specification bounding envelope and that its MCNP model can be used to predict future core configuration changes.

NETL Neutronic Analysis 24 Feb 2023

REFERENCES

[1] NUREG-1282, Safety Evaluation Report on High-Uranium Content, Low-Enriched Uranium-Zirconium Hydride Fuels for TRIGA Reactors, USNRC, August 1987.

[2] C.J. Werner, et al., "MCNP6.2 Release Notes", Los Alamos National Laboratory, report LA-UR-18-20808 (2018).

[3] Safety Analysis Report for the Conversion of the Oregon State University TRIGA Reactor from HEU to LEU Fuel, Submitted by the Oregon State University TRIGA Reactor (2007).

[4] Analysis of the Neutronic Behavior of the Maryland University Training Reactor, Submitted by the Oregon State University Radiation Center to the Department of Energy (July 2017).

[5] GA-7882, Kinetic Behavior of TRIGA Reactors, General Atomics (1967).

[6] Safety Analysis Report Submitted by the University of Texas at Austin Nuclear Engineering Teaching Laboratory (January 2012).

[7] Engineering Toolbox. Web. Accessed May 3rd, 2017.

Link: http://www.engineeringtoolbox.com/water-thermal-properties-d_162.html NETL Neutronic Analysis 25 Feb 2023

APPENDIX A - NETL Control Rod Calibration Procedure NETL Neutronic Analysis 26 Feb 2023

P2Format.doc Date: 3/2/09 d:\Procedures\Surv\surv6int.doc Number* Rev.: SURV-6: 1.00 d:\Procedures\Surv\surv6pro.doc Procedure Title Control Rod Calibration

]

2 3 PROCEDURE 4

5 SURV-6 6

7 Control Rod Calibration 8

9 Version 10 1.00 I1 12 13 14 15 16 17 18 19 20 21 22 Approvals :

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Number of Pages: 8 38 Number of Words: 2396 39 Number of Characters: 11929 40 41 NUCLEAR ENGINEERING TEACHING LABORATORY 42 J. J. PICKLE RESEARCH CAMPUS 43 THE UNIVERSITY OF TEXAS AT AUSTIN NETL Procedure Stamp (Original-Red, Copy-Blue)

Page 1 of 8

P2Format.doc Date: 3/2/09 d:\Procedures\Surv\surv6int.doc Number- Rev.: SURV-6: 1.00 d:\Procedures\Surv\surv6pro.doc Procedure Title . Control Rod Calibration

]

2 I. INTRODUCTION 3

4 A. Purpose 5 The Control Rod Calibration Procedure benchmarks the primary system for reactor 6 control and safety.

7 8 B. Description 9 Knowledge of control rod worth is necessary to assure the appropriate performance of the 10 reactor control system and demonstrate compliance with Technical Specification limits.

11 Both routine operating conditions and the safety functions of the control rods depend on 12 accurate calibration data. Two separate methods of measurement are available to provide 13 calibration data. The Rod Drop Experiment determines the approximate integral control 14 rod worth by observation of the change in reactor power level as a function of time after 15 the rod drop. This experiment provides the initial estimate of a rod's worth and may be 16 used after major core rearrangements to predict approximate rod worth. The experiment 17 may also verify the total rod worth after minor core changes. The second method of rod 18 calibration is the Positive Period Experiment. This method provides the most accurate 19 measurement of the differential rod worth. This experiment determines both the total 20 control rod worth and the shape of the control rod position versus control rod worth 21 curve. The Positive Period method should be used for normal control rod calibration 22 23 Measurement of the rod drop times verify the performance of the system safety function 24 per Technical Specification requirement. The SCRAM switch or relay in the safety 25 circuit initiates the safety circuit action dropping all control rods. Individual rod switches 26 initiate the drop of each individual rod. Rod position switches sense when the rods reach 27 the full down position. Proper performance of the safety system is indicated if all rods 28 reach the full down position in the specified time limit.

29 30 Measurement of the control system rod removal rate coupled with the control rod peak 31 differential worth establishes the maximum reactivity insertion rate of each control rod.

32 This rate is limited as specified in the Technical Specifications to allow the safe control 33 of the reactor in manual or auto mode.

34 35 36 C. Schedule 37 Control rod calibrations are to be done at least once each year and after any significant 38 change to the reactor core configuration. Annual calibration measurements should be 39 done in January or July but shall not exceed longer than 15 months from preceding 40 measurement.

41 42 Measurement of control rod drop time and reactivity insertion rate should be done 43 annually, not to exceed 15 months from preceding measurement and/or after control rod Date of Change: I I I I I I Stamp (Original-Red, Copy-Blue)

Change Approval: I I

I I I

I I

NETL Director Page 2 of 8

P2Format.doc Date: 3/2/09 d : \Procedures\Surv\surv6int . doc Number - Rev.: SURV-6: 1.00 d : \Procedures\Surv\surv6pro . doc Procedure Title : Control Rod Calibration l or drive maintenance, reactor core reconfiguration, or movement of fuel adjacent to the 2 standard rod drives.

3 4

5 D. Contents 6 A. Rod Drop Procedure page 4 7 B. Positive Period Procedure page 5 8 C. Control Rod Drop Time and Removal Rate Measurement page 7 9

10 E. Attachments 11 1. Reactivity vs. Power Ratio Plot 1 Page 12 2. Positive Period Data Sheet 1 Page 13 3. Stable Period Wait Time 1 Page 14 4. Inhour curve - Reactivity vs. Period Plot 1 Page 15 5. Rod Drop Time / Withdrawal Rate Data Sheet 1 Page 16 17 F. Equipment, Materials 18 TRIGA ICS System with control rod drives 19 Data Analysis Software such as "MathCAD" 20 Digital Stopwatch 21 Digital Storage Oscilloscope 22 23 G. References, Other Procedures 24 MAIN-6, Rod & Drive Maintenance, Inspection 25 Attachments 1 & 3 :

26 A. Edward Profio "Experimental Reactor Physics", John Wiley and Sons 27 Inc., 1976 pp 712, 716 28 Attachment 4 :

29 General Atomics Data Sheet.

30 31 Date of Change: Stamp (Original-Red , Copy-Blue)

Change Approval:

NETL Director Page 3 of 8

P2Format.doc Date: 3/2/09 d : \Procedures\Surv \ surv6int . doc Number- Rev.: SURV-6: 1.00 d : \Procedures\Surv\surv6pro . doc Procedure Title : Control Rod Calibration I

2 I

I. PROCEDURE

3 4

5 A. Control Rod Worth Estimate by Rod Drop Method:

6 7 Use to estimate initial control rod worth following new core start-up.

8 May be useful folJowing substantial core reconfigurations.

9 10 I. The reactor core condition should be cold and clean prior to measurement of rod worth.

I1 Perform JCS system pre-start checks. The reactor coolant system pumps should be off 12 during control rod calibration.

13 14 2. Commence Startup of the reactor:

15 16 a Position the control rod being evaluated at the desired position - full up if the 17 entire rod worth is desired to be estimated in one step or partially withdrawn at 18 selected increasing withdrawn locations if several drops are to made.

19 20 b. Position the two rods closest to the rod being evaluated at a banked elevation, 21 position the control rod farthest from the rod being calibrated at about 900 units to 22 allow fine control of its reactivity for achieving criticality.

23 24 c. Adjust control rods for criticality at a low power level such as 50 to 500 watts.

25 The power should not be so high as to see a fuel temperature increase above 26 ambient i.e. less than 1 Ki lowatt.

27 28 d. Remove the neutron source and readjust for criticality. The delayed neutrons 29 should be allowed to come into equilibrium as evidenced by the indicated power 30 remaining constant to within +/- 2% for a minimum of 3 to 5 minutes without 31 further rod movement.

32 33 3. Setup data recording system to record reactor linear power as a function of time or use 34 stopwatch and indication on linear power display to tabulate initial power and the 35 indicated power after the control rod is dropped .

36 37 4. Drop control rod being evaluated by actuation of magnet button (standard rod drives) or 38 air button (transient rod drive) and document the power vs. time data. Select times to 39 record data based on time data plotted on the graph in Attachment J - Ratio of neutron 40 density after a rod drop to the initial density (at critical) as a function of subcritical 41 reactivity.

42 I

Date of Change: I I

I StamP, {Original-Red, Copy-Blue)

I Change Approval I I

I I

I I

I I

I I

I NETL Director Page 4 of 8

P2Format.doc Date: 3/2/09 d : \Procedures \Surv\surv6int . doc Number* Rev.: SURV-6: 1.00 d : \ Procedures \Surv\surv6pro . doc Procedure Title : Control Rod Calibration 1 5. Using the data in Attachment 1 determine the reactivity associated with the rod drop 2 based on the measured neutron density ratio (power ratio) at the specified time after the 3 rod drop.

4 5 B. Control Rod Worth Measurement by Positive Period Method:

6 7 Use for the annual rod worth calibration.

8 Use as the primary rod calibration method.

9 10 11 1. The reactor core condition should be cold and clean prior to measurement of rod worth .

12 Perform ICS system pre-start checks. The reactor coolant system pumps should be off 13 during control rod calibration.

14 15 2. Commence Startup of the reactor:

16 17 a. Position the control rod being evaluated at the desired position - full down if the 18 entire rod worth is to be evaluated, or at predetermined locations if the shape of 19 the differential rod worth curve has already been established.

20 21 1. Initial control rod calibrations or calibrations after major core 22 reconfigurations should evaluate the entire rod worth by stepwise pulling 23 the rod in increments correlating to reactivity steps of 15 to 20 cents over 24 its entire travel. This will require taking 10 to 20 measurements per 25 control rod depending on its total worth.

26 11. Once the initial control rod calibration curve shape has been established, 27 subsequent routine control rod calibrations may be made by using only 5 28 or 6 appropriately selected insertions of the same reactivity magnitude as 29 above. One or two points should be selected near the rod height 30 correlating to the peak differential rod worth. Four additional points 31 should be selected, two in the lower and two in the upper parts of the rod 32 travel correlating to areas spaced roughly equally on the slope portions of 33 the differential worth curve. The data from these measurements can then 34 be curve fit to the shape of the differential control rod worth curve to 35 determine the actual rod worth.

36 37 b. Position the two rods closest to the rod being evaluated at a banked elevation, 38 position the control rod farthest from the rod being calibrated at about 900 units to 39 allow fine control of its reactivity for achieving criticality.

40 41 C. Adjust control rods for criticality at a low power level of 1 to 3 Watts.

42 I I Date of Change: I I I

I Stamp (Original-Red , Copy-Blue)

I I Change Approval I l NETL Director Page 5 of 8

P2Format.doc Date: 3/2/09 d : \Procedures \Surv\surv6int . doc Number- Rev.: SURV-6: 1.00 d : \Procedures \Surv\surv6pro . doc Procedure

Title:

Control Rod Calibration

] 3. Remove the neutron source and readjust for criticality. The delayed neutrons should be 2 allowed to come into equilibrium as evidenced by the indicated power remaining 3 constant to within +/- 4% for a minimum of 3 .5 to 5 minutes without further rod 4 movement. This constraint will limit measurement errors of criticality to +/- ~0.25 ¢ per 5 measurement.

6 7 4. Record the Control rod positions on the Control Rod Calibration Data Sheet in 8 Attachment 2.

9 JO 5. Pull the control rod being calibrated in one smooth movement a distance correlating to an 11 estimated reactivity worth of 15 to 20 cents which correlates to a stable period between 12 58 and 37 seconds. Record the rod position stop point on data sheet. (To minimize rod 13 position hysteresis, if you inadvertently pull the rod too far, quickly move the rod back 14 down slightly below the target point, then raise the rod to the target point.) Refer to 15 previous calibration data to estimate the number of units to move the rod. Typical 16 movements are 90 to I 00 units for the initial and final pull at the full down or full up 17 endpoints, decreasing rapidly to 20 to 40 units per pull in the mid range of rod travel.

18 The reactivity per pull is limited to allow the reactor to attain a stable period prior to 19 taking the power vs. time data thus reducing measurement errors. The time to reach a 20 stable period is called the wait time. The wait times for 5% error are 20 to 35 seconds, 21 for a 1% error they increase to 50 to 65 seconds respectively for 37 to 58 second stable 22 periods. A table showing measurement errors as a function of the wait time required to 23 attain a stable period is shown in Attachment 3.

24 25 6. Observe the power increase as indicated on the digital readout of the auto ranging linear 26 power channel on the Animation Window. Use a stopwatch set to measure time intervals 27 with respect to the start time. Start the primary stopwatch when the power passes the 60 28 watt point. Record the time when the power passes the 90 watt level, the 600 watt level 29 and the 900 watt level (time points should be marked at the first instant the power reaches 30 the target value on the digital display). Time data at powers above the I Kilowatt level 31 shall not be used as temperature feedback will create errors above this level.

32 33 7. Drive control rods other that the rod being calibrated down to decrease the reactor power.

34 Leave the control rod being calibrated at the point to which it was withdrawn in step 5 if 35 the entire rod is being stepwise calibrated. If the curve fit method is being used, 36 reposition the rod being calibrated to its next starting point.

37 38 8. Repeat steps 2 through 7 until the remainder of the rod is completed or sufficient data 39 points for curve fitting are obtained.

40 41 Notes: As long as the power level is not allowed to fall below the source interlock the source 42 may be continuously left out of the core until all the data points desired are obtained.

I I Date of Change: I I

I I I I

I I

I I Stamp (Original-Red, Copy-Blue)

I Change Approval I NETL Director Page 6 of 8

P2Format.doc Date: 3/2/09 d : \Procedures\Surv\surv6int .doc Number* Rev.: SURV-6: 1.00 d : \Procedures\Surv\surv6pro . doc Procedure Title : Control Rod Calibration 1 Each sequence through this process will take approximately 15 minutes if everything is 2 done with attention to detail so plan accordingly.

3 4 9. Analyze data either manually or via software program. Using the time data recorded, 5 calculate the stable period resulting from each rod pull. Then use the reactivity equation 6 or inhour curve in Attachment 4 to determine the reactivity associated with each rod pull.

7 8 10. A senior operator should review and approve the rod calibration data. If significant 9 changes in rod worth are indicated, a review of the implications on excess reactivity and IO shutdown margins should also be initiated.

11 12 13 C. Control Rod Withdrawal, insertion, and drop time measurements.

14 15 l. Perform ICS system pre-start checks if not already completed.

16 17 2. Setup drop time measurement system. The magnet power supply voltage level controlled 18 by the console scram switch should be used to start the timing. A signal from the control 19 rod down limit switch should be monitored to indicate when the rod has reached the full 20 down position 21 a. Measurement equipment should be a storage oscilloscope or an electronic timer 22 with signal start-stop features. Use of a stopwatch to measure rod drop time, 23 manually started at the time the scram button is depressed and stopped at the time 24 the rod visually hits bottom is also acceptable but not the preferred method of 25 measurement.

26 b. Measurement resolution for oscilloscope sweep should be set to 100 ms/div, 27 vertical gain should be set to 5 V/div. Vertical signal probe should be set to X 10 28 for the transient rod, and X 1 for all other rods. Scope should be set to Auto 29 trigger mode while setting up, and changed to single trigger or normal mode when 30 taking the data.

31 c. Connect start signal (scope trigger) to the Regulating rod positive magnet power 32 (see table below for connection location). Set the scope trigger to DC coupling on 33 a negative slope at a level of about 10 volts. The nominal magnet power high side 34 is + 13 volts and the low side is -6volts.

35 d. Connect the signal (Channel 1) to the rod drive down limit switch (see table 36 below for connection location) of the drive being evaluated.

37 Scope Input Channel DAC Tie Bar Description Trigger TB 5-3 Reg Magnet Pwr (+ 13 V)

CH I TB 8-8 TR rod down limit CH l TB 8-16 Shim 1 rod down limit CH 1 TB 8-24 Shim 2 rod down limit CH 1 TB 9-32 Reg rod down limit Date of Change: I I

I I

I I

I I

I I

I I Stamp (Original-Red, Copy-Blue)

I Change Approval I I I

I I I

I I

NETL Director Page 7 of 8

P2Format.doc Date: 3/2/09 d:\Procedures\Surv\surv6int.doc Number* Rev.: SURV-6: 1.00 d:\Procedures\Surv\surv6pro.doc Procedure Title . Control Rod Calibration 1

2 3 3. Withdraw control rod being measured about 60 units and test drop the rod to verify the 4 scope setup.

5 6 4. Fully withdraw the rod being evaluated measuring the time it takes to move from full 7 down to full up using a stopwatch. Record data on Attachment 5.

8 9 5. Drop the control rod to trigger and record a trace by initiation of the scram button. Drop 10 time is measured from the time the scope triggered until the rod reaches full down, as 11 evidenced by the transition of the signal on the rod down switch. Some rods may show a 12 bounce after the initial bottom transition ' typical drop time recorded is the time measured 13 to when the rod remains full down as indicated on the trace. Record data on Attachment 14 5.

15 16 6. Repeat steps 2d through 5 for each remaining rod.

17 18 7. Calculate measured reactivity insertion rate and record data on Attachment 5:

19 20 a.. Obtain peak differential rod worth near rod midpoint for each rod from the control 21 rod calibration data.

22 b. Calculate insertion rate(< 0.2 % k/k/sec) as follows:

23 24 rate(% k/k/sec)=rate (units/sec)* worth(¢/unit)*(0.7% L\k/k/100¢ )

25 26 8. Document any relevant notes, comments, or observations on Attachment 5 data sheet.

27 28 Date of Change: I

!  !  !  !  ! Stamp (Original-Red, Copy-Blue)

Change Approval I I I I I I NETL Director Page 8 of 8

PAFormat.doc Date: 3/2/09 Attachment Number- Rev.: SURV-6 1.00 d : \Procedures\Surv\surv6 - al . doc Procedure

Title:

Control Rod Calibration 0II)

.,,0.

II?

fA 0 a:

II) 0 0

Q~

I-(.)

II?

N IIJ a:

q N (.)

an.

- Ir m

0 C? U)

'o (0)U/(1) U 'Al.lSNlO NOYJ.nlN 2

3 Reactivity vs . Power Ratio Stamp(Original-Red, Copy-Blue)

Date of Change NETL Director Approval Page 1 of 1

PAFormat.doc Date: 3/2/09 Attachment Number- Rev.: SURV-6 1.00 d : \Procedures\Surv\surv6-a2 . doc Procedure

Title:

Control Rod Calibration Rod: P('t)=P 0 e-rrr 2 Date: T=~-c/ln(P(1:)/P 0 )

3 =~-c/ln (10) 4 = .434 ~'t Steady-state: ~3 watts Rod positions Times at power TR SI S2 Reg start stop 60 watts 90 watts 600 watts 900 watts l

2 3

4 5

6 7

8 9

10 11 12 13 14 15 16 17 18 19 20 5

Positive Period Data Starnp(Original-Red, Copy-Blue)

Date of Change I I I I I NETL Director Approval I I I I I Page 1 of 1

PAFormat.doc Date: 3/2/09 Attachment Number- Rev.: SURV-6 1.00 d : \Procedures\Surv\surv 6- a3 . doc Procedure

Title:

Control Rod Calibration 160 2 .5 140 120 2 NO SOURCE ( ko = I) 100 fl, =4 .0 X I0- 5 SEC .

Cl) 0 1.5 (/)

z .-.::>w Sao z w

C/)  ::E z z 60 w w

E  ::E

~ ~

40 10-X, 0 .5 20 3

10 POSITIVE ASYMPTOTIC PERIOD IN SECONDS 2

3 Stable Period Wait Time Stamp(Original-Red, Copy-Blue)

Date of Change _ _ _ _ J, _ _ J _ _ I NETL Director Approval l_~_ _I Page 1 of 1

PAFormat.doc Date: 3/2/09 Attachment Number- Rev.: SURY-6 1.00 d : \Procedures\Surv\surv6-a4 . doc Procedure

Title:

Control Rod Calibration u

I 0

0 2

1..0, x.-1 sun, NI .1.111\11,.,1.,

~

Inhour Curve Stamp(Original-Red, Copy-Blue)

Date of Change I I I NETL Director Approval Page 1 of 1

PAFormat.doc Date: 3/2/09 Attachment Number- Rev.: SURV-6 1.00 d : \Procedures\Surv\surv6-a5 . doc Procedure

Title:

Control Rod Calibration I

2 Rod Drop/Withdrawal Data 3

4 5 Rod Drop Time:

6 (limit : less than 1 second) 7 Time Verified OK Rod (Sec) (initial)

Transient Shim 1 Shim 2 Regulating 8

9 10 11 Maximum Reactivity Insertion Rate :

12 (limit : less than 0 . 2 % l'iK/K/sec) 13 Peak Withdrawal Insertion Rate Rod Differential Time ( sec) (% l'iK/K/ sec)

Worth (¢/unit)

Transient Shim 1 Shim 2 Regulating 14 15 Comments:

16 17 18 19 20 21 SRO Approval : Date: _ / _ / _

22 Rod Drop/Insertion Rate Data Stamp(Original-Red , Copy-Blue)

Date of Change I I I I I NETL Director Approval I I I I I Page 1 of 1

THERMAL HYDRAULIC ANALYSIS OF THE UNIVERSITY OF TEXAS (UT) TRIGA REACTOR Paul (Michael) Whaley and William S. Charlton Nuclear Engineering Teaching Laboratory University of Texas at Austin Austin, TX 78758 February 2023

1.0 Introduction This report documents analysis of the thermal hydraulic characteristics of the UT TRIGA in support of renewal of the U.S. Nuclear Regulatory Commission facility operating license.

The UTAustin TRIGA Research Reactor (UT TRIGA) is a TRIGA MarkII nuclear research reactor licensed to The University of Texas at Austin for operation up to 1.1 MW steadystate thermal power level. The geometry of the UT TRIGA core is based on seven concentric hexagons (designated as rings) that fix locations for fuel elements, graphite filled elements, and various experiment facilities. The core is surrounded by a modified cylindrical annulus in an aluminum container filled with graphite (neutron reflector), a rotary specimen rack (RSR), four beam port penetrations, and void spaces accommodating the RSR and beam port facilities. The core and reflector are located in an aluminum tank (pool) filled with highpurity water. The water acts as a neutron moderator, coolant, and radiation shield.

Thermal hydraulic modeling of the UT TRIGA was performed with TRAC/RELAP Advanced Computational Engine (TRACE). Thermal hydraulic characteristics were developed from classical methods and corrections for UT TRIGA geometry using the computational fluid dynamics code FLUENT. Distribution of fission activity was developed from transport calculations in MCNP.

The thermal hydraulic codes TRACE and RELAP are designed to perform bestestimate analyses of operational transients and accident scenarios by modeling physical geometry and thermodynamic conditions. TRACE and RELAP were developed for commercial nuclear reactor applications, and RELAP has been widely used in characterizing research reactor thermal hydraulic performance. TRACE is the NRC's flagship thermal hydraulics analysis tool consolidating and extending the capabilities of NRC's 3 legacy safety codes: TRACP, TRACB and RELAP.

NRC guidance1 defines a limiting core configuration (LCC) as the core that would yield the highest power density using the fuel specified for the reactor, with all other core configurations demonstrated to be encompassed by safety analysis for the limiting core configuration. In this report, hot channel analysis was used to determine the power level and thermal hydraulic characteristics of the fuel element generating the highest power.

The guidance references an operational core. Analytical methods used to define the LCC were applied to the operational core, providing confidence that the model adequately supports LCC analysis.

2.0 General Description of Heat Transfer at the UT TRIGA Heat is generated in the fuel by the fission process. Cooling is required to maintain fuel temperature low enough to prevent challenges to cladding integrity. Fuel cladding is the principal safety feature of the TRIGA reactor, preventing radioactive fission products from release that could result in possible hazardous exposure to radiation for facility personnel and the general public. The UT TRIGA reactor operates in a natural convectioncooling mode. Heat transfer from fuel to the coolant in the core area is developed by generation of heat in the fission process, conduction of the heat to external surface of the fuel element, and heat transfer by convection from the fuel element surface to water in the core area.

Temperature increase of the water in the core area develops buoyancy forces that drive flow. The flow is diminished by momentum changes and friction (across the gird plates, fuel element end fittings, and fuel 1

NUREG 1537, Guidelines for Preparing and Reviewing Applications for the Licensing of NonPower Reactors, Format and Content 1

element cladding surfaces). Above a critical heat flux, coolant flow will not be adequate to prevent thermal hydraulic conditions from exceeding limits. This analysis demonstrates that operation at the maximum licensed power level has adequate margin to the fuel temperature limits and critical heat flux.

3.0 Power Distribution The distribution of heat generation across the fuel elements in the core and the ratio of peaktoaverage power in the hot channel fuel element (i.e., the fuel element producing the most power) are described in the Neutronics Report2. The ratio of the hot channel peak power to element average power was derived from MCNP calculations for a 2dimensional (axial and radial) mesh tally. Mesh tally results were used to develop power distribution profiles for a fuel element to support TRACE calculations.

4.0. Thermal Hydraulic Modeling, Unit Cell Geometry and Thermal Hydraulic Characteristics 4.1 UTTRIGA Unit Cell Geometry The flow channel unit cell cross section is based on the typical fuel element geometry, as illustrated in Fig. 4.1 (unit cell and the surrounding fuel elements). Some unit cell locations in the grid plate have different structures. The central thimble is not fueled, the transient rod does not contain fuel, and the fuel followers (which are generally not fully inserted in the core) have 80% of the fuel mass contained in standard fuel elements (which are generally not fully inserted. This analysis uses the hot channel identified in the neutronics report and assumes no interaction between adjacent unit cells. Any interaction between unit cells with fuel and adjacent unit cells with less or no fuel contributes a larger area where convection flow is the result of heat transfer from the fully fueled cell, resulting in enhanced heat removal from the fully fueled cell. Thus, from this standpoint the analysis here is conservative. As illustrated, the unit cell analysis is based on a fuel element and the surrounding flow area (end fittings have more complex geometry) circumscribed by a hexagon with an inner radius of 1/2 of the pintopin pitch. The complex geometries of the fuel element end fittings are approximated as hydrodynamic characteristics.

Figure 4.1, Flow Channel for UT TRIGA Fuel Elements 2

Analysis of the Neutronic Behavior of the Nuclear Engineering Teaching Laboratory at The University of Texas, Radiation Center - Oregon State University (March 2021), submitted concurrently with this report 2

Since a regular hexagon can be decomposed into six equilateral triangles, a triangular unit cell is the smallest possible unit cell. However, TRACE heat structures (described in a following section) have limited options for temperature analysis of solid structures; a cylinder can be used to develop a fully symmetric heat source, but this is not possible with a halfcylinder. This does not limit fluid analysis in thermal hydraulic calculations with a triangular unit cell but limits the ability to calculate temperatures in the fuel element since the geometry of a triangular unit is 1/2 of the heat contribution from a single fuel element. Intrinsic properties used to calculate thermal hydraulic conditions are fully represented, but total heat for the cylindrical fuel element (used in material temperature calculations) in a triangular unit cell is not.

The volume of the flow channel is calculated as the product of the flow area and length. The length of the TRIGA flow channel is defined for the heated (adjacent to fuel) and unheated surfaces of fuel element cladding. The heated length is divided into smaller sections for analysis. The geometries and thermal hydraulic parameters of the upper and lower grid plate/fuel element are calculated through equations 4.1 4.9, with results summarized in Table 4.2.

The area of a regular polygon is calculated using the interior radius (ri) and perimeter (P) as:

1 A ri P 4.1 2

The unit cell is a hexagon (i.e., 6sided perimeter) with each side one leg of an equilateral triangle; the height of the triangle is the hexagons interior radius. The hexagon/triangle dimension (a) in terms of the internal radius is calculated:

2 a ri 4.2 3

Substituting 4.1 into 4.2, the crosssectional area of the hexagonal unit cell (AUC) using the interior radius is therefore:

1 2 AUC ri 6 ri 2 3 ri 2 4.3 2 3 The inner radius of the unit cell is 1/2 the distance between two fuel elements or 1/2 of the fuel element pitch (pe) so that:

3 2 AUC pe 4.4 2

The crosssectional area of a fuel element (AF) is calculated:

2 D

AF F 4.5 2

The area of the flow channel in the unit cell (AFC) is the difference between the unit cell area (eq. 4.1) and the area occupied by fuel (eq. 4.2). Since the interior radius is 1/2 of the pitch, the area of the flow channel is calculated by:

3

2 3 2 D AFC pe F 4.6 2 2 The wetted perimeter is the length of the flow channel in contact with channel wall surfaces (i.e., the perimeter of the fuel element):

PW DF 4.7 Noncircular pipes are approximated as a pipe with an equivalent hydraulic diameter (Dh) with a wetted perimeter (PW), where the hydraulic diameter is calculated as:

4 AFC Dh 4.8 PW Substituting equations 4.6 and 4.7 for flow area and perimeter into equation 4.8, the hydraulic diameter is given by:

4 3 2 DF 2 3 pe 2 2 3 pe2 Dh pe DF DF 2 1 4.9 DF 2 2 DF 2 DF A summary of primary and calculated parameters using the equations above is provided in Table 4.1.

Table 4.1, Summary of Principle Thermal Hydraulic Values Description Var. Value Fuel Element Pitch P 1.7149 in 0.1428 ft 4.3535 cm 0.04353 m Fuel Element Diameter Dfuel 1.4784 in 0.1232 ft 3.7551 cm 0.03755 m Wetted Perimeter PW 4.6445 in 0.3870 ft 11.7971 cm 0.1179 m Fuel Cross Section/Area AFC 1.7166 in2 0.01192 ft2 11.0749 cm2 0.001107 m2 Unit Cell Area ACell 2.5442 in2 0.01766 ft2 16.4142 cm2 0.001641 m2 Flow Channel Area AFC 0.8275 in2 0.005747 ft2 5.3392 cm2 0.000534 m2 Hydraulic Diameter Dh 0.7127 in. 0.05939 ft. 1.8103 cm 0.01810 m 4.2 UT TRIGA Thermal Hydraulic Model TRACE analysis is based on modeling a set of representative TRACE components with characteristics specified by the user to model the system. The UT TRIGA model uses Break, Pipe, Heat Structure, and Power components. These TRACE components were assembled as shown in Fig. 4.2 to model the thermal hydraulic performance of the unit cell flow channel.

4.2.1 Break

4

A break is a boundary component normally used as a sink for liquid flows exiting the system3, here simulating pool water entering and exiting the flow channel. Break pressure and temperature specifications are based on local environmental conditions (barometric pressure, confinement control), pool level, and water temperature. Pressure boundary conditions for limiting and nominal cases are provided in Table 4.2. UT TRIGA flow is calculated by TRACE as developed by convection in reactor operation.

Figure 4.2, TRACE Model The NETL building is approximately 240 m above sea level, corresponding to 96 kPa at standard atmospheric conditions. The reactor bay confinement system is designed to control differential pressure to 0.06 in. (14.9 Pa) below atmospheric (minimal compared to atmospheric pressure). Total pressure at the top of the core is therefore:

pT 96 KPa p H 2O 4.10 Pool water is a minimum of 5.25 m above the core, nominally 7.25 m. Constant pressure is established by setting the rate of change variable to zero in the break, and with a single value for pressure over RELAP time intervals. Pool water temperature is limited to less than 49 °C, nominally 2527 °C. Where g denotes the gravitational constant (9.8 ms2), the pressure (pH2O) exerted by a column of water (at density in kgm3 and height h in m) is given by:

p H 2O g h 4.11 Table 4.2, Pressure Boundary Condition Hydrostatic Temperature Density Height Pressure Pressure Condition Pressure

°C kgm3 m kPa kPa Psia Limiting 49 988.4881 5.25 50.9 146.9 21.3 25 997.0479 7.25 70.8 166.8 24.2 Nominal 27 996.5162 7.25 70.8 166.8 24.2 3

TRACE has a Fill component, but Break component calculates flow rate while Fill requires specifying a flow rate.

5

4.2.2 Pipe The pipe component is a cylindrical volume for water flow with userspecified geometric and hydrodynamic properties. One pipe (downcomer) represents movement of cooling flow from the inlet break to the bottom of the flow channel. A second pipe (connector) moves flow to the entrance of the flow channel and connects the down comer to the third pipe (unit cell flow channel). The third pipe discharges to the outlet break.

4.2.3 Downcomer/Cold Leg Conservation requirements for calculations require balanced elevation changes, with a downcomer at the same length and area as the fuel element region. Instabilities can occur in TRACE calculations if adjacent volumes are sufficiently different, and the downcomer is segmented to meet the ratio criteria (for convenience, segmenting has equal lengths). Dimensions for the downcomer pipe are provided in Table 4.3. The direction of flow is down.

Table 4.3, Downcomer Pipe Length (segments) 0.09985 m Length (total) 0.5991 m Flow area 5.39E4 m2 Volume (segments) 5.38E5 m3 Volume (total) 3.23E4 m3 Hydraulic diameter 0.0183 m Height Change (segments) 0.09985 m Height Change (total) 0.5991 m 4.2.4 Connector A pipe with two elbows (Fig. 4.3) connects flow from the downcomer to the unit cell flow channel.

Dimensions of the connecting pipe are provided in Table 4.4. Inlet flow is down, outlet flow is up, and the intermediate flow is horizontal.

Table 4.4, Connecting Pipe Flow Height Segment Volume Length Area Change m3 m m2 M 1 5.38E05 0.01 5.39E04 5.0E3 2 5.38E05 0.01 5.39E04 0.0 3 5.38E05 0.01 5.39E04 5.0E3 6

Figure 4.3, Cold Leg to Flow Channel Connector 4.2.5 Unit Cell Flow Channel/Fuel Element Region The flow channel for the fuel element region in the unit cell is modeled as a pipe with heated lengths connected to a heat structure. Specifications for the simulated fuel element cooling channel are provided in Table 4.5. The K factors are applied to the 2nd and the 19th segments. Flow is up.

Traditional K factors are discussed, followed by development of UT TRIGA specific K factors.

Table 4.5, Specifications for Unit Cell Flow Channel Segment Volume Length Flow Area z m3 m m2 M 1 5.14E06 0.01905 2.70E04 0.01905 2 2.43E05 0.09 2.70E04 0.09 3 6.86E06 0.0254 2.70E04 0.0254 4 6.86E06 0.0254 2.70E04 0.0254 5 6.86E06 0.0254 2.70E04 0.0254 6 6.86E06 0.0254 2.70E04 0.0254 7 6.86E06 0.0254 2.70E04 0.0254 8 6.86E06 0.0254 2.70E04 0.0254 9 6.86E06 0.0254 2.70E04 0.0254 10 6.86E06 0.0254 2.70E04 0.0254 11 6.86E06 0.0254 2.70E04 0.0254 12 6.86E06 0.0254 2.70E04 0.0254 13 6.86E06 0.0254 2.70E04 0.0254 14 6.86E06 0.0254 2.70E04 0.0254 15 6.86E06 0.0254 2.70E04 0.0254 16 6.86E06 0.0254 2.70E04 0.0254 17 6.86E06 0.0254 2.70E04 0.0254 18 2.43E05 0.09 2.70E04 0.09 19 5.14E06 0.01905 2.70E04 0.01905 Total 1.62E04 0.5991 5.13E03 0.5991 4.2.6 Headloss (K) factors Pressure drops (head loss) across hydraulic components are the product of the fluid flow and factors such as the coefficient of friction between the fluid and the pipe wall, changes in flow area and diameter, flow channel surface roughness, and/or flow channel length. Within limits, the factors (K factors) are constant, the sum of the pressure drops in linear flow is additive. This analysis provides a traditional approach to evaluating the loss factors and loss factors reported by analysis and 7

experiments conducted at the UT reactor, followed by the results of analysis and experiments conducted specific to the UTTRIGA facility.

4.2.7 Classic K factors The impact of sudden expansion or contraction is principally in velocity changes. Traditional K factors for sudden expansions or contractions are based on the ratio of inlet and outlet flow areas (Table 4.6, Equation 4.12).

d2 A K e 1 12 1 1 4.12 d 2 A2 Other K factors are based on the magnitude of the direction change, the pipe surface roughness, and flow mode (turbulent, laminar, etc.). Calculations are simplified by using the DarcyWeisback friction factor (f) as a multiplier on applicable aspects of system geometry. The friction factor is a function of the Reynolds number, wall surface roughness, and flow channel. The relationship is described in the Colebrook formula:

1 2.51 2.0 log10 4.13 f 3.7 D Re f In practice, the Moody chart is frequently used to determine the friction factor. For reasonable and expected flow rates at the TRIGA reactor, the Reynolds number is between 1X104 and 3X105. Over this range, convergence exists for wall surface roughness values between 5X107 to 1X103. The broad range of surface roughness values indicates a very low sensitivity for roughness, and that any surface roughness within this range can be used without affecting the friction factor significantly. For comparison, RELAP analysis conducted for DOW Chemical4 reactor used surface roughness of 2.13E 6.

For losses in a straight pipe:

L K f 4.14 D

For a 45° turn:

K 45 f 16 4.15 For a 90° turn:

K90 f 30 4.16 Table 4.6:

Location Component Eff. Area Bottom Entrance Lower grid plate 1.2 cm2 Bottom Exit Lower End fitting/Channel 3.9 cm2 Top Entrance Upper End Fitting/ Channel 3.9 cm2 Top Exit Upper Grid Plate 1.2 cm2 4

ANALYSIS OF THE THERMAL HYDRAULIC AND REACTIVITY INSERTION BEHAVIOR OF THE DOW TRIGA RESEARCH REACTOR, Submitted to the NRC in support of the DTRR License Renewal (M. H. Hartman, 03/12/2011).

8

The K factor for elevations above the flow channel is based on a 45° turn out of the main channel and sudden contraction at the upper grid plate. The K factor below the flow channel is based on a sudden expansion exiting the grid plate and a 45° turn into the main channel.

4.2.8 UT TRIGA Specific K factors Correlation of K factors to flow are based on historical, experimental measurements with cylindrical pipes, with additional work validating this approach for rectangular ducts. In practice, noncircular cross sections are reduced to a flow area and a hydraulic diameter, with length as measured for the pipe. However, the complexity of the TRIGA inlet and exit flow channel geometry is challenging. As fluid interacts with noncircular structures (or components), nonuniform surfaces can result in forces leading to secondary and/or internal flow paths that affect head loss/pressure drops. This suggests two potential issues using K factors calculated classically in analyzing thermal hydraulic response of the TRIGA reactor:

The actual entrance and exit to the flow channel between the grid plates is directed by fins mounted on a conical shape that terminates in cylindrical alignment (bottom end fitting) and handling (upper end fitting) structures. The wetted perimeters and flow areas vary continuously from entrance and exit for each end fitting.

The interface between adjacent fuel channels is not separated by a physical boundary.

Differential pressure between adjacent flow channels at interfaces can support crosschannel flow.

Therefore, thermal hydraulic analysis to support relicensing was developed5 to:

(1) Model the UT TRIGA reactor using TRACE (2) Develop an independent solution tool using MATLAB to calculate thermal hydraulic performance based on mass and energy balance and K factors, (3) Develop a computational fluid dynamics model using FLUENT, and (4) Conduct experiments to develop a UT TRIGA specific heat transfer correlation.

These methodologies were used to independently model thermal hydraulic performance from (1) first principles, (2) TRACE thermal hydraulics code, and (3) FLUENT computational fluid dynamics code.

The results of experiments in the TRIGA core were used to evaluate UT TRIGAspecific K factors based on actual fuel element geometry. A summary of K values determined from both the traditional/classical method and the UT FLUENT analysis is provided in Table 4.7, with a fractional deviation between factors provided. For comparison, RELAP work6 performed for DOW Chemical facility used K factors of 2.26 and 0.63 for the lower and upper channels respectively. The values determined from the UT FLUENT analysis were used in modeling for TRACE calculations.

Table 4.7, K Factors 5

Development of Thermal Hydraulic Correlations for the University of Texas at Austin TRIGA Reactor Using Computational Fluid Dynamics and InCore Measurements, A. D. Brand 6

ANALYSIS OF THE THERMAL HYDRAULIC AND REACTIVITY INSERTION BEHAVIOR OF THE DOW TRIGA RESEARCH REACTOR, Submitted to the NRC in support of the DTRR License Renewal (M. H. Hartman, 03/12/2011).

9

APPLICATION CLASSICAL FLUENT7 DEVIATION Lower Channel 1.244 1.63 23.7%

Upper Channel 0.844 1.12 33.6%

4.2.9 Heat Structure TRACE defines heat structures as rigid components that absorb, transfer, or radiate heat. The heat structure is specified by geometry, inner and outer radial boundary conditions, and material information. Boundary conditions for heat transfer are specified for axial nodes/surfaces, linking the heat source to the heated lengths of the pipe to represent the active (fueled) part of the fuel element.

Heat structure cells simulate the zirconium fill rod at the center of the fuel element, ZrHU fuel, the gap between the fuel and cladding, and the cladding. Heat structure material properties are used to calculate temperature distribution for fuel element components (zirconium fill rod, UZrH matrix, gas gap, and cladding). The material in the UT TRIGA model includes:

Zirconium from a radius of 0 cm to 0.3175 cm (3.175E3 m)

Zirconiumhydride from a radius of 0.3175 cm to 1.74117 cm (0.0174117 m), subdivided into 15 segments, 13 equal volume segments with one segment boundary at thermocouple location Gap gases from a radius of 1.74117 cm to 1.8161 cm (0.018161 m)

Stainless steel 304 cladding from a radius of 1.8161 cm to 1.8671 cm (0.018671 m)

The effects of core reactivity on the hot channel are simulated by defining a heat structure for the flow channel. When this option is selected it is necessary to set initial conditions for the average fuel element and implement the hot channel power peaking factor from the Neutronics Analysis.

4.2.10 GasGap Heat Transfer Coefficient There is a small difference in the outer radius of the fuel element matrix and the inner radius of the fuel element cladding. This annulus contains hydrogen and fission product gases in a balance between evolution from and reabsorption into fuel matrix. The heat transfer coefficient (HTC) of the gap is a complicated function of geometry, surface roughness at solid to gaseous interfaces, differential pressure, and constituent gas properties8, all of which are variable with temperature and unknown.

Correlations for power reactors are built into TRACE, but for applications other than power reactors only a single HTC value can be used in a calculation.

The UT TRIGA reactor has two fuel elements instrumented with thermocouples that monitor fuel temperature in positions that produce power levels at or near the hot channel power. The construction of the instrumented elements varies from standard fuel elements with (small) channels for thermocouple leads and (small) penetrations for the thermocouples. The instrumented elements are designed to be representative of standard fuel elements for initiation of protective action. The fuel mass is roughly 97% of the mass of a standard fuel element, and the gap geometry of the instrumented element is therefore similar to the gap of standard fuel elements.

8 Elements of Nuclear Reactor Design, 2nd Edition (J. Weisman) & TRACE V5 Theory Manual, pages 438441 10

4.2.11 Power Component The power component includes fundamental nuclear data such as fissile isotope Q values and fission fractions, delayed neutron data, decay heat model, power distribution, reactivity coefficients, and timebased profiles for power or reactivity in transient problems. The ANSI/ANS5.12014 standard (Decay Heat Power in Light Water Reactors) was used as the decay heat model.

Nuclear data was provided by the MCNP calculations in the Neutronics Report. The ratio of decay heat power to initial reactor power depends on a reactor specific energy generation per fission (MeV/fission) used in TRACE transient calculations. The TRACE point reactor kinetics treatment uses fission product nuclear characteristics (precursor fractions, decay constants, and the generation time) from MCNP adjoint calculations for the current core configuration (i.e., in validation) and the LCC. The MCNP burnup analysis output supporting the Neutronics Report tabulates fission energy yield data (Q value) as given in Table 4.8.

Table 4.8: Fission Energy Yield from MCNP Analysis Nuclide QValue (MeV) 92235 180.99 92238 181.31 94239 189.44 94241 189.99 The fraction of energy produced by each fissionable material (Table 4.9) for TRACE analysis is taken from the MCNP burnup calculations. The estimate of the fraction of isotope 92235 fissions at greater than thermal energy is assumed to have a Q value consistent with isotope 92238.

Table 4.9: Fission Isotope Nuclear Characteristics9 Fission Fissions at Fraction Nuclide Crosssection Energy Range Energy (b) 585.1 <0.625 eV 94.28%

92235 19.8%

274.4 0.625 ev - 100 kev 4.99%

1.241 92238 80.2%

0.03064 >100 kev 0.72%

MCNP adjoint calculations provided estimates of the prompt neutron generation time and effective delayed neutron fractions. The promptneutron generation time was reported as 43.81+/-0.53 s. The 1992 Safety Analysis Report for the UT TRIGA cited a prompt generation time of 41 s, which is reasonably consistent with that calculated by MCNP. The MCNP calculated value utilizing current parameters is used in this report.

An MCNP calculation incorporating the KOPTS option was used to provide delayed neutron precursor data (Table 4.10) for the current core described in the Neutronics report.

Table 4.10: Nominal Core Delayed Neutron Precursor Group Characteristics 9

https://wwwndc.jaea.go.jp; Fission at Energy from MCNP burnup calculation 11

Standard Energy Standard i Standard T1/2 Group i Deviation (MeV) Deviation (s1) Deviation (s) 1 0.00024 0.00004 0.41037 0.00353 0.01334 0.00000 51.974 2 0.00117 0.00009 0.47158 0.00158 0.03271 0.00000 21.189 3 0.00098 0.00007 0.44336 0.00159 0.12074 0.00000 5.741 4 0.00287 0.00013 0.55738 0.00142 0.30288 0.00001 2.288 5 0.00116 0.00009 0.51857 0.00230 0.85033 0.00003 0.815 6 0.00047 0.00006 0.54495 0.00392 2.85505 0.00020 0.232 The time dependent behavior of power following shutdown was evaluated and developed using reactor kinetics for fission power and the method of ANSI/ANS5.12014 (equations 7 and 8) for decay heat from fission products. For an irradiation interval of time T, a decay time of t, and irradiation intervals i and a constant fission rate of unity, decay heat power (F, units of MeV/fission) is represented analytically as a function (equation 4.17) using analytic fit constants and (provided for all 23 components in the standard):

, exp , 1 ,

, 4.17 Assuming a single continuous power generation interval and a long operating time, equation 4.17 reduces to equation 4.18:

exp 4.18 Power distribution based on the results of the Neutronics Report MCNP model and power distribution in TRACE is implemented as the fraction of power generated between the inner and outer radial boundaries at each axial node. The power distribution in the UT TRACE model is a 2 dimensional, radial and axial, distribution based on the Neutronics Report MCNP model.

Steady state calculations were performed to establish operational characteristics for evaluating the critical heat flux ratio and also to provide initial conditions for some transient calculations. Transient power calculations were performed to establish initial conditions for loss of cooling analysis.

Transient reactivity calculations were performed for pulsing from low power, pulsing from power above the point of sensible heat production, continuous reactivity additions, and a fuel element discharged from the core and air cooled.

4.2.12 Materials TRACE has a limited set of material characteristics applicable to nuclear power plants, with only gap gases and stainless steel 304 applicable to TRIGA reactors. The default set of materials were user augmented with zirconium and uranium zirconium. User defined materials are defined in a data table specified over a range of temperature, and include (1) density, (2) specific heat capacity, (3) thermal conductivity, and (4) emissivity.

12

The thermal conductivity of TRIGA fuel is noted to be 0.042 cals1cm1°C1 (17.573 Wm1°K1) 10, and is assumed to be insensitive to temperature. The volumetric heat capacity (Cv, referenced to temperature T in °C) was calculated using:

2.04 4.1710 4.19 Specific heat capacity is calculated by normalizing the volumetric heat capacity to density (), with the density of the fuel matrix. The density of ZrH for hydrogen to zirconium ratio (R) of 1.6 or greater is given by:

1 0.1706 0.0042 4.20 The uranium density is 19.07 g/cm3. The weight per cent for the two components is indicated by subscripts, U for Uranium ZrH for ZirconiumHydride so that the fuel matrix density is:

1

, . 4.21 Since the fuel heat capacity is a linear function with respect to temperature, only two values that bound the calculations were used (Table 4.8).

Thermal conductivity for the zirconium fill rod at the center of the fuel element was taken (even 100 temperature values) from the Journal of Physical and Chemical reference Data (Volume 3, 1974, Supplement 1, Table 184), with intermittent values interpolated. Volumetric heat capacity data was taken from a compilation11, with data interpolated by a curve fit. Massspecific heat capacity used in TRACE is calculated as the ratio of the volumetric heat capacity to the density. Zirconium and uraniumzirconiumhydride data is provided in Table 4.11.

Table 4.11, User Supplied Material Data Material T Cp Conductivity Emissivity

°K kgm3 Wskg1K1 Wm1K1 UraniumZirconiumHydride 293.15 5998.595 353.88 18.01 0.8 3033.15 5998.595 3815.01 18.01 0.8 Zirconium 200.15 5256.94 344.276 25.19226 0.8 400.15 5256.94 427.5999 21.59248 0.8 600.15 5256.94 510.9237 20.68942 0.8 800.15 5256.94 594.2475 21.59248 0.8 1000.15 5256.94 677.5714 23.69131 0.8 1200.15 5256.94 760.8952 25.98944 0.8 1500.15 5256.94 885.881 28.78582 0.8 2744.928 5256.94 1404.479 37.43582 0.8 10 Simnad, The UZrHx Alloy: Its Properties and Use in TRIGA Fuel (August 1980) 11 http://www.efunda.com 13

4.2.13 Temperature Coefficients of Reactivity The Fuel Temperature Coefficient of Reactivity (FTC) was developed from the Neutronics Reports MCNP model with crosssections at temperatures taken from the distributed libraries for scattering data and isotopes for comparison to the FTC from General Atomics. There are 8 sets of scattering data and 4 sets of isotope cross sections corresponding to the scattering data within MCNP. Where scattering data at specific temperatures do not have corresponding scattering cross section data, an auxiliary program (MAKXSF) was used to develop interpolated cross sections at scattering data temperatures. Calculations using the fuel temperature data were performed at three water temperatures: nominal operating temperature, maximum permitted pool temperature, and a value approximately halfway between the two. The values for keff were fit to a curve (first varying the fuel temperature, then varying the moderator temperature), and the formulae used to generate data 1/2 degree above and 1/2 degree below the temperature of interest. The FTC, reactivity associated with a 1 degree change in temperature, was calculated as:

, 1/2 , 1/2

, 1/2 , 1/2 4.22 The moderator temperature coefficient was very small compared to the FTC, and the difference between the keff values at each moderator temperature was therefore not significant to the FTC. The MCNPbased FTC is shown in Figure 4.4. Also, shown in the figure is historical GA data. General Atomics12 indicates that the fuel temperature coefficient (water reflected) is convex, with a minimum occurring about 300°C. The MCNPbased FTC for the UT TRIGA agrees well with the GA data above 200oC. At low temperatures the MCNPbased FTC deviates from the General Atomics FTC, but the general shapes are similar. The distribution of cross sections temperatures is too broad to reflect the changes when keff changes quickly with respect to temperature, which affects the FTC calculation. There is a decrease to a minimum value of FTC and then nearly a linear increase above the minimum FTC for both MCNPbased and the General Atomics provided data (Table 4.12 and Fig. 4.4). The moderator temperature coefficient (Table 4.13 and Fig. 4.5) was also developed from the MCNP data.

Figure 4.4: MCNPBased FTC and GA FTC 12 Simnad op. cit.

14

Table 4.12, MCNPBased Fuel Temperature Coefficient Fuel Temperature FTC

°C k/°C 26.85 4.96E08 126.85 5.41E05 226.85 1.01E04 326.85 1.14E04 426.85 1.11E04 526.85 1.05E04 726.85 8.89E05 926.85 7.50E05 Table 4.13, Moderator Temperature Coefficient (k/°C)

Fuel Temperature 24 °C 427 °C 927 °C Moderator 24 °C 1.65E0 1.76E5 1.93E5 (Pool) 51 °C 5.47E0 5.84E6 6.44E6 Temperature 101 °C 1.75E0 1.87E5 2.05E5 Figure 4.5: Moderator Temperature Coefficient Analysis at AFRRI13, based on DIF3D (Argonne National Laboratory, diffusion, and transport theory code) shows convex fuel temperature reactivity structure from 101000°C for TRIGA fuel. Calculations based on DIF3D indicate fuel coefficient reactivity range from 8x105 to 1.2x104 k/k °C1 for a TRIGA core with a circular grid plate (the UT TRIGA has a hexagonal pitch). The values in Fig. 4.4 are in reasonable agreement with this analysis.

13 AFRRI op. cit.

15

4.2.14 Critical Heat Flux Ratio Critical heat flux ratio (CHFR) is the ratio of the critical heat flux (CHF) to the actual heat flux. ANL/RERTR/TM 0701 provides a series of calculations with different correlations for critical heat flux, including the Bernath correlation. The correlation for critical heat flux developed by Bernath is recommended by the reference in evaluating TRIGA fuel performance. The Bernath correlation (where CHFBO is the heat flux that results in burnout hBO is the convection heat transfer correlation at burnout, TW,BO is the temperature of the cladding surface at burnout, and V is the fluid velocity, Tb is the cooling water bulk temperature, and dimensional variables as described in Table 4.14) determines the critical heat flux that results in burnout as:

CHFBO hBO TW ,BO Tb 4.23 where the heat transfer coefficient for burnout conditions is calculated using:

De 48 hBO 10890 V 0.6 4.24 De Di De The formula predicting wall temperature at burnout is:

P V TW ,BO 57 ln P 54 4.25 P 15 4 Substituting equations 4.23 and 4.24 into equation 4.25 results in:

De 48 P V CHFBO 10890 V 0.6 57 ln P 54 Tb 4.26 De Di De P 15 4 The Bernath formulation is in pound centigrade units, converted to BTU h1 ft2 by multiplying by a factor of 1.8:

De 48 P V WCHF 1.8 10890 V 0.6 57 ln P 54 Tb 4.27 De Di De P 15 4 Table 4.14, Bernath Correlation Variables De Hydraulic diameter Ft Previous formula Di Heater surface diameter Ft Fuel element diameter3 V Pressure Psia Calculated by TRACE 1

P Velocity fts Calculated by TRACE TB Coolant Temperature °C Calculated by TRACE 5.0 Model Validation The model described above was validated by comparison of model results to measured parameters from the UT TRIGA reactor. Nominal values indicated in Table 5.1 are used in validation.

16

Table 5.1, Model Input Values Description Source or Value Pressure and Temperature Table 2.4 KFactors (Fluent) Table 4.7 Fission Energy Yields Table 4.10 Thermal and Fast Fission Fractions Table 4.11 Delayed Neutron Precursor Data Table 4.12 User Supplied Material Data Table 4.13 Fuel Temperature Coefficient Table 4.14 Moderator Temperature Coefficient Table 4.15 Prompt Neutron Generation Time 47 s14 Core Radial Peaking Factor 1.6815 Wall Surface roughness 6.998032E6 5.1 Operating Data The Integrated Control System (ICS) reports peak power and peak temperature from a thermocouple in an Instrumented Fuel Element during normal and pulsing operations. Measured fuel temperature channel indications at varying power level are compared to thermocouple location temperatures calculated by TRACE. There is no actual measurement of heat fluxes available for comparison with calculations of critical heat flux ratio, CHFR (the ratio of fuel element local heat flux to the heat flux that could result in departure form nucleate boiling). During pulsing, core peak power is reported. TRACE calculations of an average fuel element were compared to ICS data (peak power distributed over all elements in the core) and the TRACE calculated temperature at a location similar to the IFE thermocouple location is compared to ICS data.

5.2 Temperature Instrumentation Instrumented fuel elements (IFEs) are located in the B ring. Three thermocouples are installed in each IFE, with one thermocouple in each IFE normally instrumented. The thermocouples are located 0.762 cm from the center of the fuel element, with one an inch above the midplane, a second at the midplane, and a third an inch below the midplane. TRACE calculations show the thermocouple response varies only slightly between the three elevations.

5.3 SteadyState Validation The temperatures of the thermocouple locations used in the measuring channel were calculated using the TRACE model during steadystate operation. One thermocouple was replaced in February 2016.

Consequently, the baseline temperature data for validation is taken from December 2015 through February 2016 with IFE load configurations as specified in Table 5.2.

The Neutroincs Report was developed using an MCNP burncard to generate an estimate of the material composition for each fuel element in the TRIGA core at specified burn intervals. The burn intervals correlate to specific dates when core inventory was altered. Fuel element material specifications for the burn interval (corresponding to the date of the validation data in Table 5.2) were used with the Neutroincs Report MCNP model to develop peaking factors for the IFEs. The peaking factors and power levels from reactor operating records were used to calculate IFE power levels for TRACE calculations. The temperatures corresponding to 14 Neutronics Report 15 Derived from Neutronics Report, Figure 7, Current Core PowerPerElement (in kW) Distribution at 1.1 MW 17

the location of the thermocouples in the IFEs were compared to operating data (Table 5.2). Comparison of the calculated and observed data indicates TRACE can predict steadystate behavior with reasonable accuracy (within +/-7.8% of measured temperature values).

Table 5.2, TRACE CALCULATED AND IFE MEASURED FUEL TEMPERATURE COMPARISON IFE TRACE  % DIFFERENCE FUEL ELEMENT INDICATED CALCULATED BETWEEN POSITION ELEMENT POWER (kW) FUEL TEMP FUEL TEMP MEASURED AND (oC) (oC) CALCULATED INITIAL CONFIGURATION (12/18/2015)

B03 10878 13.24 325 345 5.80%

B06 10708 13.61 364 354 2.82%

IFE 10708 REPLACED WITH IFE 10809 (02/01/2016)

B03 10878 13.30 319 346 7.80%

B06 10809 15.30 420 394 6.60%

5.4 Pulsing Validation There were over 300 pulses conducted at the UT TRIGA between initial criticality and 2016. Records of pulse data includes operatorcalculated reactivity addition, control rod positions, core configuration (i.e.,

indicating the number of fuel elements in the core), peak power, total pulse energy, and peak temperature from the fuel temperature measuring channel. Pulses with reactivity values approaching $1.00 have very large pulse widths, and assumption of adiabatic pulsing and a transient much smaller than some of the delayed neutron precursors decay constants may be affected by heat transfer.

The TRACE calculations were compared to historical data to validate the accuracy of the method used.

Historical pulse data (reactivity addition, peak pulse power, and maximum temperature from the fuel temperature measuring channels) was compiled, with incomplete data purged. Figures 5.1 and 5.2 show the TRACE calculated power level and temperature compared to the historical data. While there is significant scatter in power level and temperature data with some outliers, the results overall show excellent agreement and provide a basis for validation.

18

Figure 5.1: Peak Element Power Level Versus Pulse Reactivity Addition from UT TRACE Calculation Compared to Observed Historical Data Figure 5.2: Peak Fuel Temperature Versus Pulse Reactivity Addition from UT TRACE Calculation Compared to Historical Data 19

Pulse records do not include factors that affect pulse characteristics such as initial fuel temperature, pool temperature, or recent operating history which might explain some of the scatter in the data. Although there is significant scatter and outliers in historical pulse power level data, it is clear that qualitatively the TRACE data agrees well with historical data. This indicates that TRACE can predict transient behavior with reasonable accuracy.

Average fuel element power and peak fuel temperature indicated from the IFE thermocouple location were compared to TRACE simulated values for pulses from $1.25 to $3.00 in $0.25 increments. The results are shown in Table 5.3. As noted, the results show good agreement especially for pulses above $2.00 in reactivity addition.

Table 5.3: Comparison of Observed Pulse Data with TRACE Simulation Pulsed Deviation in Deviation 114 Element Core TRACE Simulation Reactivity Temperature in Power k($) FT(°C) FT(°K) POWER (W) TC (°K) POWER (W) (°K) (%)

$1.25 158 431.32 4.40E+05 410.58 3.49E+05 20.75 26.04%

$1.50 200 472.79 1.39E+06 460.50 1.23E+06 12.30 12.73%

$2.00 283 555.73 5.17E+06 547.79 4.76E+06 7.94 8.50%

$2.50 366 638.66 1.15E+07 627.83 1.08E+07 10.83 6.12%

$2.75 407 680.13 1.56E+07 668.99 1.49E+07 11.14 4.77%

$3.00 448 721.60 2.03E+07 716.87 1.95E+07 4.73 4.18%

5.5 Conclusion on Model Validation Fuel temperatures indicated by the fuel temperature measuring channels, and power and temperature from historical records of pulsing, were compared to data generated with TRACE calculations. The comparison demonstrates that the TRACE model predicts thermal hydraulic performance of the UT TRIGA reactor with reasonable accuracy.

6.0 Results A limiting case and a nominal case were defined for analysis of the fuel element producing the maximum power in the core, based on Technical Specifications and parameters as identified in Table 4.2. The maximum peaking factor for the minimum number of fuel elements in the core was calculated in the Neutronics Report and used to identify and characterize the fuel element that produces the maximum power (and therefore the maximum power density). Table 5.1 values are used in LCC calculations except for the LCC Specific Values identified in Table 6.1 The maximum peaking factor for the LCC hot channel, i.e., channel producing the most power, is calculated as the ratio of the hot channel power to the average of the power of all elements based on the Neutronics Report (Figure 12 - Limiting Core Configuration PowerPerElement Distribution at 1.1 MW). Reactor power is assumed to be 1.21 MW, the maximum licensed power with maximum instrument error identified in the 1992 Technical Specifications of 10%. The hot channel power is calculated using the peaking factor applied to the core power distributed over the number of fuel elements.

20

Table 6.1, LCC Specific Values Physics Factors Value Prompt Neutron Generation Time 44 s Core Radial Peaking Factor 1.68 Number of Fuel Elements16 84 Maximum Hot Channel Power 24.34 kW Pool Water Conditions Value Temperature 120.2 °C Pressure 23.1 psia Calculations of temperature across the fuel element were performed by TRACE in hot channel analyses. The results of temperature calculations are used to demonstrate that the maximum hot channel fuel temperature in limiting core conditions will remain less than 830°C for pulsing operations, 950°C if cladding is greater than or equal to 500°C, and 1150°C if cladding temperature is less than 500°C. The results of thermodynamic analysis for the hot channel were used in the Bernath correlation to demonstrate that the heat flux for the limiting and nominal cases will not result in departure from nucleate boiling for the hot channel.

6.1 Critical Heat Flux TRACE was used to calculate heat transfer, water temperature, water density, pressure, and mass flow rate at 15 elevations in the heated length of the flow channel across a range of fuel element power levels (Table 6.2).

These values were used to calculate the critical heat flux using the Bernath correlation (Eq. 4.27) with results illustrated in Fig. 6.1. The coolant mass flow rate is shown in Fig. 6.2. The radial and axial maximum fuel temperatures for the fuel element operating at the maximum power level are shown in Figs. 6.3 and 6.4, respectively.

Table 6.2, Steady State Calculations Max Fuel Power CHFRmin Temp kW °C 11 5.82 367 14 4.45 414 17 3.56 459 20 2.94 502 23 2.48 559 24 2.39 559 24.34 2.37 564 27 2.43 599 30 2.26 639 16 Includes standard fuel elements, FFCRs and IFEs.

21

Figure 6.1: Critical Heat Flux, Bernath Correlation Figure 6.2: Coolant Flow Rate at Element Power 22

Figure 6.3, Hot Channel Radial Temperature Distribution Figure 6.4: Hot Channel Axial Temperature Distribution Limiting the critical heat flux ratio (CHFR) in the hot channel assures that departure from nucleate boiling will not occur. Therefore, CHFR minimum of 2.0 is not achieved at element power levels less than 30 kW, and operation at a maximum element power level of 24.34 kW has a sufficient margin to the 2.0 limit.

6.2 Pulsing from Low Power Simulation of pulsing from shutdown power levels was performed. On initiation of a pulse, the ICS shifts instrumentation to a pulse monitoring mode for about 15 seconds. Pulsing reactivity calculations were performed in TRACE for the LCC hot channel using $1, $2, $3, and $4 insertions with initial conditions approximating shutdown power levels. The peak power as a function of pulsed reactivity is shown in Fig. 6.5 and the maximum fuel temperature is shown in Fig. 6.6.

23

Figure 6.5: Hot Channel LCC Peak Power Level Versus Pulsed Reactivity Insertion of $1, $2, $3 and $4 Figure 6.6: LCC Peak Fuel Temperature for Varying Reactivity Insertions A $3 insertion from low power levels (i.e., that do not generate sensible heat) resulted in a peak temperature of 530°C, with significant margin below the pulsing temperature limit. As shown in Fig. 6.6, the temperature limit was not exceeded for any reactivity insertions below $4.

To study the pulsing limits in more detail, pulses from $3 to $4.40 were simulated in TRACE with the results given in Table 6.3. The pulse power, maximum fuel temperature, and fuel temperature at 3 locations in the fuel versus time following the pulse is shown in Fig. 6.7, 6.8, and 6.9, respectively.

24

Table 6.3: Pulse Response to 15 s Pulsed k Peak Power Peak Temp.

$ W °C

$3.00 3.43E+07 530

$3.50 5.51E+07 610

$4.20 9.43E+07 778

$4.30 1.01E+08 795

$4.40 1.08E+08 824 For a maximum pulse of $3.00, the peak maximum fuel temperature is 530°C occurring near the thermocouple position about 13 seconds after the pulse. Temperature decreases after the peak.

Evolution of temperature is shown at the position adjacent to a thermocouple (position 4), an intermediate position (position 13) and position near the fuel outer surface (position 16) in Fig. 6.7. A pulse of $4.40 reaches 824°C in 15 seconds (Table 6.3) and is on an increasing trend.

Figure 6.7: Peak Hot Channel Power Performance in Pulsing 25

Figure 6.8: Maximum Hot Channel Temperature Performance in Pulsing Figure 6.9: Evolution of Temperature for a $3.00 Pulse, 3 Locations Since pulsing to $4.40 does not result in exceeding the temperature limit for pulsing, pulsing to $3.00 will assure the fuel temperature remains below the 830°C limit by a large margin.

6.3 Pulsing from Power Pulsing from power was simulated by a transient reactivity, with the reactivity applied to establish initial conditions for 200 seconds followed by the addition of $3.00 of additional reactivity at 200 seconds. With initial element power level of 1.33 kW (corresponding to core power of 111 kW as indicated in Table 6.4, $0.55 above the cold clean critical position), a $3.00 pulse results in:

maximum hot channel fuel temperature of 724 °C and final steady state power level of 24.02 kW.

Pulsing from an initial element power level of 2.08 kW (corresponding to core power of 174 kW, $0.80 above the cold clean critical position), a $3.00 pulse approaches the temperature limit, resulting in:

maximum hot channel temperature of 826 °C and final steady state power level of 28.8 kW.

Table 6.4, Pulsing from Power Summary Init Core Power (kW) 111 kW 124 kW 174 kW Initial Ave. Element Power (kW) 1.33 kW 1.47 kW 2.08 kW Final Element Power (kW) 24.02 kW 26.92 kW 28.80 kW Final Core Power (kW) 1193 kW 1216 kW 1337 kW Max Hot Channel Temperature (°C) 724 °C 747 °C 826 °C Results for two cases approaching limits are shown in Fig. 6.10. Figure 6.10 shows the response for the average 26

fuel element (the bases for establishing core power level) and the hot channel peak temperature for the immediate response through the peak temperatures for initial core power levels of 111 kW and 174 kW.

Figure 6.11 shows the longterm temperature response for initial core power of 111 kW, 124 kW (where the final steadystate power level will be 0.5% higher than the licensed power limit for steady state power operation), and 174 kW. In all cases, the peak temperatures occur 3035 seconds following the pulse.

Tmp Figure 6.10, Responses to $3.00 Pulse from Power within Maximum Pulse Temperature Limit (Transient) and Licensed Power Level (Steady State) 27

Figure 6.11, Temperatures Following Pulses from Power If a pulse is initiated at 111 kW and allowed to come to equilibrium, the maximum hot channel power level is slightly less than the power for operation at the licensed power limit for steadystate power operation. If a pulse is initiated at 174 kW core power and allowed to come to equilibrium, although the pulsing temperature limit is met, the power level will be greater than the power for operation at the licensed power limit for steady state power operation. At 30.0 kW, CHFR has been shown to be greater than 2.0. However, a control rod interlock prevents pulsing if power level is greater than 1 kW. A TRACE calculation was performed with pulsing to $3.00 initiated from operations with core power at 1.021 kW (3.95x104 k/k compensating for temperature at power). The maximum hot channel fuel temperature for this transient is 522°C. A control rod inhibit interlock that prevents pulsing at greater than 1 kW is adequate by a large margin to ensure pulsing from power does not exceed the pulsing safety limit.

6.4 Analysis of Continuous Reactivity Addition from Full Power An analysis of reactivity insertion at power was accomplished in three steps: (1) identification of the reactivity required to support full power operation at steadystate conditions (i.e., average fuel element power multiplied by 84 elements), (2) identify the time that the average fuel element operating at the reactivity in item 1 followed by a continuous reactivity addition reaches the scram setpoint, and (3) hot channel temperature response to a scram at varying continuous reactivity insertion rates, varying delays for scram initiation after reaching the LSSS, and onesecond from initiation of scram to full insertion.

Reactivity insertions were varied to establish initial conditions of average element power corresponding to maximum licensed core power at the maximum error assumed in the 1992 Safety Analysis Report. A reactivity of $3.72 (0.0260 k/k) resulted in a steadystate core power of 1.258 MW, slightly higher than the nominal value, at 585 seconds of operation.

Calculations were performed for reactivity additions from the operating condition at various rates with varying scramtime delays (before the scram occurs after power reaches the scram setpoint). The minimum value for the reactivity addition rate is based on the 1992 Technical Specification for maximum control rod reactivity insertion rate (0.2% k/k per second) up to 0.7% k/k per second. The minimum value for scram response times is based on the 1992 Technical Specification value for full insertion no more than 1 second after initiation of the scram up to a maximum of 5 seconds. The shutdown reactivity values ($8.04, 28

0.05628 k/k) were based on Neutronics Report values for the worth of all control rods ($14.97) and the core excess reactivity ($6.93). The 5 second delay at reactivity addition rates of 0.6% and 0.7% k/k per second resulted in fatal run time errors.

Table 6.5, PEAK TEMPERATURE FOLLOWING CONTROL ROD FULLINSERT DELAY Reactivity addition rate 0.2%/s 0.4%/s 0.5%/s 0.6%/s 0.7%/s Delay (seconds) Tmax (°C) 1 573 589 608 627 651 2 585 639 679 726 778 3 609 709 773 863 993 4 630 772 878 1050 1448 5 634 800 992 N/A N/A The maximum hot channel temperature during the transient was 573°C for limiting conditions of the 1992 Technical Specifications. This is a conservative calculation, using an initial power level marginally higher than the maximum license power level with the maximum instrument error and a reactivity addition rate at the 1992 license limiting reactivity addition rate for control rods. A 4second delay in scram initiation after power reaches 1.1 MW and does not exceed the safety limit. Therefore, a maximum scram setpoint of 1.1 MW is adequate to prevent exceeding the safety limit for an event where a continuous reactivity addition occurs while operating at full power with the maximum power level instrument error and a maximum reactivity rate of 0.2% k/ks1 with a maximum of 1 second for full insertion of control rods.

Figure 6.12, Response to 0.2% k/k per Second Reactivity Addition at Full Power, Full Control Rod Insertion 1 Second after 1.1 MW 29

6.5 Loss of Coolant Event (LOCA)

The LOCA analysis was a 2step process with first a TRACE calculation to establish initial conditions at 25 kW per element with a series of shutdown and decay intervals with water cooling. This was followed by a TRACE restart case initiated as a transient calculation with air cooling. Air temperature was assumed to be 77°F. The method of ANSI/ANS5.12014, Decay Heat Power in Light Water Reactors was used to evaluate fission and fission product power decay in time. Four cases were calculated (Table 6.6) using four intervals between shutdown and instantaneous replacement of water cooing with air cooling. The timedependent behavior is shown in Fig. 6.15 over four hours following shutdown. The maximum cladding temperature occurred with 1 s delay before air cooling, and was 784°C. Therefore, on a loss of cooling event following steadystate operation at 25 kW per element, the maximum fuel temperature remains at acceptable levels.

Table 6.6, LOSS OF WATERCOOLING ANALYSIS DELAY FOR AIR MAXIMUM COOLING (s) TEMPERATURE (°C) 1 787 60 780 600 753 1200 733 Figure 6.13, Maximum Fuel Temperature with Loss of Cooling Following SteadyState Operation at 25 kW Per Element 7.0 Summary The minimum LCC is 84 elements for steadystate operations at 1210 kW (the licensed limit with maximum measuring channel error). A summary of the results from the thermalhydraulics analysis for the LCC is given in Table 7.1. The following conclusions are made:

1. Power levels up to 1210 kW demonstrate that a minimum CHFR of 2.0 is assured in the limiting core configuration (minimum pool water level, maximum pool temperature).

2. Pulsing from shutdown to $3.00 will remain within the pulse temperature limit. Pulsing to $3.00 from operation at power levels up to 174 kW will result in maximum fuel temperatures that remain within 30

the pulse temperature limit.

3. Fuel temperature during continuous reactivity addition from full power operations for delays in control rod insertion up to 3 seconds at reactivity addition rates up to 0.7% k/k per second will remain within the steadystate temperature limit.

4. Fuel temperatures following a loss of coolant after steadystate operation will remain within limits.

Table 7.1: Final Summary Analysis Value Max/Min Limit 564 °C 950°C /1150°C Steady State Power Level 24.34 kW per element 2.37 (CHFR) 2.0 (CHFR)

Reactivity Transients Pulse from Shutdown $3.00 530°C 830°C

$3.00 Pulse from Operation 522°C 830°C 1.02kW Continuous Reactivity Addition 0.2%k/k/s 573°C 1150°C from Full Power 1 s scram Loss of Coolant Event 25.00 kW per element17 787°C 950°C 17 Initial Steady State Condition 31