Text

University of Texas, Austin, Analysis of the Neutronic Behavior
	ML20115E287
Person / Time
Site:	University of Texas at Austin
Issue date:	12/31/2019
From:	; Oregon State University
To:	; Office of Nuclear Reactor Regulation
References
	Download: ML20115E287 (50)
	v • d • e

{{#Wiki_filter:ANALYSIS OF THE NEUTRONIC BEHAVIOR OF THE NUCLEAR ENGINEERING TEACHING LABORATORY REACTOR AT NETL Neutronic Analysis THE UNIVERSITY OF TEXAS Submitted By: Radiation Center Oregon State University Corvallis, Oregon December 2019 April 2020

Table of Contents

1.

Introduction.....................................................................................................................,... 4

2.

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

3.

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

4.

Reactor Core....................................................................................................................... 5

5.

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

6.

Bum up Calculations........................................................................................................... 9

7.

Current Core Configuration.............................................................................................. 10

8.

Limiting Core Configuration............................................................................................ 16

9.

Summary........................................................................................................................... 22 NETL Neutronic Analysis 2 April 2020

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......................... 11 Figure 7 - Current Core Power-Per-Element (in kW) Distribution at 1.1 MW............................ 11 Figure 8 - Current Core Configuration Prompt Temperature Coefficient, aF, as a Function of Temperature......................*........................................................................................... 14 Figure 9 - Current Core Configuration Moderator Void Coefficient............................................ 15 Figure 10- Current Core Configuration Moderator Temperature Coefficient............................. 15 Figure 11 - Vertical Cross-section of Limiting Core Configuration MCNP Model..................... 17 Figure 12 - Limiting Core Configuration Power-Per-Element Distribution at 1.1 MW............... 17 Figure 13 - Limiting Core Configuration Prompt Temperature Coefficient, aF, as a Function of Temperature................................................................................................................. 19 Figure 14-Limiting Core Configuration Moderator Void Coefficient........................................ 20 Figure 15 - Limiting Core Configuration Moderator Temperature Coefficient........................... 21 List of Tables Table 1 - Characteristics of Stainless Steel Clad Fuel Elements.................................................... 5 Table 2 - Summary of Bumup Steps............................................................................................. 10 Table 3 - Peff and Prompt Neutron Lifetimes for Current Core Configuration............................. 12 Table 4-Current Core Rod Worth Calculations.......................................................................... 13 Table 5 - Current Core Configuration Prompt Temperature Coefficient...................................... 14 Table 6-K-Effective Calculations Used to Determine Current Core Power Defect................... 16 Table 7 - Peff and Prompt Neutron Lifetimes for Limiting Core Configuration........................... 18 Table 8 - Limiting Core Configuration Rod Worth Calculations................................................. 19 Table 9 - Prompt Neutron Lifetimes in Limiting Core Configuration......... Error! Bqokmark not defined. Table 10- Limiting Core Configuration Prompt Temperature Coefficient.............. :................... 20 Table 11 - K-Effective Calculations Used to Determine Limiting Core Power Defect................ 21 Table 12 - Limiting Core Hot Channel Power Summary............................................................. 22 NETL Neutronic Analysis 3 April 2020

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 Sa,fety 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 TRI GA 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 April 2020

STAINLESS STEEL TOP END OF FITTING~ i -~03~ CLADDING lHICKNESS 0.021N. 1.475 IN. STAINLESS STEEL TUBE ZIRCONIUM HYDRIDE* 8.5WT%

URANIU~

MOLYDISC. 0.08 MM THlcf<'= j ,. -~O,;.. STAINLESS STEa / BOTTOM END FITTING 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 BOL 235U 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 J 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

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 TRI GA 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-NETL Neutronic Analysis 5 April 2020

fueled location that allow for a larger irradiation facility (in po ition 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. G23 Rabbit G2l G35 2928 G20 10815 G36 2925 F01 3504 G18 34~ G2 6142 G17 G3 5919 G16 3700 G6 2952 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 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. NETL Neutronic Analysi 6 April 2020

Figure 3-Horizontal and Vertical Cross-sections of the NETL MCNP Model at BOL The NETL reactor initially achieved criticality in March of 1992, however all of the fuel ( except for the fresh FFCRs) was previously used at other facilities. Most ofit 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 bum the fuel in conjunction with the given bumup records. The SCALE outputs were used to create BOL fuel isotopics for the MCNP runs.

5.

Model 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 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 3/23/1 992. The k-effective of this configuration was 0.99393 +/- 0.0001 3, or -$0.87 +/- $0.04. Eighty different critical core configurations were then analyzed to determine how they bounded ETL Neutronic Analysis 7 April 2020

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). $1.00 $( 1.00) $(2.00) $(3.00) $(4.00) Reactivity (including bias) ...,......,,,... -*--. --... ~..__...... $(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 but some are at high power. 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. If the first 44 runs are ignored (if runs after 5/5/92 are observed), the data looks more accurate (see Figure 5), with an average of -$0.23. 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. NETL Neutronic Analysis 8 April 2020

$0.60 $0.40 $0.20 $(0.20) $(0.40) * $(0.60) $(0.80) 45 50 Reactivity (including bias) 55 60 65 Critical Configuration Number 70 75 80 Figure 5 - Reactivity (including bias) of 36 Different BOL Critical Core Configurations Thus the model bias that will be used for this study is -$1.10 (the -$0.23 bias plus -$0.87 bias). This 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 bumup and previous grid location of these elements, it is nearly impossible to accurately determine their fuel compositions. This bias will be used to determine reactivity values in the following sections.

6.

Burnup Calculations After performing the initial model bias calculations, a series of MCNP BURN calculations were performed to bum 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 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 2). Each bumup step involved the fuel bumup for the specified amount of MW-days, parsing of the output fuel isotopics, then subsequent core model reconfiguration. 9 April 2020

Table 2 - Summary of Burnup Steps Burnup From To MW-days Total 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 NewIFE 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 31.3 13 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 45.8 14 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 NewIFE

7.

Current Core Configuration Once the bumup 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. NETL Neutronic Analysis 10 April 2020

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. G27 5.91 G28 5.17 G29 G30 14.1 7.82 F23 8.12 F24 8.2 7.55 F26 6.8 G32 Source G23 E18 10.4 E19 10.7 E20 E21 9.33 F27 7.63 G33 6.49 G22 12 D14 11 D15 12.9 E22 10.6 F28 8.679 G34 Rabbi G21 5.5 (09 13.9 (10 14.3 en D17 13.5 E23 11.1 F29 8.61 G35 6.6 G20 5.2 F17 14.1 805 16 806 15.8 15.6 D18 13.6 E24 10.9 F30 8.014 G36 5.9 F16 6.3 804 15.9 A01 CT 801 COl Trans 001 13.1 EOl 9.9 FOl 7.4 G18 5.1 HS 13.9 803 16.3 802 16.4 14.5 002 13.6 E02 11.71 F02 8.471 G2 6.1 G17 5.8 cos 14.1 (04 15.3 (03 003 13.4 E03 11.6 F03 9.18 G3 6.7 G16 11.6 006 11.3 DOS 13.4 12.7 E04 11.4 F04 9.329 G4 7.037 EOB 10.3 E07 11.3 E06 EOS 10.7 FOS 8.63 GS 6.76 7.68 F09 8.57 FOB 8.78 8.29 F06 7.56 G6 6.123 GlO 6.32 G9 GB 5.88 Figure 7 - Current Core Power-Per-Element (in kW) Distribution at I.I MW The red highlighting indicates the hottest fuel element location, in B-1, with a maximum power of 16.39 kW (at a total maximum core power of 1.1 MW). NETL Neutronic Analysis I I April 2020

Effective Delayed Neutron Fraction and Prompt Neutron Generation Time MCNP outputs effective delayed neutron fraction (Pelf) 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 Pelf and prompt neutron lifetime (see Table 3). Table 3 - Peff and Prompt Neutron Lifetimes for Current Core Configuration Case Prompt Neutron Error (s) ~elf 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 PeffWas calculated to be 0.00710 +/- 0.00007. This is in reasonable agreement with values predicted in other LEU TRI GA cores (i.e., Oregon State University Peff= 0.0076 [3], University of Maryland Pelf= 0.007 [4]) and also the value historically used for the NETL of Peff = 0.007. The value Pelf= 0.007 will be used to express all dollar values of reactivities in this report. The average prompt neutron generation time for the cases I Table 3 is 4 7.191 +/- 7.284 seconds. NETL Neutronic Analysis 12 April 2020

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 4 - Current Core Rod Worth Calculations MCNP MCNP MCNP Experimental Case k-effective k-effective Difference Rod Full-In Rod Full-Out Rod Worth Reactivity 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 Shim2 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 I 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 ofreasons, 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 13 April 2020

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, a.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]. ~

i.

-$0.025 =,-. f ~ -$0.020 ~ i. Cl.~ ~ ~ -$0.015 E--1.... - = ~.Sl -$0.010 i;.. ~ -.~ 9" ~ -$0.005 ~------- =U ~ $0.000 200 700 1200 Temperature (Kelvin) + NETL __

_G=A~-l 1700 2200 Figure 8-Current Core Configuration Prompt Temperature Coefficient, aF, as a Function of Temperature Table 5 - Current Core Configuration Prompt Temperature Coefficient Fuel Temperature rK]

Prompt Temperature Coefficient [$/°Cl 446.8 -$0.0130 750 -$0.0208 1050 -$0.0092 1850 -$0.0010 NETL Neutronic Analysis 14 April 2020

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 I 0% intervals. The void coefficient was negative for every interval and steadily decreased, as can be seen in Figure 9. $0.00 -c ,-.. -$0.20 +------- "C ~ c *= -$0.40 ~ -~ > -$0.60 J -~~ ~ ~ "',.. -$0.80 +----------------- ------< Cl> 0 Cl> "g U C..-$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 ofreactivity, aM, was determined by varying the moderator density with respect to temperature within the MCNP model from the expected operating temperature range of 20°C to 50°C (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 25°C 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 -$0.0 15 -$0.020 +-'~~-'---1~~~-~~~+-'~~-'--+~~~-~~-'---i 20 25 30 35 40 45 50 Moderator Temperature (°C) Figure 10 - Current Core Configuration Moderator Temperature Coefficient 15 April 2020

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 6-K-Effective Calculations Used to Determine Current Core Power Defect Case MCNP k-effective Standard Deviation Reactivity Error (2-sigma) Low Power 1.04118 0.0001 2 $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.

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. NETL Neutronic Analysis 16 April 2020

Figure 11-Vertical Cross-section of Limiting Core Configuration MCNP Model Core Power Distribution Figure 12 shows the power-per-element (in kW) in the suggested limiting core configuration. G29 5.90 G30 G21 4.72 4.67 G17 cos 18.01 C04 18.79 (03 17.52 003 14.90 14.34 006 13.87 DOS 15.90 004 14.04 E04 EDS E07 13.02 ED6 12.02 EDS 8.90 F09 9.20 FD8 9.20 F07 8.90 Gll 5.45 GlO 5.45 G9 5.46 G8 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 at 1.1 MW is now 22.14 +/- 0.06 kW, which is NETL Neutronic Analy i 17 April 2020

higher than the current core hot channel, due to a lower fuel loading concentrating more power at the center of the core. Effective Delayed Neutron Fraction and Prompt Neutron Generation Time Once again using the "KOPTS" card and running nine cases, the effective delayed neutron fraction Peff and prompt neutron generation times were calculated Table 7 - f}etr 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 The average PeffWas calculated to be 0.00735 +/- 0.00007. There is a slight increase in PeffCOmpared 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. NETL Neutronic Analysis 18 April 2020

Table 8-Limiting Core Configuration Rod Worth Calculations Case MCNP k-effective MCNP k-effective MCNPRod 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 Shim2 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. 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. -$0.025 4> -$0.020 + NETL

GA J,,

=,-. co ~ J,, J,, ~ ~ -$0.015. e~ 4>,._,

=

-$0.010 4> 4> = *- l;a;.~ -$0.005 Q, 4> e = =U J,, $0.000 ~ 200 700 1200 1700 2200 Temperature (Kelvin) Figure 13 - Limiting Core Configuration Prompt Temperature Coefficient, aF, as a Function of Temperature NETL Neutronic Analysis 19 April 2020

Table 9 - Limiting Core Configuration Prompt Temperature Coefficient Fuel Temperature rKl Prompt Temperature Coefficient f$/°Cl 446.8 -$0.01302 750 -$0.02081 1050 -$0.00928 1850 -$0.00105 These values are similar to the original BOL coefficients. Moderator Void Coefficient Figure 14 shows the moderator void coefficient in the suggested limiting core configuration. $0.00 ~----------- ------- -$0.20 'Cl

§... :9 -$0.40

...... = Q S -~ ;... -$0.60 +-------- ... =~ CIS.,_ o t Cl.I a.. -$0.80 'Cl Q Cl.I ~ U ; -$1.00 -$1.20 -$1.40 +-....__,___,____._ __ _.__~~-~~~--+---'-~~......_............ ___..____._*-'----i 0 20 40 60 80 100 Percent Void Figure 14-Limiting Core Configuration Moderator Void Coefficient 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. NETL Neutronic Analysis 20 April 2020

Q,l s {....-. $0.010 5.§ ~ $0.005 -+---------------+-----------I ~ S t $0.000 -+------+-----+----------t-----t------t a Q,l c.. ~ 8 e$o.oo5 t "C 0 ~ -$0.010 -l-------<1---------1--------lf----------j -$0.015 --l--+------------------------; -$0. 020 --J-L.___l..-'--'--f--l---'-------'-----'-+-'------'----l'---'--+--'--'--'---'--l--'--'---'---'----1---1..-'-------'-----L-f 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 all observed temperature ranges. Power Coefficient of Reactivity The power coefficient of reactivity results are seen in Table 10. Table 10 - K-Effective Calculations Used to Determine Limiting Core Power-Defect Case MCNP k-effective Standard Deviation Reactivity Error (2-sigrna) Low Power 1.04231 0.00015 $6.90 $0.04 Full Power 1.01921 0.00010 $3.79 $0.03 Thus the power defect is $3.11 +/- $0.05. This is lower than the current core configuration's 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 :finesh 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. NETL Neutronic Analysis 21 April 2020

Table 11 - Limiting Core Hot Channel Power Summary Hot Rod Hot Rod Hot Rod Hot Rod Core Hot Rod Thermal Peak Factor Axial Peak Radial Peak Effective Configuration Location Power [kW] [Pmax!Pavg] Factor Factor Peak Factor rPmax!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 reactor's 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 model's 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. REFERENCES [1] NUREG-1282, "Safety Evaluation Report on High-Uranium Content, Low-Enriched Uranium-Zirconium Hydride Fuels for TRI GA 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 TRI GA Reactor from HEU to LEU Fuel," Submitted by the Oregon State University TRI GA 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 TRI GA Reactors, General Atomics (1967). [6] "Safety Analysis Report" Submitted by the University of Texas at Austin Nuclear Engineering Teaching Laboratory (~anuary 2012). [7] Engineering Toolbox. Web. Accessed May 3rd, 2017. NETL Neutronic Analysis 22 April 2020

Link: http://www.engineeringtoolbox.com/water-thermal-properties-d _162.html NETL Neutronic Analysis 23 April 2020

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 April 15, 2020 1

1.0 INTRODUCTION

This report documents analysis of the thermal hydraulic characteristics of the UT TRIGA nuclear research reactor in support of renewal of the U.S. Nuclear Regulatory Commission facility operating license. NRC guidance1 specifies definition of 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 encompassed by safety analysis for the LCC. Coupled analyses for neutronic and thermal hydraulic behavior are used to characterize thermal hydraulic performance of the LCC. This report describes the thermal hydraulic analysis, the neutronic analysis is a separate report2* Heat generated by fission in operation of the UT TRIGA reactor is transferred by conduction from the fuel to the cladding. The 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 integrity of the fuel cladding is assured if the fuel temperature remains below specific values (830°C during pulsing3, 950°C if the cladding temperature is less than S00°C, and 1150°C if the cladding temperature is above S00°c4). Heat is transferred to water [or air in the case of a loss of coolant accident (LOCA)] surrounding tbe fuel element cladding. The cooling media temperature increase develops buoyancy forces that drive convection flow. Cooling flow is impeded by momentum changes and friction (across the grid plates, fuel element end fittings, and fuel element cladding surfaces). Above a "critical" heat flux (CHF), cooling flow may not be adequate to prevent exceeding the temperature limits, and 'burnout' may occur. The NRC guidance limits the ratio of heat flux to critical heat flux (CHFR) to a value greater than 2.0. A correlation for CHFR applicable to TRIGA reactors has been developed by Bernaths and is used in this analysis. The fuel element in the LCC that operates closest to the limits is the element generating the most power, 'hot channel.' Analysis of the hot channel demonstrates that operation at the maximum licensed power level and within licensed reactivity limits maintains the reactor fuel below the temperature limits, and heat transfer to the surrounding cooling media remains above the CHFR limit. TRACE (TRAC/RELAP Advanced Computational Engine) is the NRC's flagship thermal-hydraulics analysis tool consolidating and extending the capabilities of N~C's three legacy safety codes: TRAC-P, TRAC-B and RELAP. These codes are designed to perform best-estimate analyses of operational transients and accident scenarios by modeling physical geometry and thermodynamic conditions. TRACE and RELAP were developed for commercial nuclear reactors applications, and RELAP has also been widely used in characterizing non-power/research reactor thermal hydraulic performance. Thermal hydraulic modeling of the hot channel in the UT TRIGA reactor was performed using TRACE. The hot channel model was developed for TRACE using standard, classical methods applied to the fuel element geometry and fuel element pitch. However, fuel element construction includes complex end fittings not well represented in modeling flow resistance classically, and coefficients characterizing the resistance to 1 NURGE 1537, Guidelines for Preparing and Reviewing Applications for the Licensing of Non-Power Reactors, Format and Content 2 Analysis of the Neutronic Behavior of the Nuclear Engineering Teaching Laboratory at The University of Texas, Radiation Center - Oregon State University 3 TRD-070-0.1006.05 Rev A, Pulsing Temperature Limit for TR/GA Fuel (TRIGA Reactors Division of General Atomics-ESI, 2008) 4 NUREG/CR-2387 Credible Accident Analyses for TR/GA and TR/GA-Fueled Reactors (PNL-4028, April 1962) 5 ANL RETR TM 07 01, Fundamental Approach to TRIA Steady-State Thermal-Hydraulic CHF Analysis, E. E. Feldman (2007) 2

cooling flow for the UTTRIGA were therefore developed in the computational fluid dynamics code FLUENT.6 Data in TRACE output files were parsed to identify maximum fuel temperature and variables used in the Bernath correlation. TRACE has an internal library of standard material properties relevant to power reactors. The library does not include properties for zirconium and TRIGA fuel; these material characteristics were supplied as manual input. The ratio of the maximum power in the hot channel to average core-wide element power (peaking factor, developed from the neutronics analysis} is used to extend transient calculations from elements generating core average power to calculations for the element generating the maximum power in the core. The MCNP model developed in the neutronics analysis was used to calculate power generated in each fuel element, fractions and energy of fast and thermal fissions, ~eff, and (in an adjoint calculation} delayed* neutron precursor constants and the prompt neutron lifetime. An auxiliary program distributed with MCNP (MAKXS} was used to generate neutron interaction cross sections at specific fuel and water temperatures. These cross sections were used in MCNP criticality calculations to develop reactivity temperature-coefficients for TRACE reactivity calculations. The delayed neutron precursor constant2 were used to develop time-dependent heat generation from fission following shutdown for LOCA' analysis. The heat* generation from fission product decay for LOCA analysis was developed from the method described in ANSI/ANS-5.1-2014, Decay Heat Power in Light Water Reactors. Heat generated from neutron interaction with fission products was neglected. The model was used characterize performance of the LCC following validation. Analyses were developed to characterize performance associated with steady-state operation, reactor pulsing, continuous control rod withdrawal transients, and a LOCA. Validation of the modeling was performed by comparing the results of analyses to peak temperatures during steady state operations, and power levels and temperatures that occurred during pulsing operation. 2.0 TRACE MODEL The TRACE model uses 'break' elements that establish pressure and temperature fluid boundary conditions. Pool water is delivered from break through a 'downcomer' with a horizontal structure connecting the downcomer to the inlet of the fuel element flow channel. The fuel element flow channel discharges to another break.. A conceptual diagram of the model is provided in Fig. 1. 2.1 Flow Channel Geometry The inner diameter of the TRIGA fuel (an annulus filled with a 0.225 in. diameter zirconium rod} is 0.25 in. The outer diameter of the fuel is 1.435 in. There is a small gap between the fuel and the cladding; the radial gap for production elements is estimated to be less than 0.0025 in.7 and was assumed to be 0.00126 in. for this analysis. The cladding is 0.02 in. thick. The outer diameter of standard TRIGA fuel is nominally 1.475 in. and was modeled as slightly expanded in this analysis to 1.478 (DF} in order to model the gap, which is a major contributor to temperature elevation in the fuel compared to the cladding. (With cladding 1.475 in. diameter and thickness of 0.02 in., the inner cladding diameter of 1.435 in. does not have a gap between cladding and the 1.435 in. diameter fuel element.} The center-to-center distance to adjacent elements (pitch, Pe} is 1.714 in. 6 (Doctoral Dissertation) Development of thermal Hydraulic Correlations for the University of Texas at Austin TRIGA Reactor Using Computational Fluid Dynamics and In-Core measurements, A. D. Brand (2013), https://repositories.lib.utexas.edu/handle/2152/23039 7 Fission Product Release from TRIGA-LEU Reactor Fuels, N. L. Baldwin, F. C. Fousheee and J. S. Greenwood (10/1980) 3

cooling flow for the UTTRIGA were therefore developed in the computational fluid dynamics code FLUENT.6 Data in TRACE output files were parsed to identify maximum fuel temperature and variables used in the Bernath correlation.. TRACE has an internal library of standard material properties relevant to power reactors. The library does not include properties for zirconium and TRIGA fuel; these material characteristics were supplied as manual input. The ratio of the maximum power in the hot channel to average core-wide element power (peaking factor, developed from the neutronics analysis) is used to extend transient calculations from elements generating core average power to calculations for the element generating the maximum power in the core. The MCNP model developed in the neutronics analysis was used to calculate power generated in each fuel element, fractions and energy of fast and thermal fissions, ~ett, and (in an adjoint calculation) delayed neutron precursor constants and the prompt neutron lifetime. An auxiliary program distributed with MCNP (MAKXS) was used to generate neutron interaction cross sections at specific fuel and water temperatures. These cross sections were used in MCNP criticality calculations to develop reactivity temperature-coefficients for TRACE reactivity calculations. The delayed neutron precursor constants were used to develop time-dependent heat generation from fission following shutdown for LOCA analysis. The heat generation from fission product decay for LOCA analysis was developed from the method described in ANSI/ANS-5.1-2014, Decay Heat Power in Light Water Reactors. Heat generated from neutron interaction with fission products was neglected. The model was used characterize performance of the LCC following validation. Analyses were developed to characterize performance associated with steady-state operation, reactor pulsing, continuous control rod withdrawal transients, and a LOCA. Validation of the modeling was performed by comparing the results of analyses to peak temperatures during steady state operations, and power levels and temperatures that occurred during pulsing operation. 2.0 TRACE MODEL The TRACE model uses 'break' elements that establish pressure and temperature fluid boundary conditions. Pool water is delivered from break through a 'downcomer' with a horizontal structure connecting the downcomer to the inlet of the fuel element flow channel. The fuel element flow channel discharges to another break. A conceptual diagram of the model is provided in Fig. 1. 2.1 Flow Channel Geometry The cooling flow channel is modeled as a heated pipe with thermodynamic characteristics based on physical dimensions and properties of the coolant surrounding the fuel elements. The outer diameter of the fuel element cladding (DF) is 1.475 in. The cladding is 0.020 in. thick, with a small gap between the inner cladding diameter and the fuel outer diameter. The outer diameter of the fuel is 1.435 in. with an inner diameter of 0.25 in. that is filled with a 0.225 in. diameter zirconium rod. The center-to-center distance to adjacent elements (pitch, Pe) is 1.714 in. 6 (Doctoral Dissertation) Development of thermal Hydraulic Correlations for the University of Texas at Austin TRIGA Reactor Using Computational Fluid Dynamics and In-Core measurements, A. D. Brand (2013), https://repositories.lib.utexas.edu/handle/2152/23039 3

Inlet Break Figure 1: UT TRACE Model Outlet Break The geometry of the fuel element and surrounding flow area (Figs. 2 and 3) is modeled as a hexagon with an inner radius (i.e., the largest circle centered in and bounded by the hexagon) Yi of the pitch. A large fraction of the model is occupied by the fuel element, leaving a relatively small flow area. End fittings have more complex geometry and are approximated using hydrodynamic characteristics. Equations 1-4 (following) provide data required in the TRACE input file. COOLING WATER CLADDING Zr FILL ROD Figure 2: Flow Channel for UT TRIGA The area of the flow channel (A) is the area of the hexagon less the area of a fuel element. Using Pe as pitch and DF as the diameter of the fuel element, the area of the flow channel is given by: A = V:. p~ - rr. ( D;) 2 Eqn 1 The wetted perimeter (Pw) is the perimeter ofthe fuel element: Eqn 2 4

Flow Channel with Area A Adjacent Fuel Element Figure 3: Pitch and Fuel Element Diameter And Relation to Flow Channel Non-circular pipes are approximated from the flow area and the wetted perimeter (Pw) with an equivalent hydraulic diameter (D,.) calculated using: 4*A Dh=-- Pw Eqn 3 Substituting equations 1 and 2 into equation 3 yields the following expression for the hydraulic diameter: D =-4 __ [..{3 *p2-1r*(DF)2]= 2.../3. p; -D =D *[Jj. p; -1] Jr* DF 2 e 2 Jr DF F F 2 D; Eqn 4 2.2 Thermal Hydraulic Loss 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. Standard practice defines the characteristics that determine head loss as coefficients, or K factors based on experimentally derived correlations to support conservation of energy and momentum calculations for fluid flow. The surface roughness for TRIGA fuel elements is 6.998E-6 ft.7 Correlation of K factors with geometry and flow are based on historical, experimental measurements with cylindrical pipes. The correlations are extended to rectangular ducts; non-circular cross sections are reduced to a flow area and hydraulic diameter developed by assuming the area is a circular geometry. These correlations apply after flow is fully developed following geometry perturbations that create turbulence. The geometry inherent in TRIGA fuel challenges application of standard formula to calculate the K factors: The fuel element end fittings are partially inserted in grid plate penetrations that fix the position of the fuel elements in the core, causing abrupt flow area expansion and contraction at upper and lower grid plates that affects end fitting flow conditions. Flow entrance to and exit from the flow channel between the grid plates is directed by fins, a center 7 Oregon State University, Model Editor File for the OSTR thermal hydraulic Model (private communication) 5

pin, and a conical shape (with part of the structure extending into the volume bounded by grid plate penetrations); the wetted perimeters vary continuously from entrance to exit for each end fitting. The interface between adjacent fuel channels is not separated by a physical boundary, and differential pressure between adjacent flow channels can support cross-channel flow, although analysis has shown that any cross-flow is small and has "no effect on fuel temperature" with a slight increase in critical heat flux ratio8; cross flow is neglected in this analysis. Therefore, thermal hydraulic analysis to support relicensing was developed to independently model thermal hydraulic performance at the end fittings from (1) first principles, (2) the TRACE thermal hydraulics code, and (3) the FLUENT computational fluid dynamics code. The results of experiments in the TRIGA core were used to evaluate UTTRIGA-specificKfactors based on actual fuel element geometry. The values determined from the UT research program (Table 1) were used in modeling for TRACE calculations. Table 1: UT TRIGA Specific K Factors APPLICATION K Factor TRIGA Lower Channel 1.63 TRIGA Upper Channel 1.12 2.3 Hydrostatic Pressure Cooling media pressure and temperature specifications associated with break components are based on local environmental conditions (barometric pressure, confinement pressure regulation) and the pool water (level and temperature). The NETL building is approximately 240 m above sea level, and the reactor bay confinement system is designed to control differential pressure to 0.06 in. below atmospheric pressure (minimal compared to atmospheric pressure). Total pressure at the top of the core is: Pr= 96{.KPa}+ PHp Eqn 5 Pool water is nominally 7.25 m above the core, with a minimum of 5.25 m. Pool water temperature is nominally 25-27 °C, and administratively limited to less than 49 °C. Where g denotes the gravitational constant (9.8 m*s-2), the pressure (pH20) exerted by a column of water (at density pin kg*m-2and height h in m) is given by: Eqn 6 Breaks using the same parameters are connected to the entrance of the downcomer and exit from the flow channel (representing the exit of the core). Pressure boundary conditions for the limiting cases are provided in Table 2, converted to British units for TRACE input. Table 2: LCC Temperature & Pressure Condition Limiting Nominal Temp Pressure (°F) (Psia) 120.2 21.3 77 24.2 8 Armed Forces Radiobiology Research Institute (AFRRI) submittal of Safety Analysis Chapters 4 and 13 to the USN RC (ML101650422, submitted on March 4, 2010) 6

2.4 Heat Structure Modeling The heated length of the fuel element is modeled as a TRACE heat structure. The heat structure simulates the zirconium fill rod at the center of the fuel element, the fuel matrix, the gap between the fuel and cladding, and the cladding. The UT TRIGA heat structure is segmented by 15 radial and 15 axial coordinates into cells that transmit heat from the fuel element to the cooling channel. The UT TRl<;:iA model uses a gas gap heat transfer coefficient9 of 500 Btu h-1n--2 °F-1, (2840 W m-2 K-1). The flow channel includes non-heated lengths that are not part of the heat structure above and below the heated length representing integral graphite reflector and end fittings (with a space in the upper end fitting for fission gas collection incorporated above the upper graphite reflector). 2.5 Power/Reactivity Application Fuel element power was used as an input for steady-state and LOCA analysis, supporting cases with time-dependent reactivity variation, and as an input to extend average fuel element power analysis to the hot channel. Reactivity was used as an input for transient analysis, supporting pulsing, control rod withdrawal, and LOCA analyses. Reactivity transients were used to generate time dependent power levels for a fuel element representing the core average in pulsing and control rod withdrawal analysis. The time dependent average element power level was scaled by the core peaking factor to represent the hot channel time dependent power calculations. For LOCA analysis, the time dependent behavior of power following shutdown was evaluated and developed using reactor kinetics for fission power and the method of ANSI/ ANS-5.1-2014 (equations 7 and 8) for decay heat from fission products. For an irradiation interval of time T, a decay time oft, and irradiation intervals i and a constant fission rate of unity, decay heat power (F, units of MeV/fission) can be represented analytically as a function using analytic fit constants a and p (the fit constants are provided for all 23 components in the standard): 23 Fi ( t, Ta) = L :~,~

exp ( -Ai,j
t) * [ 1 - exp ( -Ai,j
T)]

j=1 l,] Eqn 7 Assuming a single power generation interval and a long operating time, thi~ reduces to: 23 ~a-F(t) =LA~* exp(-Aj

t) j=1 J Eqn 8 The ratio of decay heat power to initial reactor power depends on a reactor specific energy generation per fission (MeV/fission) that is also used in TRACE transient calculations. The MCNP burnup analysis output tabulates fission energy yield data (Q value) as given in Table 3.

Table 3: Fission Energy Yield from MCNP Analysis Nuclide Q-Value (MeV) 92235 180.88 92238 94239 94241 181.31 189.44 188.99 9 The University of Texas Safety Analysis Report, Revision 1.01 (5/91) 7

The fraction of energy produced by each fissionable material (Table 4) used in TRACE analysis is taken from an MCNP burnup calculation supporting the neutronics report. Estimates of the fraction of isotope 92235 fissions at greater than thermal energy is assumed to have Q values consistent with isotope 92238. Table 4: Fission Isotope Nuclear Characteristics10 Fission Fissions at Nuclide Fraction Cross-section Energy Range Energy (b) 92235 19.8% 585.1 <0.625 eV 94.28% 274.4 0.625 ev - 100 kev 4.99% 1.241 92238 80.2% 0.03064 >100 kev 0.72% The kinetics treatment used fission product nuclear characteristics (precursor fractions, decay constants, and the generation time) from MCNP adjoint calculations given in Table 5. The'rmal/non-thermal fission fractions are taken from the MCNP burn up calculations. Table 5: Delayed Neutron Precursor Group Characteristics p; Standard Energy Standard A; Standard T112 Standard Deviation (MeV) Deviation (s-1) Deviation (s) Deviation 1 0.00022 0.00005 0.40605 0.00460 0.01334 0.00000 51.960 0.004 2 0.00138 0.00013 0.47218 0.00202 0.03273 0.00000 21.178 0.001 3 0.00138 0.00012 0.44217 0.00202 0.12080 0.00000 5.73797 0.00005 4 0.00296 0.00019 0.55867 0.00181 0.30295 0.00001 2.28799 0.00008 5 0.00113 0.00011 0.51895 0.00294 0.85032 0.00004 0.81516 0.00004 6 0.00059 0.00009 0.53926 0.00495 2.85537 0.00021 0.24227 0.00002 Heat generation distribution in the heat structure is based on mesh tally calculations from MCNP modeling described in the neutronics report. The MCNP mesh tally provides average fission density at 225 locations throughout the fuel rod. These 225 locations consist of equal volume segments of the fuel with 15 radial segments and 15 axial segments. The normalized mesh values were applied to locations in TRACE.. This 15x15 matrix was used as a power distribution profile in TRACE. Equal volume segments allow direct averaging of temperatures in the fuel MCNP adjoint calculations were used to generate estimates of the prompt neutron generation time and effective delayed neutron fractions. The MCNP calculations produced an estimated prompt-neutron generation time of 43.81+/-0.53 µs (Table 6). The 1992 Safety Analysis Report for the UTTRIGA cited a prompt generation time of 41 µs, which is reasonably consistent with that calculated by MCNP; the MCNP calculated value utilizing current parameters is used in this report. The effective delayed neutron fraction has historically been used as 0.007 for the UT TRIGA. The MCNP calculated value for Petr Was 0.00737+/-0.00006 (Table 6). Calculating pulsing reactivity in units of$ based on Petr of 0.007 represents 91% of the nominal pulsed reactivity. Data from two series of reactor pulses with varying reactivities were analyzed using the Fuchs-Hansen pulse model modified to evaluate Petr from the convergence of iterative calculations. Pulses from November 2018 were found to have a Petr of 0.007306, and from March 2020 to have a Petr of 0.007382. 10 https:ljwwwndc.jaea.go.jp, Fission at Energy from MCNP burnup calculation 8

2.6 User Defined Materials Table 6: MCNP Calculated Reactivity Parameters Parameter Generation Time ~eff Estimate 43.81 µs 0.00737 Standard Deviation 0.53 µs 0.00006 Zirconium properties are provided in Table 7. The zirconium fill rod is slightly smaller than the center hole of the fuel element. Therefore, the standard zirconium density was reduced by the ratio of the zirconium rod volume to the whole center-hole volume. Table 7: Zirconium Properties Temperature Density Heat Capacity (Cp) Thermal Conductivity Emissivity OF lbm/ft3 Btu/lbm-°F Btu/h-ft-°F -99.4 328.18 0.082284 14.562 0.8 260.6 328.18 0.102199 12.4812 0.8 620.6. 328.18 0.122114 11.9592 0.8 980.6 328.18 0.142029 12.4812 0.8 1340.6 328.18 0.161943 13.6944 0.8 1700.6 328.18 0.181858 15.0228 0.8 2240.6 328.18 0.211731 16.6392 0.8 4481.2 328.18 0.335679 21.6392 0.8 Reference values11 for specific heat capacity (Cp) and thermal conductivity (kc)of TRIGA fuel are: Cp = 0.018 +/- 0.009 watts [cm-10C-1] kc= 2.04 + 0.00417 x T [W s-1 cm-3 0 C-1] (note that this is given as a function of material temperature T). These formulae were converted to British units and used to generate tabular values (Table 8) for TRACE. Table 8: Uranium-Zirconium Hydride (Fuel) Temperature Density Heat Capacity (Cp) Thermal Conductivity Emissivity OF lbm/ft3 Btu/lbm-°F Btu/h-ft-°F -65.6 374.5 0.029754 10.16021 0.8 294.4 374.5 0.091403 10.16021 0.8 654.4 374.5 0.153051 10.16021 0.8 1014.4 374.5 0.2147 10.16021 0.8 1374.4 374.5 0.276349 10.16021 0.8 1734.4 374.5 0.337997 10.16021 0.8 2274.4 374.5 0.43047 10.16021 0.8 4548.8 374.5 0.814165 10.16021 0.8 11 E-117-833, The U-Zrx Alloy: Its Properties and Use in TRIGA Fuel (Simnad, General Atomics Project No. 4314) 9

2.7 Temperature Feedback Temperature feedback is incorporated in the reactivity transient analysis to support pulsing and control rod withdrawal event analyses. General Atomics 12 indicates that the fuel temperature coefficient (water reflected) is convex, with a minimum occurring about 300°C. An analysis at AFRRl 13 based on DIF3D (Argonne National Laboratory, diffusion and transport theory code) calculations show convex fuel temperature reactivity structure from 10-1000°C for TRIGA fuel. Standard MCNP cross sections are available in intervals of about 300°C; it is not possible to get data from calculations where cross-sections are separated by 300°C that would provide accurate data for derivation of reactivity coefficients over the range of operation. Therefore, an auxiliary code distributed with MCNP (MAKXSF) was used to generate data over the range of interest. Although fuel temperature and water temperature coefficients are generally considered independently, they are coupled. The temperatures used for generating these cross sections are shown in Table 9. Fuel 98.33 Water 76.73 118.13 76.73 Table 9: Cross Section Temperatures (°F) 344.93 80.33 458.33 107.33 571.73 134.33 841.73 168.53 1111.73 1291.73 188.33 215.33 1471.73 1651.73 Criticality calculations with the MCNP model of neutronics analysis were performed using all permutations of temperatures indicated. Fuel temperature, water temperature, and the associated excess reactivity for each calculation was used to generate a reactivity function of the two temperatures. The function was used to evaluate values Yi degree above and Yi degree below specific fuel temperatures with water temperature held constant, and then varying water temperature with constant fuel temperatures. The results are provided graphically at selected fuel temperatures, fuel temperature reactivity coefficients for a series of constant moderator temperatures (Fig. 4) and moderator temperature coefficients for a series of constant fuel temperatures (Fig. 5). -6.0E-05 u -7.0E-05 0.... Q)

a.

~ -8.0E-05 (.() - -9.0E-05 I-z UJ 0 -l.OE-04 iI UJ 0 -l.lE-04 u i::: > -l.2E-04 ~ u <t: -l.3E-04 UJ Cl: -1.4E-04 0 200 400 600 800 1000 FUEL TEMPERATURE (0 () -+--24.85 -+--26.85 -+-41.85 -+-56.85 -+--75.85 -+-86.85 -+--101.85 -+--116.85 Figure 4: Temperature Coefficient of Reactivity for the Fuel Versus Fuel Temperature (shown at a variety of fixed coolant temperatures) 12 Simnad op. cit. 13 AFRRI op. cit. 10

1.0E-05 'u O.OE+OO 0... QJ a. -1.0E-05 .llii: i -2.0E-05 1-fl:i -3.0E-05 u i:i: -4.0E-05

u.

w 8 -5.0E-05 ~ -6.0E-05 ~ -7.0E-05 a: -8.0E-05

*- *~-

-*-- --- ----- *-----r--*--**----------*--- ;--------*---*--!----- - 20 40 60 80 100 120 WATER MODERATOR TEMPERATURE (°C) +FT,25C -+-FT,48C +FT,800C +FT,900C Figure 5: Water Temperature Coefficient of Reactivity Versus Water Moderator Temperature (shown at a variety of fixed fuel temperatures) The maximum variation in the fuel temperature reactivity coefficient at low temperature (with respect to water temperature) is on the order of 10% from nominal water temperature to temperatures near boiling in the core. The water temperature change is dominated by the decrease in water density as temperature increases; lower water density increases scattering length and reduces moderation in the cooling channels. As fuel temperature increases there is less absorption in the fuel matrix and consequently more neutron scattering out ofthe fuel matrix into the water. These competing effects cause the difference with respect to water temperature to converge and then reverse the difference at higher fuel temperatures. At constant fuel temperature, the moderator temperature coefficient change is almost linear with respect to moderator temperature change from ambient temperature values to near boiling at the core hydrostatic pressure. The calculations based on DIF3D14 suggest fuel coefficient reactivity ranges from -8x10-5 to -1.2xl0-4 Llk/k 0 C-1 for a TRIGA core with a circular grid plate (the UT TRIGA has hexagonal pitch), or -4.5x10-5 and -6.8xl0-5 Llk/k °F-1. This is in reasonable agreement with this analysis. The temperature coefficient data is used in TRACE to determine temperature reactivity coefficient feedback for transient analysis. 3.0 VALIDATION The neutronics report identifies instrumented fuel elements (IFEs) in the current core (1.1 MW core, 113 elements) in positions B03 and BOG generating 15.82 kW and 15.39. kW respectively. The UT TRIGA administratively limits power operation to 950 kW, 86.4% of the licensed power level; therefore, the power levels generated in the thermocouple elements at nominal full power operations are 13.66 kW and 13.29 kW. Normal operations occur with a pool (cooling water) temperature of 25°C (77°F). Pool level during normal operations is 7.25 m {23.8 ft) above the core; therefore, hydrostatic pressure is 24.2 psia for normal operations. A series of calculations with nominal conditions was performed to allow comparison of observed power levels and temperatures to those generated by TRACE model simulating nominal operations. 14 AFRRI 11

3.1 Steady State Each IFE has three thermocouples embed in the fuel matrix, with the thermocouple exhibiting the highest temperature connected to the fuel temperature measuring channels. The temperatures of the thermocouple locations used in the measuring channel were calculated using the TRACE model. One thermocouple was replaced in 'February 2016 and indicated high temperature indication from installation to the present. This is typical behavior for replacing a partially depleted IFE with one containing fresh fuel: the amount of 235U is higher in the replacement element, there are no fission product poisons, and the fuel tolerance that allows insertion into the cladding impedes heat transfer (fuel swelling during operation eventually closes the gap). These transient conditions have a significant effect on power distribution and heat transfer. However, the material composition for the IFE in the neutronics analysis was based on average isotopic values at initial fuel loading, which included fission products and transuranic isotopes (all standard fuel in initial loading had prior power history). To avoid associated complications, the baseline data for validation is, taken from December 2015 through February 2016 covering IFE load configurations in Table 10. Temperature calculations are extremely sensitive to power generation, so a correction was made for MCNP material composition with the fresh IFE to more accurately reflect actual fuel loading with no fission products or transuranics in the MCNP material. Power distribution was developed from calculation of fissions in each element using the MCNP model as modified. Table 10: IFE CONFIGURATIONS AND INSTRUMENT INDICATIONS DATE POWER IFE FTl IFE FT2 12/18/15 0.92 10878 (B03) 325°c 10708 (B06) 364°C 01/16/16 /FE 10708 replaced with /FE 10809 01/27/16 0.92 10878 (B03) 325°C 10809 (B06) 427°C 02/01/16 0.93 10878 (B03) 319°C 10809 (B06) 420°C 02/01/16 FT1 and FT leads exchanged 2/2/2016 0.94 10809 (B06) 419°C 10878 (B03) 322°C P_ower distribution data from the neutronics analysis is assumed to apply prior to the instrumented fuel element replacement and scaled to 92% consistent with the power level when tbe temperature data was taken. TRACE calculations indicate temperature 10°C below the measuring channel that used IFE 10708 and 36°C above the channel with IFE 10848 before element 10708 was removed After 10708 was removed and 10809 installed, the TRACE calculations show temperature 26°C elevated over 10848 and 21 °C below 10809. Table 11: TRACE AND INSTRUMENT COMPARISON INITIAL CONFIGURATION ELEMENT POWER TRACE/MCNP FT 10708 13.61 354°c 364°C 10848 13.24 345°c 319°c FOLLOWING IFE EXCHANGE ELEMENT POWER TRACE/MCNP FT 10809 15.30 394°C 420°C 10848 13.30 346°C 325°C Comparison of the TRACE calculation and observed data shows reasonably close agreement (within +/-8.5% of the measured value), with differences principally attributed to the IFE material composition in the MCNP model used to calculate the power generation rate. 12

3.2 Pulsing Operations 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. There is significant scatter in power level (Fig. 6) and temperature (Fig. 7) data with some outliers but, the results overall show good agreement AND provide a basis for validation. Pulse records do not include factors with potential to affect pulse characteristics such as initial fuel temperature, pool temperature, or recent operating history that might explain some of the scatter in the data. Since TRACE calculations are for individual elements, peak pulse power level was distributed across the core for comparison to historical data. Although there is significant scatter and outliers in historical pulse power level data, it is clear that qualitatively the TRACE data 1agrees well with historical data. A comparison of peak temperature to historical pulsing data required adjusting the average pulse power to the power generated in the IFE by the ratio of the power in the instrumented fuel elements (identified in the neutronics report) to the average element power. The fuel matrix in IFEs is fabricated with milled channels to accommodate thermocouple leads, and the thermocouples are installed in fuel penetrations extending from the outer surface to near the inner surface of the fuel. Consequently, the volume of the fuel in the IFE is a few percent lower than a standard fuel element. One IFE was installed in 2016 while the MCNP model did not account for higher loading of fresh fuel. All fuel material specifications in the MCNP model used to calculate power generated in the IFE assumed the same density as standard fuel elements, while initial loading varied. Even with the complications created by MCNP modeling, data scatter, and outliers the TRACE calculations agree well with observed data. 2.50E+07 2.00E+07

l.50E+07

~ t E CIJ w 1.00E+07 !ti :. 5.00E+06 D D !o d~ a~" - 0 oclD Joa~ i ~. O.OOE+OO o-=-'i'E--~---~--~---~----'----~--~ $1.00 $1.50 $2.00 Pulsed Reactivity Addition ($) o Historical Observed Data OTRACE $2.50 $3.00 Figure 6: Peak Element Power Level Versus Pulse Reactivity Addition from UT TRACE Calculation Compared to Observed Historical Data 13

450 400 350 ~ 300 ~

i

~ 250 QJ c.. E 200 ~

lJ 150 c..

100 so IJ 1 - c f a ao p1 c51 -- ( -- D. --~IJ~ dl9" i --- ---, --- -o-~Ji~----:--- ---- - I IJ IJ IJ IJI I IJ - i!iJil;I ~*IJ . --------- ---- ----1----- ---- --- ----- -*----*f---- ---* -

* ------ --~- - ----1------

~---~--~---~---~--~---_J*---~--____J 0 $1.00 $1.50 $2.00 $2.50 $3.00 Pulse Reactivity Addition($) c Historical Observed Data O TRACE Figure 7: Peak Fuel Temperature Versus Pulse Reactivity Addition from UT TRACE Calculation Compared to Historical Data 3.3 Conclusion 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. 4.0 LIMITING CORE CONFIGURATION 4.1 Steady-State Power Level The neutronic analysis indicates that for a licensed power of 1.1 MW, the maximum power in any element is 22.14 kW with ~4 elements. With a maximum potential instrument error of 10%, the maximum power in the hot channel could be as high as 24.34 kW. This UT TRIGA hot channel power level is well within TRIGA operating experience, less than reported values for the AFFRI TRIGA reactor (35.3 kW) and the TRIGA conversion reactor at the University of Wisconsin (26.04 kW)15, and less than the 30 kW value cited in ANL RERTR TM 07 01 (which references the MNCR reactor at 33.2 kW). To evaluate steady-state performance of the LCC, a series of TRACE calculations were performed for the UT TRIGA for fuel element power levels between 5 kW and 60 kW under limiting core conditions. Temperature dependence of the fuel versus power is shown in Fig. 8 for the hot channel, while the radial and axial temperature distribution across the hot channel are shown in Figs. 9 and 10. Fuel element power at 44.9 kW is found to result in a maximum fuel temperature during steady state operations of 1149.9°C, essentially 15 University of Wisconsin LEU Conversion Report (Request for Amendment No. 17 to Facility License No. R-17, 08/25/2008) 14

the maximum safety limit if cladding temperate is greater than S00°C. Fuel element power at 36.6 kW results in 948.1 °C, approaching the 950°C safety limit if cladding temperature is less than S00°C. ---~~---! 1200 1100 1000 u 900 0 - 800 ~ 700 +-' ~- 600 QJ

a.

E ~ 500 400 300 200 I:: 100 i: I* 0 0 5 10

.. i

1.

'*I! l 15 20 25 30 35 40 Fuel Element Power Level (kW)

.. >-e; I

I -o-IFE-TCl

- - 6-* IFE - TC2 45 50

,JIii -e-MAX FT Figure 8: LCC Steady-State Fuel Temperature Versus Fuel Element Power at Each Thermocouple and at the Maximum Fuel Temperature Location 800 700 600 usoo .2.- (]) L. t; 400 L. (]) Cl. E ~ 300 200 I o-------- 1 II Zirc ,- Fill-+ [ Rod I 0 0.1 - - - - - - -.- - -0. --o_ -o, -o_ Fuel 0.2 0.3 0.4 'O Gas 'o, 0.5 'O Gap 'o 'o.. 0.6 'a. 'o Distance from Rod Center_ (in.) Clad 'o., 0.7 0.8 Figure 9: LCC Hot Channel Radial Fuel Temperature Profile at Hottest Axial Position 15

800 700 600 G 500

l ro ai 400 Q.

E (lJ '.:::'.: 300 (lJ

l u..

200, I 100 I. i l o L 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 Elevation from Bottom of the Element (in.) o Hottest Channel

Average Channel Figure 10: LCC Hot Channel Axial Radial Fuel Temperature Profile 16.0 The critical heat flux that leads to burnout is calculated by the Bernath correlation (with variables defined in Table 11):

CHF80 = [10890

De; Di+ V * :~~6] * ([57
lnP-54
P: 15 -i]-rs )FCta)

Eqn 9 Table 11: Bernath Correlation Variables Hydraulic diameter ft Previous formula Heater surface diameter ft Fuel element diameter Pressure Psia Calculated by TRACE Velocity ft*s-1 Calculated by TRACE Coolant Temperature °C Calculated by TRACE A series of calculations were performed using the Bernath correlation to determine the relationship between element power and the minimum CHFR along a rod in limiting core conditions (Fig. 11). The limiting value of 2.0 for CHFR occurs at 56 kW generated in the hot channel, in agreement with analysis cited in ANL RERTR TM 07 01. With the UT TRIGA hot channel operating at 24.35, the critical heat flux ratio is 4.69. 16

25 II 19 a 17 I ~ 15 X 13 u::: ~ 11 I 9 7 3 I 1L~---- 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Element Power Level (kW} Figure 11: CHFR (Bernath Correlation) Versus Fuel Element Power The critical heat flux ratio varies along the length of the hot channel fuel element at full power as indicated in Fig 12. The minimum CHFR is slightly above the center of the section of the fuel element that contains fuel. 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 Elevation Along Heated Length (in.) Figure 12: CHFR Hot Channel Axial Values from the Bernath Correlation for UTTRIGA Fuel Element Operating at 24.34 kW 17

With the hot channel operating at 24.34 kW, the maximum temperature in the core is calculated by TRACE at 650°C, approximately 300°C below the 950°C safety limit. The hot channel power is 12.3 kW below the 36.6 kW that would approach the fuel temperature safety limit. The hot channel element CHFR with power level of 24.34 kW is 4.69 - significantly above the CHFR limit of 2.0. The LCC hot channel is 33.7 kW below 55 kW, the power level that would result in a CHFR of 2.0. Based on these results the LCC characteristics are well within the safety margins. 4.2 Pulsing Operations Pulsing reactivity calculations were performed in TRACE for the ~CC hot channel using $3, $4, $5, $6, and $7 insertions with initial conditions approximating shutdown power levels. The calculations were performed in two steps: (1) reactivity addition was used to calculate time-dependent core power level of an average fuel element and (2) the time/power profile was used in second step, with the power of the second step scaled by the hot channel peaking factor to represent the LCC hot channel fuel element. The peak power as a function of pulsed reactivity is shown in Fig. 13, and the time-development of the power pulse for the reactivity insertions in Fig. 14. The $7 insertion is not included so that the remaining data can be shown on a reasonable scale. vi' ~ ro L.. QJ 5 0 CL QJ Vl

s CL E

s E

x ro

~ 1.6E+08 1.4E+08 1.2E+08 1.0E+08 8.0E+07 6.0E+07 4.0E+07 2.0E+07 O.OE+OO $2.50 $3.00 $3.50 $4.00 $4.50 $5.00 $5.50 $6.00 $6.50 Pulse Reactivity Insertion Figure 13: Hot Channel LCC Peak Power Level Versus Reactivity Insertion 18

1.60E+08 1.40E+08 1.20E+08

E 1.00E+08 Cll

~8.00E+07 QJ ~ ~ 6.00E+07 4.00E+07 2.00E+07 O.OOE+OO 1.02 I \\ I \\ J

I I

\\. - \\ \\ ' \\. V _,*\\ \\ 1.04 ____ '..-:.'.'_,7-------------------------- 1.06 Time (s) 1.08

$3.00 -*- $4.00

---$5.00 -$6.00 Figure 14: LCC Pulse Power Versus Time for Varying Reactivity Insertions A $3 insertion resulted in a peak temperature of 378°C. As shown in Fig. 15, the temperature limit was not exceeded for any reactivity insertions below $6 (with a 130°C or 16% margin at $6). The temperature safety limit was not also exceeded at a $7 reactivity insertion. Time dependent behavior is shown in Fig. 16. 750 700 650 I p-600 (l}

i I

1-1 ~ 550 I (l}

0.

E 500 ~ w 450 [ _

i LL I

400 350 I 300 I L__ $2.50 ,,,//,,,,,,",,, ,,o' - o' ,.0 $3.00 $3.50 $4.00 $4.50 $5.00 $5.50 $6.00 $6.50 Pulse Reactivity Figure 15: LCC Hot Channel Peak Fuel Temperature Versus Pulse Reactivity Addition 19

825 725

i.

625 p-525 Q.) L.. +J 425 L.. !1l L.. Q.) c.. 325 E

~------

,/ ~ I I l 1 I i -{/ -~*-~ ~*~~ _-_ -~~T. --*., I I - ~-~ ~-~-- ~ ~ ~ ~ - ~ - ~--i- - ____ I_.. ____ -- -- -- *: ** - I ~ iii 225 LL.. 125 j --.,,.*----------+----*---------*-*-- ---------** ---. -- ----- -t------- ---- ---- --- i I J 1* ..... r. _ 25 ---**------~-----* 0 10 20 30 40 so 60 Time(s) ---$3.00 -*- $4.00 -----$5.00 -$6.00

~~---------------**-*-*-------

Figure 16: LCC Hot Channel Fuel Temperature Versus Time from Pulse Initiation for Several Pulse Reactivity Additions The maximum effect of pulsing at power was evaluated. The average channel was characterized with (1) simulated incremental insertion to $4, (2) a period at $4 to establish initial conditions, and (3) total reactivity addition of $7. The resulting time dependent. power was taken from the calculation, scaled by the hot channel peaking factor to simulate the hot channel, and used in a time dependent power calculation. The maximum fuel temperature in the hot channel prior to the pulse was 252 °Cat 8.25 kW (corresponding to a core power of 43.3 kW). Following the pulsed insertion of $3, the peak hot channel fuel temperature was 579°C and peak power 74.4 kW (Fig. 17). 700 u 1 90 f--------------------------11* 80 600 f *

- - - - ~* *.. ---~-- --

70 .Q.) 500 I

. ~/_/_--_* *_---_-_.. _. _-*-_*-*_--_*--_-_-_-**_-.. _______

-__,_-1 60 ~ L.. +J ~ Q.) c.. E ~ w ',I / ) 400 ' I *

- 50 5 I

~ ctH----------------------!- 40 +-'

\\ ----*-*-,--- ----------*-* *-----*--*---

-1

~ ct-_~_ ~\\----_ __,--.---*--!***--... --.. -..... -. __ -_-_.... -___ -__ -_ ----------!,- 30 ~ rt---~.;;;::::::::::::::::::::-------.,......~~c:::'.::~~~c:::'.::~1-20 -- ---1 300 200 100 I f I w 0 ' *--*-*-*-*---**-----*--*--*----------*--------------~ 0 0 10 20 Time (s)

Temperature -Power 30 40 Figure 17: LCC Hot Channel Fuel Temperature and Element Power Versus Time for a $3 Pulse from Steady-State Operation with Element at an Initial Power of 8.25 kW and Fuel Temperature of 252 °C (Note that pulse was initiated at 0.5 seconds on the time.scale shown) 20

A maximum pulsed reactivity addition of $3.00 from low power is adequate to maintain fuel element temperature less than 50% of the limit. Maximum pulsed reactivity addition with the reactor operating at the balance of excess reactivity will not exceed the limiting fuel temperature at any initial achievable power level. 4.3 Continuous Reactivity Addition Transient (Control Rod Withdrawal Event) The ne.utronics report indicates the maximum worth of the control rods in the LCC is $4.43 (for the regulating rod). Historical data indicates the greatest integral reactivity worth of any control rod since initial criticality was $4.50. Annual measurement of control rod speeds support evaluation of the maximum reactivity addition rate; control rod drive speeds over the 15 in. span are typically about Yi in. per second (30 in. per minute). A set of TRACE calculations was performed to simulate the effect of a continuous control rod withdrawal to t~e full out position for various integral control rod worths. Reactivity addition was modeled based on the fraction of total reactivity added as a function of position,.and a constant speed was assumed to determine the fraction of the total integral reactivity added as a function of time for withdraw-intervals (full-in to full-out) of 5, 15, 30, 45; 75, 90, 120 and 135 seconds. Reactivity scale factors of $3, $4, $5, $6 and $7 were applied to each interval to calculate average fuel element response to the transient. The time dependent power resulting from this set of calculations was modified with the hot channel peaking factor to simulate the LCC hot channel, and a time dependent power calculation was performed. The characteristic of a 5 second interval are essentially a reactor pulse with a small perturbation on the pulse-tail. The CHFR (using the Bernath correlation) was calculated as a minimum of 2.41 for the $7 reactivity additions to a minimum of 5.76 for a $3 addition, both at 135 seconds (2.25 minutes) for full reactivity addition. Power, temperature, and reactivity characteristics for a reactivity addition to $3 over 45 seconds is displayed in Figs. 18 and 19. 20 18 16 -i 14 ~ - ,_ 12 QJ 0 10 a.. QJ 8 C C co .c 6 u 0 4

c 2

0 ,, '/ /it /,t ___./_!_ \\ __ ,' I ,\\ ~---~------~------- -i,','- . \\,........... ------------ --l ------- i, / /

\\ / /

r...., -l I - : - /-. i / .IL i I ! / -~r 0 50 100 150 Run Time (s)

Hot Channel Power

- -Max Fuel Temp. 400 350 u 0 300 QJ,_

J 250 co,_

QJ c.. 200 E ~ C 150 QJ E QJ 100 UJ E

J 50 E

x co 0

~ 200 Figure 18: Hot Channel Power and Maximum Fuel Temperature Versus Run Time for a $3 Reactivity Addition in 45 Seconds 21

2.SE-02 ~ 2.0E-02 ~ u l.SE-02 J:I -a Ill Ill l.OE-02

J..

I!! 5.0E-03 Ill Q. E ~ 0.0E+OO "ii

J...

-5.0E-03 ail C 0 t: -1.0E-02 Ill "' .E ~ -l.SE-02 .i! t:.. -2.0E-02 Ill a: -2.SE-02 ' ' ' ' ' t \\. i / \\.I /\\.. I \\ \\ \\ I \\ ' \\. ' \\ '\\ / *- **- **- **- **- **- **- **- **- **- **- **- ** I I \\\\ 20 ' \\ \\ \\ ~\\ ~-, \\ '-------------------- 40 bO 100 120 140 160 18') Time (s) - *

6k Add -

- 6k FTC

- - - - 6k MTC O.E+OO

-2.E-06 -1.E-05 -1.E-05 Figure 19: Reactivity Addition, Fuel Temperature Reactivity Feedback, and Moderator Temperature Reactivity Feedback Reactivity Versus Time for a Continuous Reactivity Addition for $3 in 45 Seconds Power level response versus time is illustrated in Figs. 20 and 21 for $3 and $7 insertions that occur over a range of intervals. 40000 35000 30000 !2sooo "a; .!!f 20000 CII ~ 15000 Q. 10000 5000 0 0 so 100 Time after Initiation of Transient (s) 30 - 45 - 60 - 75 120 - 135 150 Figure 20: Power Level Versus Time After Initiation of Transient for a Continuous Rod Withdrawal of $3 in Reactivity 22

120000 100000 80000 4i Cl/ 60000 Cl/ ~ 0

0. 40000 20000 0

0 50 100 Time after lntiiation of Transient (s) 30 - 45 - 60 - 75 - 90 - 105 - 120 150 Figure 21: Power Level Versus Time After Initiation of Transient for a Continuous Rod Withdrawal of $7 in Reactivity Notably, as the interval time for the reactivity addition increases, the maximum element power over the interval decreases. The element power level following the transient [Table 12, 'Element Power (kW)] is strictly a function of the total reactivity addition regardless of the reactivity insertion rate. The maximum fuel temperature resulting from specific reactivity additions occurs after the pulse as a function of the total reactivity added [Table 12, 'Peak Temp ( 0C)'] and therefore did not vary significantly with varying time intervals. Table 12, Response to Reactivity Addition ilk Peak Temp Final Element Addition (*q Power (kW) $1.00 226 7.15 $2.00 399 14.60 $3.00 572 21.21 $4.00 748 28.42 $4.75 835 33.99 $5.00 923 35.75 $6.00 1090 43.17 $7.00 1252 50.38 The maximum fuel temperature for the $5 reactivity addition approaches but does not exceed the 950°C fuel temperature safety limit. Maximum reactivity addition up to $4.75 results in a maximum fuel temperature 748°C, well below the fuel temperature safety limit, and is greater than the maximum control rod worth measured since initial criticality of the NETL. Following a reactivity insertion transient with a total reactivity addition at $3 or less steady state power level 23

does not exceed 1.1 MW. If no credit taken for protective action from the power level instrumentation channels, power levels may exceed 1.1 MW during the transient, but total reactivity addition up to $5 does not exceed the fuel temperature safety limit and remains well within the minimum CHFR. control rod drive systems at the UTTRIGA are independent and cannot be operated in a group. The control rod drives are controlled by software, and interlocks (including prevention of withdrawing multiple control rod simultaneously) are tested annually. The regulating rod and the two shim/safety rods are powered by stepper motors (with individual power supplies for each rod drive system) while the transient rod uses an analog motor. The regulating rod drive can be operated to automatically maintain a demand power level, the shim rod drives can only be operated manually. A failure mode that causes withdrawal of multiple control rods has not been identified. Continuous withdrawal of a single control rod with an integral worth less than $4.75 at any control rod drive speed is adequate to maintain fuel temperature less than the maximum fuel temperature limit with a 115°C margin. 4.4 LOCA Analysis The LOCA analysis was a 2-step process with the first a steady-state TRACE calculation to establish initial conditions at 25 kW. This was followed by a TRACE restart case initiated as a transient calculation, with fission and fission product power decay in time established using the method of ANSI/ANS-5.1-2014, Decay Heat Power in Light Water Reactors (as described). Four cases were calculated (Table 13) using four intervals between shutdown and instantaneous replacement of water cooing with air cooling. The time-dependent behavior is shown in Fig. 22 over four hours following shutdown. Table 13: LOSS OF WATER-COOLING ANALYSIS DELAY FOR AIR MAXIMUM COOLING (s) TEMPERATURE ( 0 C) 1 709 60 699 600 1200 663 637 In a ratiol']al scenario, an automatic shutdown will be initiated before the core is voided by a low pool water level trip. Drain down of pool water under the most exaggerated conditions will require on the order of 10 to 20 minutes before water is voided from the core area. A 10-minute delay will result in a maximum temperature of 663°C, and a 20-minute delay will result in a maximum temperature of 637°C. There are other practical considerations not considered in modeling that make the analysis c~nservative in a LOCA event. A drain-down through the beam ports (the only path for flow) will not drain below the core, and some lower portion of the fuel elements will remain in a water environment to provide heat removal with conduction through the stainless-steel cladding. The air cooling will occur in extremely humid conditions, increasing heat removal capability. 24

800 700 _ 600 ~ ~ 500 n:, lii 400 C. E ~ 300 'aj

J LI. 200 a,

C

i 100 a,

C a, 0 u -;;,~~.~~~~~=:?~t2.;~~-:* /./ / // i ~----}1~1:1:_ ______. ______ ; ____. ________ --~----------------*--* -------- --*-* , t' I/ I /1 /* i--£'.f I H: ,,/.

-*- _____ _!_ ________. *-*-. --- -- - __ i -*-* --* ----*

I i '" !( _____ ----*. -------~------------------ -:----** - -------* ----- I r * *** *** * *- * -

1
- -r- -- ---

1 l 0 5000 10000 15000 Time After Cessation of Water Cooling (s)

1 s H20 ---60 s H20 -*-600 s H20 -** 1200 s H20 20000 Figure 22: Maximum Post LOCA Temperatures for a 25 kW Element Versus Time Following a LOCA Event (shown for a variety of delay times for air cooling)

The UT TRIGA LCC hot channel fuel element temperature during a LOCA is calculated to rise to 709°C (assuming a minimal 1 second delay between reactor shutdown and displacement of water with air). Analysis of fuel performance during a LOCA for the AFFRI reactor predicated on 72 hours of full power per week indicates a maximum fuel temperature of about SS0°C if there is no delay before water displacement and about 475°C assuming a 15-minute delay. The UT analysis assumes equilibrium steady state analysis, about 2.3 times the power history of the AFFRI reactor, generating more decay heat. The maximum temperature of 709°C provides a margin of 241 °C to the limiting temperature. Therefore, the LCC is adequate mair1tain fuel integrity unchallenged by temperature in a LOCA. / 5.0

SUMMARY

AND CONCLUSION There are two limits that ensure TRIGA fuel integrity during steady-state operations and one for pulsing. During steady-state operations a fuel temperature limit of 950°C and a CHFR limit of 2.0 applies. During pulsing operations, a fuel temperature limit of 830°C applies. Thermal-hydraulic calculations using TRACE linked to a neutronics calculation (using MCNP 6.2) was performed to available safety margins in operation of the UT TRIGA reactor LCC. With the hot channel of the UT LCC producing 24.3 kW, analysis shows that the CHFR is greater than 4.0 with a maximum fuel temperature of 650°C. Therefore, steady-state operations at 1.21 MW with the LCC is adequate to assure the safety limit is met. In pulsing operations to $3 from non-power conditions, the maximum fuel temperature is calculated to be 378°C. Pulsing at power from the maximum power level (reserving $3 for the maximum pulse), fuel temperature is calculated to be 579°C. These temperatures are well below the temperature limit for pulsing operations. Therefore, a $3 pulsing limit is adequate to assure the safety limit is met for the LCC. 25 i

The maximum fuel temperature from a continuous reactivity addition event up to $4.75 at any reactivity addition rate will not exceed 835°C. The control rod drive system limits reactivity addition by multiple control rods. Therefore, limiting each integral control rod worth to less than $4.75 is adequate to assure the fuel temperatur.e to the safety limit is met in a continuous rod withdrawal event for the LCC. A lo~s of coolant accident after continuous full power operations to steady-state conditions will not result in fuel temperature greater than 709°C (with the cooling delayed by 1 second). Therefore, 1.21 MW operations with the LCC ensures cooling is adequate for fuel integrity in a loss of coolant accident. 26}}

ML20115E287

Text

1.0 INTRODUCTION

SUMMARY

Navigation menu

Search