ML12346A004
| ML12346A004 | |
| Person / Time | |
|---|---|
| Site: | University of Missouri-Columbia |
| Issue date: | 06/28/2012 |
| From: | Rhonda Butler Univ of Missouri - Columbia |
| To: | Office of Nuclear Reactor Regulation |
| Wertz, Geoffrey 301-415-0893 | |
| References | |
| TAC ME1580 | |
| Download: ML12346A004 (81) | |
Text
UNIVERSITY OF MISSOURI, COLUMBIA MISSOURI UNIVERSITY RESEARCH REACTOR LICENSE NO. R-103 DOCKET NO. 50-186 LICENSE RENEWAL APPLICATION RESPONSES TO THE REQUEST FOR ADDITIONAL INFORMATION DATED JUNE 28, 2012 REDACTED VERSION*
SECURITY-RELATED INFORMATION REMOVED
- REDACTED TEXT AND FIGURES BLACKED OUR OR DENOTED BY BRACKETS
UNIVERSITY of MISSOURI RESEARCH REACTOR CENTER June 28, 2012 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Mail Station P1-37 Washington, DC 20555-0001
Reference:
Docket 50-186 University of Missouri-Columbia Research Reactor Amended Facility License R-103 Enclosed you will find the University of Missouri-Columbia Research Reactor's (MURR) responses to the U.S. Nuclear Regulatory Commission's (NRC) request for additional information, dated May 6, 2010 (Complex Questions) and June 1, 2010 (45-Day Response Questions) regarding our renewal request for Amended Facility Operating License R-103, which was submitted to the NRC on August 31, 2006, as supplemented.
If you have any questions, please contact John L. Fruits, the facility Reactor Manager, at (573) 882-5319 or FruitsJ@missouri.edu.
Sincerely, Ralph A. Butler, P.E.
Director RAB/djr Enclosures Ao2o 1513 Research Park Drive Columbia, MO 65211 Phone: 573-882-4211 Fax: 573-882-6360 Web: www.murr.missouri.edu Fighting Cancer with Tomorrow's Technology
UNIVERSITY of MISSOURI RESEARCH REACTOR CENTER June 28, 2012 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Mail Station P1-37 Washington, DC 20555-0001
REFERENCE:
Docket 50-186 University of Missouri - Columbia Research Reactor Amended Facility License R-103
SUBJECT:
Written communication as specified by 10 CFR 50.4(b)(1) regarding responses to the "University of Missouri at Columbia - Request for Additional Information Re: License Renewal, Safety Analysis Report, Complex Questions (TAC No. MD3034)," dated May 6, 2010, and the "University of Missouri at Columbia - Request for Additional Information Re: License Renewal, Safety Analysis Report, 45-Day Response Questions (TAC No. MD3034)," dated June 1, 2010 On August 31, 2006, the University of Missouri-Columbia Research Reactor (MURR) submitted a request to the U.S. Nuclear Regulatory Commission (NRC) to renew Amended Facility Operating License R-103.
On May 6, 2010, the NRC requested additional information and clarification regarding the renewal request in the form of nineteen (19) Complex Questions. By letter dated September 3, 2010, MURR responded to seven (7) of those Complex Questions.
On June 1, 2010, the NRC requested additional information and clarification regarding the renewal request in the form of one hundred and sixty-seven (167) 45-Day Response Questions. By letter dated July 16, 2010, MURR responded to forty-seven (47) of those 45-Day Response Questions.
On July 14, 2010, via electronic mail (email), MURR requested additional time to respond to the remaining one hundred and twenty (120) 45-Day Response Questions. By letter dated August 4, 2010, the NRC granted the request. By letter dated August 31, 2010, MURR responded to fifty-three (53) of the 45-Day Response Questions.
On September 1, 2010, via email, MURR requested additional time to respond to the remaining twelve (12) Complex Questions. By letter dated September 27, 2010, the NRC granted the request.
1513 Research Park Drive Columbia, MO 65211 Phone: 573-882-4211 Fax: 573-882-6360 Web: www.murr.missouri.edu Fighting Cancer with Tomorrow's Technology
On September 29, 2010, via email, MURR requested additional time to respond to the remaining sixty-seven (67) 45-Day Response Questions. On September 30, 2010, MURR responded to sixteen (16) of the remaining 45-Day Questions. By letter dated October 13, 2010, the NRC granted the extension request.
By letter dated October 29, 2010, MURR responded to sixteen (16) of the remaining 45-Day Response Questions and two (2) of the remaining Complex Questions.
By letter dated November 30, 2010, MURR responded to twelve (12) of the remaining 45-Day Response Questions.
On December 1, 2010, via email, MURR requested additional time to respond to the remaining 45-Day Response and Complex Questions. By letter dated December 13, 2010, the NRC granted the extension request.
On January 14, 2011, via email, MURR requested additional time to respond to the remaining 45-Day Response and Complex Questions. By letter dated February 1, 2011, the NRC granted the extension request.
By letter dated March 11, 2011, MURR responded to twenty-one (21) of the remaining 45-Day Response Questions.
On May 27, 2011, via email, MURR requested additional time to respond to the remaining the remaining 45-Day Response and Complex Questions. By letter dated July 5, 2011, the NRC granted the request.
By letter dated September 8, 2011, MURR responded to six (6) of the remaining 45-Day Response and Complex Questions.
On September 30, 2011, via email, MURR requested additional time to respond to the remaining the remaining 45-Day Response and Complex Questions. By letter dated November 10, 2011, the NRC granted the request.
By letter dated January 6, 2012, MURR responded to four (4) of the remaining 45-Day Response and Complex Questions. Also submitted was an updated version of the MURR Technical Specifications.
On January 23, 2012, via email, MURR requested additional time to respond to the remaining the remaining 45-Day Response and Complex Questions. By letter dated January 26, 2012, the NRC granted the request.
On April 12, 2012, via email, MURR requested additional time to respond to the remaining the remaining 45-Day Response and Complex Questions.
Attached are MURR's responses to the remaining six (6) the remaining 45-Day Response and Complex Questions. With this" set of responses, all 45-Day Response and Complex Questions have been addressed.
2 of 28
If there are questions regarding this response, please contact me at (573) 882-5319 or FruitsJ@missouri.edu. I declare under penalty of perjury that the foregoing is true and correct.
ENDORSEMENT:
Reviewed and Approved, Sincerely, John L. Fruits Reactor Manager 6Q6?
Ralph A. Butler, P.E.
Director Enclosed: : : : :
Reference 1 for RAI 4.7, Relevant portion (Section 3.1) of MURR Hazards Summary Report, University of Missouri Research Reactor Facility, Addendum 1, February 1966.
Reference 4 for RAI 4.7, BORAL Composite Standard Specifications, SP-BORA-001en, Rev. 10-17-08, Ceradyne Canada, ULC.
Reference 5 for RAI 4.7, Govindarajan, S.G., Moreland, J.A, Solbrekken, G.L., BORAL Control Blade Thermal-Mechanical Analysis, Mechanical and Aerospace Engineering Department, University of Missouri, Columbia, Missouri, June 2012.
Reference 6 for RAI 4.7, MURR Technical Data Report TDR-0133, "MURR BORAL Control Blade Thermal-Mechanical Analysis," University of Missouri-Columbia Research Reactor, Columbia, Missouri, June 2012.
xc:
Reactor Advisory Committee Reactor Safety Subcommittee Dr. Robert Duncan, Vice Chancellor for Research Mr. Geoffrey Wertz, U.S. NRC Mr. Alexander Adams, U.S. NRC Mr. Craig Basset, U.S. NRC
_;PNOTARY.'
MARGEE P. STOUT My Commission Expires March 24,2016 Montgomery County Commission #12511436 Avý f6
/
-7 3 of 28
CHAPTER 4 4.7 Section 4.2.2.3, Evaluation of Control Blade Distortion, Page 4-18. How 860 BTU/hr/ft2 was calculated as the fuel element thermal conductivity is not discussed Back-calculating we performed results in a thermal conductivity which differs from other published sources (ORNL-981). Please provide additional detail for the calculation of this value.
The possibility of control blade binding due to thermal distortion that is discussed in Section 4.2.2.3, "Evaluation of Control Blade Thermal Distortion," of the Safety Analysis Report (SAR) was first addressed in a response to a question raised by the U.S. Nuclear Regulatory Commission (NRC) in 1966. By letter dated January 5, 1966, the NRC requested additional information based on their review of the University of Missouri's application for a Class 104 utilization facility license, which was submitted to the NRC in July of 1965.
Question 3.1 stated "Provide the assembly drawings of the control rods and control rod drive mechanisms and design drawings showing overall blade dimensions and clearances in the guide structure.
Describe the preoperational tests and periodic tests to be performed on the control drive system. Discuss the possibility of control rod binding due to thermal distortion of the control blades" (Ref. 1).
The response to that question (Ref. 1), which was accepted at that time by the NRC, is essentially the same analysis that is presented in Section 4.2.2.3 of the SAR.
There is no specific information in the original response on how a radial heat flow value of 860 BTU/hr-ft2 was determined.
In ORNL-981, "BORAL: A New Thermal Neutron Shield, Supplement 1," the specimen used to determine the thermal conductivity of BORAL, as described in the Appendix of that document, was 2 cm (0.79 inches) in diameter and 5 cm (1.97 inches) in length, or thickness, and with no mention of any type of cladding material (Ref. 2).
Compare those dimensions to that of a MURR control blade where the thickness of the BORAL core is 0.100 inches (2.54 mm) and the thickness of the aluminum cladding is 0.0375 inches (0.9525 mm) on each side, for a total nominal control blade thickness of 0.175 inches (4.445 mm).
Ceradyne Canada, ULC, the current owners of the BORAL technology, referred MURR to an Electric Power Research Institute (EPRI) publication on how to calculate overall thermal conductivity based on various thicknesses of BORAL core and cladding. Handbook of Neutron Absorber Materials for Spent Nuclear Fuel Transportation and Storage Applications (Ref. 3) provides the following equation on how to calculate the overall thermal conductivity of an aluminum clad BORAL plate:
k,
=
tt / ((2
- ta/ka) + (tc/1%)),
where:
k,
overall thermal conductivity; t
overall thickness of plate; ta
=
thickness of aluminum cladding on each side; ka =
thermal conductivity of aluminum; t,
=
thickness of core matrix; and 1c
=
thermal conductivity of core matrix.
4 of 28
This equation is also listed on SP-BORA-001en, BORAL Composite Standard Specifications, BORAL Typical Engineering Properties, Ceradyne Canada, ULC (Ref. 4). Using the above equation, with the dimensions of a MURR control blade and the thermal conductivity values stated in Reference 4, an overall thermal conductivity value of 1.02 W/cm-K (102 W/m-K) was calculated for a MURR control blade.
This value is more than twice the BORAL thermal conductivity value of 25.0 BTU/hr-ft-°F (43.2 W/m-K) stated in ORNL-981. Back-calculating a radial heat flow value of 860 BTU/hr-ft2 results in a thermal conductivity of only 1.47 BTU/hr-ft-
'F (2.54 W/m-K), which is orders of magnitude less than the values published in ORNL-981 and those provided by Ceradyne Canada, ULC.
In revisiting the analysis provided in Section 4.2.2.3 of the SAR, it was felt, based on over 45 years of operational history, that very slight radial, not axial, control blade deflections occur due to thermal gradients impressed during operation. As required by current Technical Specification (TS) surveillance requirement 5.3.b (TS 4.3.b of the relicensing TSs), one of the four shim (control) blades is inspected each six months so that every blade is inspected every two years. As part of this surveillance, the control blade and its associated offset mechanism are removed from service and a refurbished, e.g., bearings replaced, dimensions verified, etc., offset mechanism with a new or a previously used and decayed control blade is placed in operation. (Note: Control blades are typically used three to four times, i.e., six to eight years of total service.) Prior to placing a new or previously used control blade in service, the control blades are inspected to validate their proper material condition and the curvature and trueness of the blade are verified using curved templates and a flat, machined surface. Only slight "flattening" on the lower few inches of the control blade has been experienced. No axial deformation has ever been noted.
University of Missouri (MU) Mechanical and Aerospace Engineering (MAE) Department provided clarification to this question. Using the commercially available finite element code Abaqus FEA (version 6.10-2), a fully coupled thermal stress analysis on the MURR control blades was performed by MU MAE to determine the magnitude of thermally induced deflection (Ref. 5,6). The deflection will arise due to the composite structure of the control blade. As the BORAL absorbs both neutrons and gamma rays, there will be volumetric heat generation and a corresponding rise in temperature. Since the BORAL and aluminum materials have different coefficients of thermal expansion, there will be a tendency of the control blade to deform as the blade temperature changes and the materials expand at different rates.
In addition to the composite nature of the control blade, there will be spatial variations in temperature within the control blade caused by non-uniform heat generation within the BORAL meat. The high boron-10 cross-section of the B-10 (n, a) Li-7 thermal neutron reaction produces the vast majority of the heating in the control blade. This reaction primarily occurs within the first 20 mils of the BORAL meat surface and produces about 2.79 MeV of energy, of which 0.84 MeV is the reaction energy of the Li-7 and 1.47 MeV is the alpha particle. The remaining 0.48 MeV is a gamma ray. Therefore, about 80% of the nuclear reaction's energy is deposited within a few mils of the reaction location. Consequently, the major heating is in the two outer surfaces of the BORAL meat, thus making the heat generation through the blade low except for the outer 20 mils of each surface. These combine to produce a variation in the heating profile through the thickness and about the circumferential width of the control blade. The heat generation is also non-uniform along the longitudinal direction because the thermal neutron flux drops off 5 of 28
significantly from the leading edge (bottom) of the control blade to the top. With typical mixed burnup cores at cold clean startup each week, only about nine inches of the control blade is inserted past the top edge of the fuel meat in the fuel elements.
A mathematical curve-fit was generated for the non-uniforn volumetric heat generation profile caused by the B-10 (n, a) Li-7 thermal neutron reaction. Similarly, a curve-fit was also generated for the gamma absorption heating rate. The functions were applied as heating conditions within a finite element model of the control blade. A convective heat transfer coefficient was applied to the outer boundaries of the control blade and the neutral assembly temperature was assumed to be room temperature. The finite element model was solved for the temperature distribution. The temperature distribution was then used as an applied load in a mechanical deflection model. The resulting deflection was compared with the channel gap to determine if there is a significant risk of the control blade binding during operation.
The thermal conductivity of the BORAL used in the analysis was 76.8 W/m-K (Ref 4). However, experiments performed by the MU MAE Department on BORAL coupons of the same material specifications as the MURR control blades suggest that the thermal conductivity of the BORAL is closer to 11 5+/-17 W/m-K. A parametric study was performed by varying the thermal conductivity of the BORAL with a lower bound of 98 W/m-K and an upper bound of 132 W/m-K. The change in maximum radial deflection was found to be negligibly small and of the order of 4 x 10-4 inches.
In conclusion, based on the analysis performed by the MU MAE Department, and with over 45 years of operation with no instances of control blade binding, the clearances provided in the control blade gaps are adequate to accommodate, without significant interference, the maximum anticipated degree of control blade distortion due to thermal gradients produced during full-power operation.
REFERENCES
'MURR Hazards Summary Report, University of Missouri Research Reactor Facility, Addendum 1, Section 3.1, February 1966.
2Kitzes, A.S., Hullings, W.Q., ORNL-981, BORAL." A New Thermal Neutron Shield, Supplement 1, Oak Ridge National Laboratory, Oak Ridge, Tennessee, April 10, 1953.
3Handbook of Neutron Absorber Materials for Spent Nuclear Fuel Transportation and Storage Applications: 2009 Edition. EPRI, Palo Alto, CA: 2009. 1019110 4BORAL Composite Standard Specifications, SP-BORA-001en, Rev. 10-17-08, Ceradyne Canada, ULC.
5Govindarajan, S.G., Moreland, J.A, Solbrekken, G.L., BORAL Control Blade Thermal-Mechanical Analysis, Mechanical and Aerospace Engineering Department, University of Missouri, Columbia, Missouri, June 2012.
6 MURR Technical Data Report TDR-0133, "MURR BORAL Control Blade Thermal-Mechanical Analysis," University of Missouri-Columbia Research Reactor, Columbia, Missouri, June 2012.
6 of 28
CHAPTER 6 6.2 Section 6.3.8, Anti-Siphon System, Design Analysis. The analysis of the system using RELAP5 assumes the anti-siphon valves will open 85 msec after the primary coolant rupture occurs.
Discuss the basis for this assumption. Provide additional TS wording on anti-siphon valve performance orjustify why this is not needed A valve opening response time of 85 msec in the RELAP5 model for the anti-siphon isolation valves was based on the best time measurements that were obtained prior to 2006; the year the relicensing Safety Analysis Report (SAR) was submitted. However, in order to better benchmark this type of analysis, and also to hopefully document a loss of primary coolant flow event that occasionally occurs at MURR from a loss of normal site electrical power, a high-speed, portable, digital data recorder was procured in 2007. The recorder was then used to perform more precise response time measurements for the various transmitters and sensors in the reactor coolant systems that provide the signals for valve action and for valve actuation (the time from fully closed to fully open or from fully open to fully closed). Response time measurements were performed for anti-siphon isolation valves V543A and V543B, in-pool heat exchanger isolation valves V546A and V546B, and primary coolant isolation valves V507A and V507B. Anti-siphon isolation valves V543A and V543B start to open from their fully closed positions 0.2 seconds after reactor core outlet pressure transmitter PT-944A and/or PT-944B sends a signal to the valve actuators. The valves are then fully open 0.4 seconds after the valve action signal (open signal) is received. Including the measured pressure transmitter response time of 0.2 seconds, it will take 0.6 seconds for valves V543A and V543B to reach their fully open positions.
In-pool heat exchanger isolation valves V546A and V546B are fully open 0.8 seconds after core inlet flow rate decreases to 3,200 gpm (12,113 1pm), the trip set point of differential pressure across the reactor core sensor DPS-929.
The anti-siphon system analysis simulating a primary coolant system piping shear on the reactor core side of each primary coolant isolation valve was performed again with the revised opening times for valves V543A and V543B and valves V546A and V546B. The initiating event was followed by a low pressure trip which actuated the opening of valves V543A and V543B and also caused primary coolant circulation pumps P-501A and P-501B to secure (Relays 2K13 and 2K28 de-energize). When primary coolant flow through the core reduces to 3,200 gpm (12,113 lpm),
differential pressure across the reactor core sensor DPS-929 sends a signal to open in-pool heat exchanger isolation valves V546A and V546B.
It was determined that an anti-siphon system pressure of 0 psig (101.3 kPa) would be sufficient to ensure that at least 5 feet (1.52 m) of water remained above the top of the reactor core after a primary coolant piping rupture.
Primary coolant temperature was assumed to be 80 'F (27 'C) for these runs. As shown in Figure 6.5, "Simplified Diagram of the Primary Coolant System," of the SAR, the centerline of the horizontal primary coolant system inlet piping after core inlet check valve V502 is 5.937 feet (1.81 m) above the core where it enters the pressure vessel.
Table 1 provides a comparison of the amount of coolant that will remain above the reactor core based on a measured pressure transmitter response delay time of 0.2 seconds - time for a valve action signal (open signal) to be sent to anti-siphon isolation valves V543A and V543B - for the 7 of 28
following four assumed anti-siphon system pressures: 0, 10, 20 and 26 psig (101.3, 170.3, 239.2 and 280.6 kPa).
Table 1 Pressure Transmitter Response Delay Time of 0.2 Seconds Anti-Siphon System Pressure Coolant Above the Core 0 psig (101.3 kPa) 5.110 feet (1.556 in) 10 psig (170.3 kPa) 5.443 feet (1.659 m) 20 psig (239.2 kPa) 5.939 feet (1.810 m) 26 psig (280.6 kPa) 5.900 feet (1.798 m)
Section 6.3.8, Design Analysis, will be revised accordingly based on these new values.
As stated above, the horizontal primary coolant system inlet piping centerline after core inlet check valve V502 is 5.937 feet (1.81 m) above the core where it enters the pressure vessel. With 26 psig (280.6 kPa) of air pressure in the anti-siphon system (1.0 psig below Technical Specification 3.4.b), Table 2 provides the coolant levels above the top of the core for the given time interval (assumed transmitter response delay time) between when a low pressure trip from reactor core outlet pressure transmitter PT-944A and/or PT-944B occurs due to a primary coolant system shear on the reactor side of both isolation valves and when the valve action signal is sent to open the anti-siphon isolation valves from their fully closed positions.
Table 2 Transmitter Response Delay Time vs. Coolant Level Above the Core Time Interval Coolant Level Above the Core (seconds)
(feet) 0.20 5.900 (1.798 m) 1.20 6.009 (1.832 in) 2.20 5.636 (1.718 m) 4.20 5.430 (1.655 m) 6.20 5.574 (1.699 in) 8.20 4.556 (1.389 m) 10.20 4.400 (1.341 m)
A delay in the opening action time of the anti-siphon isolation valves would need to be greater than 6 seconds to result in a coolant level of less than 5 feet (1.52 m) above the top of the core during a loss of coolant accident. An increase in the nonnal valve opening response time of 0.6 seconds would have to be caused by either a malfunction of both V543A and V543B valve actuators or some sort of electronic failure in both pressure transmitters PT-944A and PT-944B.
As stated in Section 6.3.4 of the SAR, operation of either valve will perform the intended function of the anti-siphon system.
Additionally, two pressure transmitters are installed for redundancy, either of which will cause the anti-siphon isolation valves to open. As stated in the response to RAI A.22, "At least two major failures must occur before the Anti-Siphon Actuation System would be unable to perform its intended function. This is not a credible assumption."
Technical Specification (TS) 3.4.b states that a minimum pressure of 27 psig (287.5 kPa) must be maintained in the anti-siphon system for Mode I or II operation; TS 3.9.a requires the anti-siphon 8 of 28
system to be operable when the reactor is operated in Mode I or II; and TS 4.2.b states that the anti-siphon isolation valves must be tested for operability at least monthly. The response to RAI A.46 discusses how the multiple TS requirements combine to confirm the operability of the anti-siphon system.
A delay in the valve action time of either anti-siphon isolation valve would easily be detected by the operators during the weekly primary coolant system startup, even before reaching two to three seconds. The response time of the trip signal from either pressure transmitter PT-944A or PT-944B would also easily be detected by the operators well before reaching three seconds. A six month TS surveillance is performed, as required by TS 4.4.a, to confirm calibration of the transmitters as well as verify operation of the anti-siphon isolation valve interlocks. Therefore, based on the above information, MURR feels that no additional TS wording on anti-siphon valve performance.is needed.
9 of 28
13.4 Section 13.2.3, Loss of Primary Coolant.
- b. Please describe any benchmarking that has been performed on the RELAP model and conclusions as to the accuracy of the model's results. Please discuss the effect on the analysis of oxide layer build up on the fuel cladding.
A loss of primary coolant, generally referred to as a Loss of Coolant Accident (LOCA), is an accident that has rarely occurred at nuclear reactors. Where they have occurred, the reactor was usually permanently shut down and decommissioned. Fortunately, MURR has never experienced this type of accident and therefore, has never had the opportunity to collect any actual data during a LOCA that could subsequently be used to benchmark the MURR RELAP5 model. However, a loss of primary coolant flow event has occurred several times at MURR over the past 45 years of operation, primarily due to a loss of normal site electrical power. When the RELAP5 modeled analyses were completed during the summer of 2006 in support of the relicensing Safety Analysis Report (SAR), no loss of flow event [also referred to as a Loss of Flow Accident (LOFA)] had been recorded for RELAP5 benchmarking. Realizing the benefits of collecting data of this type, as also stated in the responses to RAIs 6.2 and 13.7, a high-speed, portable, digital data recorder was procured in 2007. It was connected to the two (hot and cold legs) in-pool heat exchanger resistance temperature detectors (RTDs) to record the temperature transients of a loss of primary coolant flow event, if one were to occur. On April 12, 2008, MURR experienced a loss of flow event caused by a rupturing of the diaphragm on a primary coolant system throttle valve, thereby essentially causing a loss of pressure LOFA. This valve is located on the primary coolant 'A' loop just a few feet upstream from where the pressurizer connects to the primary coolant system.
This loss of flow event, due to a reduction in primary coolant system pressure, was similar to the worst-case LOFA analyzed and described in Section 13.2.4 of the SAR.
Figure 1 is a scale drawing of the in-pool portion of the primary coolant system. The reactor core is shown between the inner and outer reactor pressure vessels. During normal operation, primary coolant flows upward through the primary coolant cold leg piping and core inlet check valve V502, and then turns horizontally through a 900 elbow into the top of the outer pressure vessel.
Flow is then redirected downward in the pressure vessel, through the core region before exiting horizontally to the right. The coolant is then redirected upward into the inverted loop where it flows through an orifice plate just before making a sharp 90' turn into a branch of the piping tee and through the short horizontal piping section of the inverted loop. The coolant then flows through a long radius 90' elbow which directs it downward and out of the reactor pool to the hot leg primary coolant isolation valve located in the mechanical equipment room (Room 114), which contains the primary coolant system circulation pumps and heat exchangers. During a LOFA, normal, forced primary coolant flow is lost and natural circulation flow through the in-pool heat exchanger is initiated by the opening of in-pool heat exchanger isolation valves V546A and V546B. (Note: Since this is a 1966 drawing of the original piping arrangement, Figure 1 only depicts one of the in-pool heat exchanger isolation valves. A second valve and associated piping were installed in 1974 in parallel to the original valve as part of the uprate in power to 10 MW.)
10 of 28
V546 iJ v46 Natural Circulation 1/2-inch Bypass Vertical Piping Line Extension Inverted Loop In-Pool Heat Exchanger 7.020-inch Orifice Plate Inlet Check Valve V502"---
V Outer Reactor Pressure Vessel Primary Coolant Reactor Core Outlet Piping Figure 1 - In-Pool Portion of the Primary Coolant System With a loss of forced flow, the coolant in and around the reactor core heats up, thus causing a flow reversal through the core. The hot coolant flows upward and out of the pressure vessel as if it were returning to core inlet check valve V502 but instead flows through 6-inch diameter piping that is connected to the inlet of the in-pool heat exchanger isolation valves which then connects to the in-pool heat exchanger upper 6-inch diameter horizontal piping header. The heated coolant then flows down through the ten vertical in-pool heat exchanger finned tubes and into the lower 6-inch diameter horizontal piping header, which connects back to the natural circulation vertical piping extension above the inverted loop. The now cooler primary coolant returns to the core by flowing downward through the normal upward flow path of the inverted loop and into the bottom of the pressure vessel. Flow is caused by natural circulation resulting from a difference in density of the coolant between the hot and cold legs and the vertical lengths of the two flow paths.
The two RTDs that are installed in the in-pool heat exchanger piping measure and display the temperature difference between in-pool heat exchanger coolant inlet (Th) and outlet (T,). In the 11 of 28
event of a LOFA, the in-pool heat exchanger inlet RTD becomes Tb after core flow reversal occurs since it is connected to the primary coolant cold leg. The high-speed, portable, digital data recorder was attached to each RTD to record wiring loop resistance and convert it to a temperature reading based on resistance vs. temperature. However, this reading also includes the resistance of the wiring and connection points from the control room down into the reactor pool to where the RTDs are located and then back to the control room. This added line resistance would increase indicated temperature by at least a few degrees.
A resistance of 1.3 ohms increases temperature by 6 'F (3.3 °C); therefore, the recorded RTD temperatures were corrected accordingly.
A 1/22-inch diameter bypass line around each of the in-pool heat exchanger isolation valves allows a very limited amount of primary coolant, at a nominal temperature of 120 'F (48.9 'C), to flow through the in-pool heat exchanger during normal operation.
The in-pool heat exchanger is located in the bulk reactor pool which is typically at a temperature of 100 to 105 'F (37.8 to 40.6
'C).
Primary coolant at a temperature of approximately 120 'F (48.9 'C) will flow directly through the open in-pool heat exchanger isolation valves immediately after a LOFA initiates.
Just before the loss of flow event occurred on April 12, 2008, the wiring loop and RTD combined resistance for the in-pool heat exchanger Th leg corresponded to a temperature reading of 126 'F (52.2 'C).
Additionally, primary coolant at a temperature of 120.1 'F (48.9 'C) was flowing through the coolant piping that is connected to the in-pool heat exchanger's inlet piping.
To better understand what the actual coolant temperature is at the in-pool heat exchanger outlet (TJ) just before the start of a LOFA, a closer examination of the piping arrangement and flow dynamics during normal reactor operation is required. Consider Figure 1 once again. On April 12, 2008, just before the loss of flow event occurred, the temperature of the primary coolant at the core outlet was 136.2 'F (57.9 'C). This is the temperature of the coolant that is flowing upward in the inverted loop and through the 7.020-inch orifice plate. The vena contracta occurs at an elevation that corresponds to a point in the vertical 12-inch diameter primary coolant piping where coolant flow takes an immediate 900 turn out of the piping tee on top of the inverted loop.
Therefore, the higher velocity center flow stream will still be expanding and pushing flow upward past the horizontal outlet of the piping tee and into the 12-inch diameter natural circulation vertical piping extension above the inverted loop.
This, combined with natural convection, promotes heavy mixing in the 12-inch vertical piping extension, which keeps the temperature of the primary coolant in that section of the piping well above the temperature of the pool coolant surrounding the piping. The actual in-pool heat exchanger coolant outlet temperature reading indicates that a sufficient amount of this coolant is mixing or flowing into the in-pool heat exchanger lower horizontal (outlet) piping header, thus making the outlet RTD (TJ) read approximately one degree higher than the inlet RTD (Th).
With a resistance corrected in-pool heat exchanger coolant inlet temperature of approximately 120 °F (48.9 °C), it appears that a natural convection flow path within the in-pool heat exchanger may be created during normal operation (valves V546A and V546B closed). The convection flow loop would start with the coolant in the primary hot leg entering the in-pool heat exchanger lower 6-inch diameter horizontal header (in the upper half of the piping cross-sectional area) and flowing upward through the first few vertical heat exchanger finned tubes. Coolant would then flow into the upper 6-inch diameter horizontal (inlet) piping header towards the closed in-pool 12 of 28
heat exchanger isolation valves. The natural convection flow loop would then be completed with downward coolant flow through the last few in-pool heat exchanger finned tubes near valves V546A and V546B and then back towards the outlet RTD in the lower 6-inch horizontal header (in the lower half of the piping cross-sectional area). Flow would be caused by a difference in density of the primary coolant created by the transfer of heat to the cooler pool coolant. The actual recorded temperatures indicate either this or a similar flow path may exist within the in-pool heat exchanger. However, this cannot be modeled in RELAP5 due to the limitations of the ID flow code.
The following table and figures provide a comparison of the temperature transients for the in-pool heat exchanger from the April 12, 2008, recorded loss of flow event and the RELAP5 modeled LOFA analysis that was performed for the relicensing SAR. In-pool heat exchanger coolant inlet (Th) and outlet (TJ) temperatures for the loss of pressure, loss of flow event are shown in Figures 2 and 3. These values were recorded using the high-speed, portable, digital data recorder and corrected for the added resistance of the wiring loop and connection points.
Figure 2 also includes results from the RELAP5 modeled worst-case LOFA using the conservative assumptions as described in Section 13.2.4.1 (and Table 1 below). Agreement of the initiating temperatures between the RELAP5 model and the loss of flow event is coincidental. For the RELAP5 LOFA model, pool coolant is at a temperature of 120 *F (48.9 'C) whereas reactor inlet water temperature is 155 *F (68.3 'C), thus in-pool heat exchanger temperatures are in equilibrium with the surrounding pool coolant. As previously described, the in-pool heat exchanger temperatures are approximately 120 *F (48.9 'C) with a pool coolant temperature of about 100 *F (37.8 'C).
Figure 3 includes the results from the RELAP5 model; however, instead of using the worst-case LOFA conservative assumptions as described in Section 13.2.4.1, the actual operating values when the loss of flow event occurred on April 12, 2008 are used. Once again the RELAP5 model has the in-pool heat exchanger starting temperatures equal to the pool coolant temperature. Table I provides a numerical comparison of the LOFA conservative assumptions and the April 12, 2008 actual operating values.
Table 1 Comparison of Conservative Assumptions and April 12, 2008 Actual Operating Values Parameter Conservative Assumptiodn April 12, 2008 Actual Operating Values Reactor Power Level 11 MW 9.95 MW Reactor Inlet Water Temperature 155 'F (68 -C) 120.1 'F (48.9 QC)
Core Flow Rate 3,800 gpm (14,385 1pm) 3,793 gpm (14,359 1pm)
Pool Coolant Temperature 120 'F (49 °C) 100.2 'F (37.9 °C)
Pressurizer Pressure 60 psig (413.7 kPa)1 69.8 psig (481.3 kPa)'
Anti-Siphon System Pressure 26 psig (179.3 kPa)'
36 psig (248.2 kPa)'
Note 1: Pressure above atmosphere.
The event starting point corresponds to 500 seconds on both graphs.
However, there is a difference in starting temperatures before the event begins in Figure 3. The RELAP5 modeled transient starts with in-pool heat exchanger coolant inlet and outlet temperatures the same as reactor pool water temperature. This is reasonable because the RELAP5 modeled temperatures 13 of 28
are based on a no flow condition in the in-pool heat exchanger - valves V546A and V546B closed (Note: The RELAP5 model does not include the 72-inch diameter bypass lines around valves V546A and V546B). With no flow of primary coolant through the in-pool heat exchanger, the coolant is at an equilibrium temperature with the pool water surrounding the heat exchanger until the in-pool heat exchanger isolation valves open.
Figure 2 shows very similar temperature response times between the loss of flow event recorded on April 12, 2008 and the RELAP5 modeled LOFA using the conservative assumptions listed in Table 1. The temperature peaks for the RELAP5 modeled LOFA are obviously much higher than the actual event because of the very conservative assumptions at the start of the transient: higher reactor power level, and higher reactor inlet and pool coolant temperatures. The higher power level and coolant temperatures also cause the temperature leaks for both in-pool heat exchanger coolant inlet and outlet to occur sooner because the temperature difference is what drives the flow rate.
Figure 2, therefore, shows how conservative the RELAP5 modeled worst-case LOFA actually is.
Benchmark of Loss of Flow Accident RELAP5 Model with Conservative Assumptions vs. 4/12/2008 Recorded Values 190.00 180.00 160.00 150.00 SRELAP HX Inlet RELAP HX Outlet R4/12/08 HX Inlet 130.00 4/12/08 HX Outlet 120.00 110.00 100.00 90.00 400 500 600 700 800 900 1000 1100 1200 1300 1400 lime is) - Event starts at 500 seconds Figure 2 - RELAP5 Model with Conservative Assumptions vs. 4/12/2008 Recorded Values Figure 3 shows very similar temperature response times between the loss of flow event recorded on April 12, 2008, and the RELAP5 modeled LOFA using the actual operating values listed in Table 1. The temperature peaks for the RELAP5 modeled LOFA are very close to the actual event. The higher temperature difference between the in-pool heat exchanger coolant inlet and outlet temperatures in the RELAP5 model indicate that there is a greater transfer of heat from the 14 of 28
primary coolant through the inner and outer pressure vessels and into the reactor pool than the RELAP5 modeling provides due to the limitations of the ID flow code. This also supports how conservative the RELAP5 modeled LOCA is. In the LOCA analysis no coolant flows through the in-pool heat exchanger. All of the heat transfer from the primary coolant around the reactor core region is through the inner and outer pressure vessels.
Benchmark of Loss of Flow Accident RELAP5 Model with Actual Operating Parameters vs. 4/12/2008 Recorded Values 190-170 160 150
-RELAP HX Inlet
-2 140
-RELAP HX Outlet
-4/12/08 HX Inlet 130
/2OH~te 120 110-----
100 400 So00 0
700 S00 900 l000 1100 1200 1300 1400 Time (s) - Event starts at 500 seconds Figure 3 - RELAP5 Model with Actual Operating Values vs. 4/12/2008 Recorded Values As discussed in the response to RAI 13.4.a, the oxide buildup on the fuel plates over fuel element lifetime only tends to maintain the peak centerline temperatures as a function of power history on the element. The highest heat flux is in a new fuel element with very little oxide buildup. The high fuel plate centerline temperature is due to the higher differential temperature across the fuel meat, cladding and coolant film coefficient caused by this higher heat flux. At the end of a fuel element lifetime the differential temperature across the oxide buildup approaches the reduction in the differential temperature across the fuel meat, cladding and coolant film coefficient caused by the reduction in the heat flux.
15 of 28
13.7 Section 13.2.4.2, page 13-55. Explain why assuming longer rather than shorter close times for the isolation valves is a conservative assumption for this accident.
The assumption of longer rather than shorter closure times for primary coolant isolation valves V507A and V507B was conservative for the Loss of Coolant Accident (LOCA), but not necessarily for the Loss of Flow Accident (LOFA). The valve closure times provided on page 13-55 of the Safety Analysis Report (SAR) were based on the best time measurements that were obtained prior to 2006; the year the relicensing SAR was submitted. Realizing the benefits of collecting data of this type, as also stated in the responses to RAIs 6.2 and 13.4.b, MURR procured a high-speed, portable, digital data recorder in 2007. The recorder was then used to perform more precise response time measurements for the various transmitters and sensors in the reactor coolant systems that provide the signals for valve action and for valve actuation (the time from fully closed to fully open or from fully open to fully closed). Response time measurements were performed for primary coolant isolation valves V507A and V507B, anti-siphon isolation valves V543A and V543B, and in-pool heat exchanger isolation valves V546A and V546B.
Primary coolant isolation valve V507A starts to close from its fully open position 4.0 seconds after reactor core outlet pressure transmitter PT-944A and/or PT-944B sends a signal to the valve actuator. The valve is then fully closed 8.8 seconds after the valve action signal is received.
Whereas primary coolant isolation valve V507B valve starts to close from its fully open position 3.7 seconds after the pressure transmitter sends a signal to the valve actuator. The valve is then fully closed 8.9 seconds after the valve action signal is received. Because of the response delay time of the pressure transmitters, it will take 0.2 seconds for a valve action signal (close signal) to be sent to valves V507A and V507B after system pressure reaches the transmitter trip set point of 28.5 psig. Consequently, valves V507A and V507B will reach their fully closed positions in (0.2
+ 8.8) 9.0 seconds and (0.2 + 8.9) 9.1 seconds, respectively. Therefore, the flow path for the primary coolant loop is interrupted when valve V507A is fully closed. Consequently, the valve closure time for valves V507A and V507B used in the LOFA RELAP5 model is based on valve V507A.
Anti-siphon isolation valves V543A and V543B start to open from their fully closed positions 0.2 seconds after pressure transmitter PT-944A and/or PT-944B sends a signal to the valve actuators.
The valves are then fully open 0.4 seconds after the valve action signal is received. In-pool heat exchanger isolation valves V546A and V546B start to open from their fully closed positions 0.2 seconds after differential pressure across the reactor core sensor DPS-929 sends a signal to the valve actuators.
The valves are then fully open 0.6 seconds after the valve action signal is received. Including the response delay time of pressure transmitters PT-944A and/or PT-944B, anti-siphon isolation valves V543A and V543B will reach their fully open positions (0.2 + 0.4) 0.6 seconds after the trip set point is reached. Including the response time of differential pressure sensor DPS-929, in-pool heat exchanger isolation valves V546A and V546B will reach their fully open positions (0.2 + 0.6) 0.8 seconds after the trip set point is reached.
In the LOFA, the source of heat is the decay heat from the fuel meat in the fuel plates. The purpose of this analysis is to verify that the fuel plates do not reach a temperature that could cause damage. Figure 1 below corresponds to Figure 13.26, CENTERLINE TEMPERATURE OF THE 24 FUEL PLATES DURING THE FIRST 60 SECONDS OF THE LOF ACCIDENT, of the 16 of 28
SAR. As stated in the third paragraph of page 13-55 of the SAR, this analysis assumes a valve closure time of 9.5 seconds for primary coolant isolation valves V507A and V507B after the LOFA starts. The RELAP5 model assumed the following: a valve action time of 4.5 seconds and valve actuation time of 5.0 seconds for valves V507A and V507B; a valve action time of 0.085 seconds and valve actuation time of 0.1 seconds for valves V543A and V543B; and a valve action times of 0.135 seconds and valve actuation time of 0.25 seconds for valves V546A and V546B.
The graph shows the centerline temperatures as a function of time in the third section of all 24 fuel plates during the LOFA. Each fuel plate is associated with four heat structures, or sections (top 7.75 inches, the 5 inches above core centerline, the 5 inches below core centerline, and the bottom 7.75 inches). This third section corresponds to the section of the fuel plate with the highest decay heat level and therefore the highest temperatures.
RELAP5 Model Centerline Temperatures of the 24 Fuel Plates C_ I fVA U16~ U7.I.,.
T.r.
V~.
V507A/B 9.5 s (dose); V543A/B 0.185 s (open); V546A/B 0.385 s (open) 290
-M ate 1.3 280-Plate 2.3 J,
Plate 3.3 270 260-Plate 5.3 Plate 6.3 250-PIat 7.3
-- Plate 8.3 240 Plate 9.3 230 Plate 10.3 Plate 11.3
-20Plate 12.3 Plate 13.3
-- Plate 14.3 200 Plate 15.3
-- Plate 16.3 Plate 17.3 IS0O Plate 18.3 Plate 19.3 170 Plate 20.3 160 Plate 22.3 0
5 10 15 20 25 30 35 40 45 50 55 60 Plate23.3 P-Plte 24.3 Time (seconds)
Figure 1 - Centerline Temperature of the 24 Fuel Plates during the First 60 Seconds of a LOFA (Figure 13.26 of the SAR)
To test the sensitivity of peak fuel plate temperatures to valve closure times, additional RELAP5 runs were performed. Figure 2 provides the results of a LOFA with valve action and actuation times that correspond to the benchmark recorded values and a transmitter response delay time of 0.2 seconds for PT-944A, PT-944B and DPS-929. The revised RELAP5 model assumed the following: a valve action time of 4.2 seconds and valve actuation time of 4.8 seconds for valves V507A and V507B; a valve action time of 0.40 seconds and valve actuation time of 0.20 seconds for valves V543A and V543B; and a valve action time of 0.40 seconds and valve actuation time of 0.40 seconds for valves V546A and V546B.
Therefore, primary coolant isolation valves V507A and V507B will close in 9.0 seconds; anti-siphon isolation valves V543A and V543B will open in 0.6 seconds; and in-pool heat exchanger isolation valves V546A and V546B will open in 17 of 28
0.8 seconds after total core flow rate decreases to 3,200 gpm (12,113 lpm).
Table 1 provides a summary of the action and actuation times for the various valves using the times assumed in the SAR and the benchmark recorded values.
Table 1 Valve Action and Actuation Times - SAR vs. Benchmark Recorded Values Component I
Figure 1 Figures 2 and 3 Valves V507A and V507B Start to Close 4.5 seconds 4.2 seconds Fully Closed J
9.5 seconds 9.0 seconds Valves V543A and V543B Start to Open 0.085 seconds 0.40 seconds Fully Open 0.185 seconds 0.60 seconds Valves V546A and V546B Start to Open 0.135 seconds 0.40 seconds Fully Open 0.385 seconds 0.40 seconds RELAP5 Model Centerline Temperature of the 24 Fuel Plates for a LOFA with Valve Times of:
V507A/B 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 (open) 290 270 20Plate 8.3 230plate2.3
-Plate 1.3 Plate 7.3 Plate 8.3 Plate 9.3 Plate 10.3 Plate 15.3 Plate 12.3 F
Plate 13.3 Plate 14.3 200 Plate 15.3 Plate 16.3 170 Plate 17.3 180
-Plate 18.3
-Plate 19.3 170 Plate 20.3
......Plate 21.3 160 Plate 22.3 150Plate 23.3 Soo 505 510 515 520 525 530 535 540 545 550 555 560 565 570 Plate 24.3 Time (seconds)
Figure 2 - Centerline Temperature of the 24 Fuel Plates during the First 70 Seconds of a LOFA (Using Benchmark Valve Action and Actuation Times)
In comparing Figures 1 (SAR LOFA analysis) and 2 (LOFA analysis using benchmark recorded values), there is minimal difference between the highest fuel plate centerline temperatures and the timing of their peaks (Note: For Figures 2 and 3, the event starts at the 500 second point). The 18 of 28
highest fuel plate centerline temperature of 278 'F (136.7 'C) occurs in plate number-22, 17 seconds after the event starts.
Fuel plate number-23 has the next highest temperature of approximately 272 'F (133.3 °C), which occurs a fraction of a second earlier. Compare these values to the values stated in the SAR where the highest fuel plate centerline temperature of 280.3
'F (137.9 'C) occurs in plate number-i, 0.3 seconds into the transient. Fuel plate centerline temperatures then decrease as reactor power decreases from the insertion of the control blades.
After the first second of the transient, the highest centerline temperature of 277.9 'F (136.6 °C) occurs in plate number-22, 17 seconds after the transient starts.
The conservative assumptions included in the SAR LOFA analysis can also be seen in Figure 3, which shows the RELAP5 calculated fuel plate centerline temperatures for the loss of flow event that occurred on April 12, 2008, using the same valve action and actuation times used in Figure 2.
The highest fuel plate centerline temperature of 246.3 'F (119.1 'C) occurs in plate number-1, 0.3 seconds into the transient. After the first second of the transient, the highest fuel plate centerline temperature still occurs in plate number-22, but is only about 235 *F (112.8 *C), which is 43 *F (23.9 *C) below the peak shown in Figure 2.
RELAP5 Model Centerline Temperatures of the 24 Fuel Plates for the Loss of Flow Event that occurred on April 12, 2008, with Valve Times of:
V507A/B 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 s (open)
-Plate 1.3 250 4Plate 3.3 240 Plate 4.3 235 2
V~
IPlate 5.3 225 Plate 6.3 220 Plate 7.3 215 Plate 8.3 2105 Plate 9.3
~200...
Plate 10.3
-Plate 11.3
-Plate 13.3
-- P 185.
Plate 14.3 175 Plate 14.3 Plate 16.3 165 160 5 Plate 18.3
+30 Plate 1.3 tnn tn5 tin tit t20 525 530 535 540 545 550 555 560 565 570 Plate 24.3 lime (seconds)
Figure 3 - Centerline Temperature of the 24 Fuel Plates during the First 70 Seconds of a LOFA (April 12, 2008 Loss of Flow Event Using Benchmark Valve Action and Actuation Times)
Therefore, changing the valve action and actuation times in the RELAP5 model to the benchmark measured values, which includes a shorter closure time for primary coolant isolation valves V507A and V507B, does not have any significant effect on peak centerline fuel plate temperatures.
19 of 28
APPENDIX C C. 1 Section C. 1, Introduction. Please describe any benchmarking that has been performed on the RELAP model and conclusions as to the accuracy of the model's results.
The responses to questions 6.2, 13.4.b and 13.7 address any benchmarking that was performed on the MURR RELAP5 model.
C.3 Section C.2.2, Modeling of the MURR. Please explain why initial conditions of the analysis are not TS and license limits. Please discuss the effect on the analysis of fuel burnup and oxide layer build up on the fuel cladding.
The conservative assumptions used in the Loss of Coolant Accident (LOCA) analysis are provided in Table 13-1 of the Safety Analysis Report (SAR) along with the normal reactor operating conditions. The same conservative assumptions and normal operating conditions for the Loss of Flow Accident (LOFA) analysis are given in Table 13-7.
Table 1 Normal Reactor Operating Conditions and Conservative Assumptions When a LOCA or LOFA Initiates (Tables 13-1 and 13-7 of the SAR)
Parameter Conservative Assumption Normal Condition Reactor Power Level 11 MW 10 MW Core Inlet Temperature 155 °F (68 °C) 120 -F (49 -C)
Core Inlet Flow Rate 3,800 gpm (14,385 1pm) 3,800 gpm (14,385 1pm)
Pool Temperature 120 -F (49 -C) 100 -F (38 -C)
Pressurizer Pressure 60 psig (414 kPa)'
62 - 66 psig (427 - 455 kPa)'
Anti-Siphon Pressure 26 psig (179 kPa)'
36 psig (248 kPa)'
Note 1: Pressure above atmosphere.
Figure 13.26 of the SAR, included as Figure 1 below, provides the centerline temperatures as a function of time in the third section of all 24 fuel plates. Each fuel plate is associated with four heat structures, or sections (top 7.75 inches, the 5 inches above core centerline, the 5 inches below core centerline, and the bottom 7.75 inches). This third section corresponds to the area of the fuel plate with the highest decay heat level and therefore the highest temperatures. Figure 13.27 of the SAR shows the temperatures of the fourth volume, which is the volume from 5 inches below core centerline to the bottom of the fuel plates, of all 25 coolant channels as a function of time. This figure is included as Figure 2.
20 of 28
RELAP5 Model Centerline Temperature of the 24 Fuel Plates for a LOFA with Valve Times of:
V507A/B 9.5 s (close); V543A/B 0.185 s (open); V546A/B 0.385 s (open) 290 Plate 1.3 280 Plate 2.3 Plate 3.3 27 Plate 4.3 260 Plate 5.3 Plate 6.3 250 Plate 7.3 Plate 8.3 240 Plate 9.3 t'
230 Plate 10.3 Plate 11.3 1 220 210 Plate 13.3 E
Plate 14.3 20Plate 1.3 Plate 16.3 100 Plate 15.3 Plate 17.3 180 Plate 18.3 Plate 19.3 170 Plate 20.3 160 Plate 21.3
-- Plate 23.3 Time (seconds)
Figure 1 - Centerline Temperature of the 24 Fuel Plates during the First 60 Seconds of a LOFA (Figure 13.26 of the SAR)
RELAP5 Model Temperature of the 25 Individual Coolant Channels for a LOFA with Valve Times of:
V507A/B 9.5 s (close); V543A/B 0.185 s (open); V546A/B 0.385 s (open) 250 Channel 1.4 245
-_"Channel 2.4 240 Channel 3.4 235 Channel 4.4 230 Channel 5.4 225 Channel 6.4 220 Channel 7.4 Channel 8.4 215 Channel 9.4 210 t6 5 Channel 10.4 205 Channel 11.4 S200 Channel 12.4 C. 195 Channel 13.4 190 Channel 14.4 185 Channel 15.4 180 Channel 16.4 Channel 17.4 175 Channel 18.4 170 Channel 19.4 165 Channel 20.4 160 Channel 21.4 155 Channel 22.4 ISO Channel 23.4 0
S 10 15 20 25 30 35 40 45 50 55 60 Channel 24.4 lime ',sernnds Channel 25.4 Figure 2 - Temperature of the 25 Individual Coolant Channels during the First 60 Seconds of a LOFA (Figure 13.27 of the SAR) 21 of 28
Figures 3 and 4 provide the same type of graphs but were generated with RELAP5 using the benchmark measured valve action and actuation times. After system pressure reaches the trip set point of reactor core outlet pressure transmitter PT-944A and/or PT-944B, primary coolant isolation valves V507A and V507B are fully closed in 9.0 seconds whereas the anti-siphon isolation valves V543A and V543B are fully open in 0.6 seconds.
In-pool heat exchanger isolation valves V546A and V546B are fully open 0.8 seconds after core inlet flow rate decreases to 3,200 gpm (12,113 1pm), the trip set point of differential pressure across the reactor core sensor DPS-929.
RELAP5 Model Centerline Temperature of the 24 Fuel Plates for LOFA with Table I Conservative Assumptions and Valve Times of:
V507A/B 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 s (open) 300
-Plate 1.3 9
Plate 2.3 280 Plate 3.3 Plate 4.3 270 Plate 5.3 260 Plate 6.3
-Plate 7.3 250 Plate 8.3 240 Plate 9.3
-Plate 10.3 20 i-Plate 11.3
-Plate 13.3 210 Plate 14.3 Plate 15.3
-Plate 16.3 Plate 17.3
-Plate 18.3
-Plate 19.3 70 Plate 20.3 Plate 21.3 Plate 22.3 1z.
Plate 23.3 500 505 510 515 520 525 530 535 540 545 550 555 560 565 570 Plate 24.3 Time iseconds)
Figure 3 - Centerline Temperature of the 24 Fuel Plates during the First 60 Seconds of a LOFA (Using Table 1 Conservative Assumptions and Benchmark Valve Action and Actuation Times) 22 of 28
RELAP5 Model Temperature of the 25 Coolant Channels for LOFA with Table 1 Conservative Assumptions and Valve Times of:
V507A/B 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 s (open) 250 11Channel 1.4 245e Channel 2.4 24Channel 3.4 Channel 4.4 235 Channel 5.4 SChannel 6.4 2-Channel 7.4 Channel 6.4 225Channel 9.4 210~-
Channel 10.4 SChannel 11.4 Channel 12.4 Channel 13.4 2-Channel 14.4 1Channel 11.4 85-Channe16.4 4o Channel 17.4 200nl 0.
1-Channel 18.4 SChannel 19.4
-Channel 20.4 180
-Channel 217.4 160
-Channel 22.4 Channel 24.4 500 510 520 530 540 550 560 570 580 590 Channel 25.4 Time (seconds)
Figure 4 - Temperature of the 25 Individual Coolant Channels during the First 60 Seconds of a LOFA (Using Table I Conservative Assumptions and Benchmark Valve Action and Actuation Times)
In comparing these two sets of graphs, there are minimal differences noted. In Figures 1 and 3, the peak centerline temperature of 278 *F (136.7 *C) occurs in fuel plate number-22, around 17 seconds after the LOFA starts. Due to the 0.5 second difference in the closure times of valves V507A and V507B for the two different RELAP5 models, fuel plate number-2 centerline temperature peaks about 0.5 seconds later in Figure 1 than it does in Figure 3 and about 3 *F (1.7
- C) higher. In Figures 2 and 4, the peak coolant temperature of 237 *F (113.9 *C) occurs in coolant channel 19, around 22 seconds after the LOFA starts. One again, due to the 0.5 second difference in the closure times of valves V507A and V507B, coolant temperature in channel 3 peaks about 0.5 seconds later in Figure 2 than it does in Figure 4 and about 3 *F (1.7 *C) higher.
The conservative assumptions listed in Table 13-7, which were used in the RELAP5 analysis to generate the results given in Figures 1 through 4, will be compared to the Technical Specifications (TSs) and license limits:
Reactor Power: Amended Facility License R-103, issued by the U.S. Nuclear Regulatory Commission (NRC) to MURR, limits reactor power to maximum steady-state level of 10 MW. To be conservative, the steady-state power level is assumed to be 11 MW, which produces 10 % more decay heat during both the LOCA and the LOFA.
23 of 28
Coolant Inlet Temperature:
TS 2.2, Limiting Safety System Settings, and TS 3.3.a, Reactor Safety System, require a Reactor Inlet Water Temperature scram with a set point of 155 'F (Max), which is used as the steady-state coolant inlet temperature in the LOCA and LOFA analyses.
Core Inlet Flow Rate: TS 2.2, Limiting Safety System Settings, and TS 3.3.a, Reactor Safety System, require a core inlet flow rate (Differential Pressure Across the Core) scram with a set point of 3,200 gpm (Min); however, 3,800 gpm is used as the steady-state core inlet flow rate in the LOCA analysis. A higher flow rate is more conservative because it will result in less coolant remaining in the U-shaped primary coolant piping containing the reactor core. For the LOFA analysis in the original SAR submittal, this same flow rate was used. However, a lower initial flow rate for the LOFA analysis might be more conservative and will be discussed later in this response.
Pool Temperature: TS 3.4.b, Instrumentation, limits Reactor Pool Temperature to 120 'F (Max). Since the reactor pool serves as the heat sink during a LOCA and LOFA, it is conservative to use this maximum temperature limit as the steady-state pool temperature at the start of both of these accidents.
Pressurizer Pressure: TS 2.2, Limiting Safety System Settings, and TS 3.3.a, Reactor Safety System, require a Primary Coolant Low Pressure scram with a set point of 75 psia (Min). The RELAP5 model uses 60 psig as the steady-state pressurizer pressure. This is conservative since the worst-case LOCA is a shear of the primary coolant inlet piping (cold leg ) between the primary coolant isolation valve and the reactor pool, which causes primary coolant system pressure to reduce to 0 psig on the core side of the inlet isolation valve. Since the shear is in the cold leg, a lower pressurizer pressure reduces the pressure in the primary coolant outlet piping (hot leg) thus decreasing the pressure differential which helps to slow down the primary coolant flowing through the core. A pressurizer pressure of 60 psig is also conservative and of no significance for the LOFA because the worst-case accident is a loss of pressure LOFA.
Anti-Siphon Pressure: TS 3.4.b, Instrumentation, limits Anti-Siphon System Pressure to 27 psig (Min); therefore, it is conservative to use 26 psig as the steady-state anti-siphon system pressure at the start of the LOCA. Since a lower anti-siphon system pressure means less air will be injected into the inverted primary coolant loop, this assumption is conservative.
Based on this review, it was first considered to model the LOFA with a core inlet flow rate at the Limiting Safety System Setting (LSSS) of 3,200 gpm (12,113 lpm). Differential pressure across the reactor core sensor DPS-929 has a set point corresponding to a flow rate of 3,200 gpm (12,113 lpm). For normal operation, DPS-929 is set at a differential pressure corresponding to a slightly higher flow rate of 3,400 gpm (12,870 1pm).
The RELAP5 model uses a value of 3,200 gpm (12,113 lpm) as the core inlet flow rate. With a new conservative assumption of 3,260 gpm (12,340 lpm) and a reactor power level of 11 MW, core outlet temperature would be 175.5 'F (79.7 *C), which is slightly greater than TS 3.3.a 24 of 28
Reactor Outlet Water Temperature scram set point of 175 'F (79.4 *C). Therefore, core inlet flow rate is set to a value of 3,260 gpm (12,340 lpm), which is just above the LSSS and safety system set point of 3,200 gpm (12,113 lpm) and results in core outlet temperature being slightly greater than the set point of TS 3.3.a (175 'F). Table 2 is the revised SAR Table 13-7 for re-evaluation of the LOFA.
Table 2 Normal Reactor Operating Conditions and Conservative Assumptions when the LOF Accident Initiates (Revised Table 13-7 of the SAR)
Parameter Conservative Assumption Normal Condition Reactor Power Level 11 MW 10 MW Core Inlet Temperature 155 -F (68 -C) 120 -F (49 -C)
Core Inlet Flow Rate 3,260 gpm (12,113 lpm) 3,800 gpm (14,385 1pm)
Pool Temperature 120 -F (49 -C) 100 -F (38 -C)
Pressurizer Pressure 60 psig (414 kPa)'
62 - 66 psig (427 - 455 kPa)'
Anti-Siphon Pressure 26 psig (179 kPa)1 36 psig (248 kPa)'
Note 1: Pressure above atmosphere.
Again, Figures 3 and 4 were generated with RELAP5 using the conservative assumptions provided in Table 1 (Table 13-7 of the SAR) and the benchmark measured valve action and actuation times. Figures 5 and 6 were generated using the same benchmark measured valve action and actuation times; however, the new conservative assumptions listed in Table 2 were used.
25 of 28
RELAP5 Model Centerline Temperature of the 24 Fuel Plates for LOFA with Table 2 Conservative Assumptions and Valve Times of:
V507A/B 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 s (open) 300
!k 290
-Plate 1.3 Plate 2.3 280 Plate 3.3 Plate 4.3
-Plate 5.3 260 Plate 6.3 Plate 7.3 Plate 8.3 240
-Plate 9.3
-Plate 10.3 2-Plate 11.3 220
-Plate 12.3
-Plate 13.3
-Plate 14.3 20o Plate 15.3 Plate 16.3 Plate 17.3 1-Plate 18.3
-Plate 19.3 170 Plate 20.3 Plate 21.3 Plate 22.3 1S0 Plate 23.3 500 505 510 515 520 525 530 535 540 545 550 555 560 565 570 Plate 24.3 Time (seconds)
Figure 5 - Centerline Temperature of the 24 Fuel Plates during the First 60 Seconds of a LOFA (Using Table 2 Conservative Assumptions and Benchmark Valve Action and Actuation Times) 26 of 28
RELAP5 Model Temperature of the 25 Coolant Channels for LOFA with Table 1 Conservative Assumptions and Valve Times of:
-Channel 1.4 V507AIB 9.0 s (close); V543A/B 0.6 s (open); V546A/B 0.8 s (open)
-Channel 2.4 250 Channel 3.4
-channel 4.4 2-Channel 5.4 23Channel 6.4
-Channel 7.4 2-Channel 6.4 SChannel 11.4 225-Channel 81.4 220 Channel 92.4 200 Channel 13.4 210
-Channel 14.4 2-Channel 15.4
-Channel 16.4 205
-Channel 17.4 195Channel 18.4 Channel 19.4 170
-Channel 20.4 165-Channel 21.4 160 Channel 22.4 1SChannel 23.4 1Channel 24.4 515 505 520 515 520 525 530 535 540 545 550 555 560 565 570 Channel 25.4 Time (seconds)
Figure 6 - Temperature of the 25 Individual Coolant Channels during the First 60 Seconds of a LOFA (Using Table 2 Conservative Assumptions and Benchmark Valve Action and Actuation Times)
In comparing these two sets of graphs, there are more differences than in the previous comparison.
In Figure 5, with a core inlet flow rate of 3,260 gpm (12,340 lpm), the peak centerline temperature of 284 'F (140 *C) occurs in fuel plate number-2, around 8 seconds after the LOFA starts. Figure 6 shows a flow reversal in coolant channels 2 through 4 as channel 3 coolant temperature peaks at 220 'F (104.4 *C) in 7.5 seconds. In Figure 3, with a core inlet flow rate of 3,800 gpm (14,385 lpm), the first peak centerline temperature of approximately 267 'F (131 *C) occurs in fuel plate number-2, around 8.5 to 9 seconds after the LOFA starts. Figure 4 shows an earlier flow reversal in channels 2 and 3 that occurs just before primary coolant isolation valves V507A and V507B close.
In Figure 3, the highest fuel plate centerline temperature of 278 'F (136.7 *C) occurs around 17 seconds after the LOFA starts. Figure 5 also shows a second later peak, but it is lower than the first one with fuel plate number-22 centerline temperature of 270 'F (132.2 *C), about 18 seconds after the start of the LOFA.
In Figure 4, the peak coolant temperature of 237 'F (113.9 *C) occurs in channel 19, around 22 seconds after the LOFA starts. The coolant temperatures peak as flow reversals occur starting with channels 1 through 6 just as primary coolant isolation valves V507A and V507B close. The highest temperatures occur in channels 18 through 24 as flow reversals occur 12 to 22 seconds after the LOFA starts. In Figure 6, the highest coolant temperature of 232 'F (111.1 *C) occurs in channel 22, around 18 seconds after the LOFA starts. With a lower core inlet flow rate of 3,260 27 of 28
gpm (12,340 1pm), the first set of flow reversals that occur in the lower numbered channels happen slightly earlier than in Figure 4 and result in a peak temperature about 10 'F (5.6 *C) higher. The later peak coolant channel temperatures, also due to the flow reversals, occur at lower temperatures over a longer time interval. They begin about 15 seconds after the LOFA initiates and they continue for another 25 seconds.
The effect of fuel burnup and oxide layer buildup on the cladding on the above results can be explained by considering the response to RAI 4.15. The analysis in support of RAI 4.15 assumed worst-case values for reactor power level, reactor inlet water temperature, primary coolant flow, and pressurizer pressure and level. For high burnup, a worst-case narrowest allowed coolant channel gap of 62 mils is assumed. Based on oxide thickness measurements taken on nineteen MURR fuel elements in 1987, it was determined that 1.27 mils would represent the "worst-case" oxide thickness, which would occur on fuel plate number-1.
Therefore, an oxide thickness of 1.27 mils is assumed on the high burnup plate. These assumptions result in higher primary coolant, fuel plate cladding and fuel plate centerline temperatures for normal operating conditions. However, as shown in Figures 1 and 2 and Table 3 of the response to RAI 4.15, the highest fuel plate temperature occurs in the hot spot of the low burnup fuel element. Burnup reduces the heat flux generated in the plate enough to reduce the differential temperature across the fuel meat, cladding and coolant film on the plate surface more than the differential temperature across the oxide layer produced as a result of the long operating history of the fuel element.
28 of 28 ADDENDUM NO.. 1 HAZARDS
SUMMARY
REPORT University of Missouri Research Reactor Facility Compiled and Edited by The Staff Research Reactor Facility Submitted by The University of Missouri Columbia, Missouri February, 1966
TABLE OF CONTENTS Section 1.0 2.0 3.0 Appendix Appendix Appendix Page Introduction Revisions to Original Application Answers to Questions I
II III -i-
TABLE OF CONTENTS Section Topic Page 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.23 Control rods and drive mechanisms...............
Fuel tolerances and acceptance checks.
Safety and process control systems and testing requirements
.4 Evaluation of safety system malfunctions Safety system environment effects.
Redundancy in building isolation system detectors Justification of 125% power level scram.
Emergency power system and tests...................
Peak-to-Average ratio analysis Description of doors 504 and 505........
Builing leak rate limits and techniques.................
Reactivity addition from melted fuel plates.
Failure of valve 546 to open Rod withdrawal accident.
Step reactivity insertion analysis Rupture of pool system Description of experiments Evacuation drills Radiation hazard from p-tubes.
Evaluation of containment isolation system....
Isolation trip adjustments Containment ventilation system Emergency shutdown system Appendix I: Preoperational checkout procedure for the process instrumentation and interlock Appendix II: Heat flux correlations applicable to section 3.7 Appendix III: Step reactivity insertion analysis: digital and analog simulation details 7
9 13 15 17 18 23 26 28 30 39 40 43 53 80 83 90 93 97 100 101 105
1.0 Introduction The purpose of this document is to make certain re-visions to and answer questions resulting from the University of Missouri's application for a Class 104 utilization facility license, submitted to the Commission in July of 1965.
The revisions are concerned mainly with the total inventory of Special Nuclear Material which the University may possess and use at any one time.
Specifically, the University wishes to revise the total quantity of Uranium-235 and Plutonium-239 to be included in the facility license.
This document also presents answers to and discussion of questions presented to us as a request for additional information necessary for the safety evaluation of the proposed operation of our research reactor.
These questions were sent to us with a letter from Dr.
R. L.
Doan dated January 5, 1966.
2.0 Revisions to Original Application In our original application, dated July 1, 1965 and directed to Dr. Richard Doan we stated that our initial fuel inventory would consist of eight fuel assemblies each containing grams of U-235 and four fuel assemblies each containing grams of U-235.
We wish to amend this quantity to include one addi-tional gram assembly.
This will bring our initial U-235 inventory, including neutron detectors, to approximately grams.
We would also like to add to our request, authoriza-tion to receive, possess and use up to 80 grams of plutonium 239 as sealed plutonium-beryllium neutron sources.
When we purchase our second core we will buy 9 rather than 8 assemblies each containing of U-235.
Eight fuel assemblies will constitute a core, the ninth assembly will be held as a spare.
Table 2.1 lists our anticipated maximum inventory of Special Nuclear Material.
Table 2.1 Maximum Special Nuclear Material Inventory I.
Uranium-235 a)
Initial inventory:
8 assemblies 1 spare assembly 4 assemblies @
Total b)
Burnup (assume 5%)
c)
Inventory at end of core life d)
New core:
8 assemblies 1 spare assembly Total e)
Neutron detector gms (260) gms gms gms gms 2 gms f)
Total maximum inventory (a)+(d)+(e)-
estimated burnup gms II.
Plutonium-239 Sealed Pu-Be sources 80 gms It should be emphasized that the inventory shown in Table 2.1 represents a maximum.
Since a core is exactly 8 fuel assemblies, four of the assemblies listed in the initial inventory will probably not be used and will eventually be returned for repro-cessing.
In this case our total inventory at the time of receipt of the second core would be reduced by an amount equal to the assemblies already returned.
Also, Table 2.1 assumes that the second core will be sized for 10 MW operation.
If the modification to the reactor plant required for operation at 10 MW has not yet been completed at the time the second core is procured this new core will also be approxi-mately 5,200 grams.
Reference is made on page 13-24 of the Hazards Summary Report to a 150 psi rupture diaphragm.
Recent design changes are incorporating pressure relief valves instead of the rupture diaphragm.
Paragraph 1.2.6 on page 1-5 is to be revised to read, the University of Missouri Research Reactor Facility will be staffed and operated by the University.
3.0 Answers to Questions In the following paragraphs, each of the 23 questions presented to us will be answered in order.
The answer to each question is complete with supporting discussion and illustrations.
- However, some detailed analysis has been relegated to appendices to avoid making the discussion unnecessarily cumbersome.
3.1 "Provide the assembly drawings of the control rods and control rod drive mechanisms and design drawings showing overall blade dimensions and clearances in the guide structure.
Describe the preoperational tests and periodic tests to be performed on the control rod drive system.
Discuss the possibility of control rod binding due to thermal distortion of the control blades."
Ten copies each of the General Electric drawings 104B2530 -
Regulating Blade 107C4660.-
Control Blade 789D880 Blade Offset Mechanism (Control) 789D884 Blade Offset Mechanism (Reg.)
are included with this submittal.
The control blades operate in a gap between the out-side of the reactor pressure vessel and the inside of the beryllium reflector.
The O.D. of the pressure vessel is between 12.546 and 12.566 inches.
The ID.
of the beryllium reflector is between 13.680 and 13.699 inches.
The gap width is maintained by vertical spacers which are set into the beryllium reflector and cross the gap to the O.D. of the reactor pressure vessel.
The minimum width of the gap in which the rods operate is (13.680 -
12.566)/2 =.557 inches.
The thickness of the control blades is 0.25 inches, leaving a gap of approximately 0.15 inches on each side of the blade.
In order for blade binding due to thermal distortion to occur, some point on a blade must be displaced 0.15 inches with respect to the constrained (upper) end.
The most probable way that this amount of displacement could occur would be as a result of differential expansion between the inner and outer longitudinal fibers accompanying a radial tempera-ture gradient through the blade.
Assuming that the inner radius of a blade is at a higher temperature than the outer radius, a calcu-lation was made to determine the differential ex-pansion needed to displace the unconstrained (bottom)
-end of a control blade 0.15 inches outward, the re-sult was 2.89 x 10-3 inches.
The coefficient of linear expansion of aluminum is 13.0 x 10-6 inches per inch degree fahrenheit.
The unconstrained length of a control blade is approxi-mately 26 inches.
The differential expansion in inches per inch is:
2.89 x 10-3/26 = 1.11 x 10-4 The corresponding temperature change is:
1.11 x 10-4/1.3 x 10-5 = 8.6 degrees fahrenheit.
The above result implies that the control blade must support a temperature gradient of 8.6 degrees.
This temperature difference would require a heat flow radially through the blade of almost 860 Btu/ft 2/hr.,
which is extremely high in view of the fact that the major heat source in the blades is from the absorption of gammia rays and neutrons.
Additionally, temperature gradients will be sup-pressed by cooling water which enters the rod gap at 100°F and passes on both sides of the blades.
It appears extremely unlikely that control rod binding due to thermal distortion of the control blades will be experienced.
The control rod drive system will receive a thorough inspection in addition to the rod drop checks following approximately every 4000 hours0.0463 days <br />1.111 hours <br />0.00661 weeks <br />0.00152 months <br /> of operation above 100 KW but not less than once per year.
During the first years operation this inspection will be performed twice at approximately six months intervals.
The inspection will include a determina-tion of any tendancy that each blade is not properly aligned in its respective slot and that adequate clearance exists between the control blade and either the Beryllium reflector or the reactor pressure vessel.
"Rod-drop times" will be measured at approximately 90 day intervals.
If these.tests show any evidence of rod binding or sticking immediate corrective action will be taken.
Prior to initial startup, the electrical systems of the control rod drive system will be thoroughly tested as part of the safety system checkout dis-cussed in Section 3.3.
In addition to these tests, the control rods will be scrammed 25 times.
Rod drop time will be measured for the first and last five scrams on each rod to determine normal statistical variations.
!ceredyne BORAL Composite STANDARD SPECIFICATIONS Well suited to many neutron absorption applications, the BORAL Composite is used for spent fuel storage pools and dry storage dual-purpose canisters and casks in the United States, Europe and Asia. Since the late 1950's, BORAL has been used in research reactors, nuclear power plants and spent fuel storage systems worldwide. It has the longest continuous service history of any neutron absorbing material, performing its intended function for very long periods in high gamma and neutron radiation fields, never failing to meet it's neutron absorption function, BORALComposite is produced in our Ceradyne Canada ISO 9001:2000 certified and NQA-1 compliant facility in a wide range of surface dimensions, areal densities and thicknesses. It is manufactured in flat sheets that can be cut, punched, bored and formed into shapes. Physical properties allow it to be designed into fabricated structures that form a solid and effective barrier against thermal neutrons.
- 1. BORAL Dimensions of Sizes Available Dimensions Minimum Maximum Note Thickness Range 0.075 1.905 mm 0.270 in 6.858 mm in For any dimension over maximum, contact Width Range 36 in 914 mm Ceradyne Canada Length 192 in 4876 mm
- 2. BORAL Typical Dimension Tolerances Per Thickness Range Thickness Width Length (in)
(mm)
(in)
(mm)
(in)
(mm) 1.905 - 2.286 mm 0.075 - 0.090 in. +/- 0.005 +/- 0.127+- 0.079 +/- 2
+-
+/- 3 0.118 2.287 - 3.810 mm 0.091 - 0.150 in. +/- 0.006 +/- 0.152+1- 0.07
+/- 2 0
+/- 3 1
0.118
+/-
3.811-6.858 mm 0.151 -0.270in.
+/-0.008+/-0.20 /-0.118 +/-3
+/-4 0.157
+/*-
Over 6.858 or 0.270 in.
+/- 0.008 +/- 0.204+A- 0.118 +/- 3 0.157
+/-4
0.1%
0.1%
0.1%
0.1%
These elements are bound beneath the protective content aluminum cladding and therefore not subject to the ambient environment. Other elements are standard constituents of 1100 series aluminum alloy.
Ceradyne Canada, ULC 2702 Boulevard Talbot Chicoutimi, Quebec G7H 5B1 Canada Tel (418) 693-0227
- Fax (418) 693-0393 info@ceradyne.com
- www.ceradyne.com
- oeradyne
- 4. BORAL Typical Engineering Properties Property Measurement Units Value Structural Modulus of Elasticity, E, ASTM E-8 Msi 9 Msi Tensile Strength, Sy, ASTM E-8, E-21 Ksi 10 Ksi Material Ductility - Elongation in 2" Coupon ASTM 0.1%
E-8 Structural In tests performed at the University of Michigan Properties Phoenix Memorial Lab; there were no significant Following Irradiation structural properties changes observed in BORAL after exposure to:
Rad 9.0Ox 10"1 rad Total dose due to gamma alone Total equivalent dose due to combined gamma and neutron exposure Rad 3.4 x 1016 rad Thermal neutron dose 2.7 x 10'9 n/cm2 (W < 0.55 eV):
n/cm 2 Fast (W > 1.0 MeV):
n/cm_
1.08 x 102u n/cmz Thermal Maximum Short-Term Temperature, Wet
°C / °F
<1000C /< 212°F Maximum Short-Term Temperature, Dry 0C /' F
<538°C/< 1000°F Maximum Long-Term Temperature, Wet
°C / 'F
<100°C/< 212°F Maximum Long-Term Temperature, Dry 0C / °F
<454°C /< 850°F Average Areal Dt = overall areal mass density =
9/cm2 Mass Density Da average areal density for aluminum g/cm' 2.713 (g/cm4) * (tclad (cm))
(tcad = clad thickness)
Dc = average areal density for core matrix g/cm' 2.481 (g/cm") * (tcore (cm))
(tcore = core thickness)
Dt (g/cm') = 2
- Da + Dc g/cmz Specific Heat Cp = overall specific heat W-s/g-K 380C 260°C Cpa = specific heat of the aluminum W-s/g-K 0.919 W-s/g-K 1.12 W-s/g-K 100°F 500°F Cpc - specific heat of the core W-s/g-K 0.936 W-s/g-K 1.38 W-s/g-K Cp = (Cpa
- 2
- Da +Cpc
- Dc) / Dt Thermal kt = overall thermal conductivity W/cm-K Conductivity ka = thermal conductivity of aluminum W/cm-K 1.621 W/cm-K 1.864 W/cm-K 380C 2600C kc = thermal conductivity of core matrix W/cm-K 0.859 W/cm-K 0.768 W/cm-K 100°F 500°F xt = overall thickness xa = thickness of aluminum cladding on each side xc = thickness of the core matrix kt = xt / ( (2
- xa / ka) + (xc / kc))
Thermal Emissivity Ref: Sparrow, E. M., & Cess, R.D. (1966).
e 0.10- 0.19 Radiation Heat Transfer (p. 44). From Brooks/Cole Publishing Company, the extrapolation between rough plate and oxidized.
Mechanical Coefficient of Thermal Expansion = a in./in.-C 1.97
- 10-5 BORAL - Typical 1 OB loading is limited to equivalent of 65% natural g/cmz 0.005 - 0.120 g/cm2 IOB Content B4C in the core volume.
However, a wide range of thicknesses from.070 inches through.400 inches have been produced with lOB areal densities ranging from.005 through
.120 gm/cm2. Please contact a Ceradyne Canada representative for a detailed explanation and specific limits.
Notes: Properties listed are typical for BORAL) for information only. Actual property values will vary based on the BORAL design necessary to achieve specific attributes. If a Customer required specific property values, conformance tests should be specified in the contract and performed by the Producer for that property.
SP-BORA-00len, Rev. 10-17-08 Ceradyne Canada, ULC 2702 Boulevard Talbot Chicoutimi, Quebec G7H 5B1 Canada Tel (418) 693-0227
- Fax (418) 693-0393 info@ceradyne.com - www.ceradyne.com BORAL Control Blade Thermal-Mechanical Analysis By Srisharan G.Govindarajan James A.Moreland Gary L.Solbrekken
Abstract A coupled thermal-mechanical analysis was completed on the BORAL-based control blade that used by the University of Missouri Research Reactor (MURR). Internal heat generation caused by radiation absorption is applied as a thermal load in a finite element model constructed in Abaqus. A convective heat transfer coefficient is applied to the outer surfaces of the blade to simulate cooling by the water coolant. The solution to the thermal model provides the temperature distribution within the composite blade. The resulting temperature profile is applied as a mechanical load on the control blade. A rigid boundary condition is applied to the mounting plate region of the control blade while the temperature profile is applied as a mechanical load.
The Abaqus finite element solver then provides the estimated blade deflection due to radiative heating. The results suggest that the amount of deflection is smaller than the size of the blade channel.
Introduction The BORAL-based control blades used by the University of Missouri Research Reactor (MURR) will experience a thermally induced deflection during reactor operation. The deflection will arise due to the composite structure of the control blade. The absorbent is BORAL, which is a boron based material. The BORAL is clad with aluminum. The interfaces between the BORAL and aluminum are pressed to establish an adherent bond. As the BORAL absorbs both alpha particles and gamma rays, there will be volumetric heat generation and a corresponding rise in temperature. Since the BORAL and aluminum materials have different coefficients of thermal expansion, there will be a tendency of the blade to deform as the blade temperature changes and the materials expand at different rates.
In addition to the composite nature of the control blade, there will be spatial variations in temperature within the control blade caused by non-uniform heat generation within the BORAL meat. The non-uniform heating develops primarily as a result of alpha particle flux being non-uniform along the longitudinal direction when then control blade is partially withdrawn. There is also variation in the heating profile through the thickness and about the circumferential width of the control blade.
A mathematical curve-fit is generated for the non-uniform volumetric heat generation profile caused by alpha particle absorption. Similarly, a curve-fit is also generated for the gamma heat generation rate. The functions are applied as heating conditions within a finite element model of the control blade. A convective heat transfer coefficient is applied to the outer boundaries of the
control blade and the neutral assembly temperature is assumed to be room temperature. The finite element model is solved for the temperature distribution. The temperature distribution is then used as an applied load in a mechanical deflection model. The resulting deflection is compared with the channel gap to determine if there is a significant risk of the plate binding during a safety shut down event.
Control Blade Geometry The commercial finite element code Abaqus FEA (version 6.10-2) was used to perform a fully coupled thermal stress analysis on the BORAL control blades to determine the magnitude of thermally induced deflection to help MURR with their relicensing. The model consists of the BORAL meat sandwiched between the aluminum cladding to form a composite structure. The BORAL core geometry was taken to be 34" long, 0.1" thick and 8.06" wide. The radius of curvature of the centerline of the BORAL was 6.555". The aluminum cladding thickness was taken to be 0.0375" on the front and back faces of the BORAL with 0.125" on both edges. The mounting region of the blade extends to 5.5" beyond the BORAL cavity. Thus the overall profile of the control blade including the mounting region was 8.31" wide x 0.175" thick x 39.75" long.
Developing a Heat Generation Rate Profile BORAL meat split into 4 quadrants.Heat Generation Rate as a Function of Thickness and Longitudinal Position.
The volumetric heat generation function used to simulate radiative decay in the BORAL was curve fit from discrete points obtained from MURR. The data at the longitudinal locations 10.2, 17, 23.8, and 30.6 inches from the fuel-side aluminum cladding and beryllium-side aluminum cladding were used for the heat generation values within the BORAL thickness. Fuel-side data points were considered constant through the first four BORAL thickness locations. Beryllium-side data points were considered constant through the last four BORAL thickness locations. The data point at the center of the BORAL was averaged from the fuel-side and beryllium-side points.
For a given azimuthal position, the volumetric heat generation data was plotted against thickness and longitudinal position using 3-D plotting software to visualize the general functional trends.
Additionally, for a given thickness, the data was plotted against azimuthal and longitudinal position to further aid in visualization. These plots suggested that the volumetric heat generation follows an even polynomial profile in the radial direction, t, and a decaying exponential in the longitudinal direction, z. Thus, a function of the form
q". = f(t, z) = A(t)eB(t)z (1) where the functions A(t) and B(t) allow for a different decaying function to be determined at a given radial position. A function of this form was determined for each of the four partitions the blade was divided into, the left edge, left middle, right middle, and right edge quadrants. The equation governing the left edge section was held constant along the azimuthal direction between
-36.3' and -18.75'. The equation governing the left middle section was held constant along the azimuthal direction for -18.15' to 00. Likewise, the right middle section applied to the 0' to 18.15' portion of the blade and the right edge equation applied to the 18.15' to 32.30 partition.
In order to derive the functions A(t) and B(t), the natural log of Eq. (1) was taken to obtain Inq'" = lnA(t) + B(t)z = B(t)z + C (2)
From here it is simple to fit a straight line relating the In q.' to z at each thickness location. An example of this process can be seen in Fig.1 where In q.' is plotted against z at the BORAL fuel-side surface (t = 0.0375 mils) at the left edge location for alpha heat generation. The slope and intercept are provided on the plot, with the slope being the value of B and intercept the value of C.
ln(q"') vs. Longitudinal Position At Left Edge, Fuel-Side Surface Alpha Heating 6
4 2
0 Y -0.39x + S.1461 R' = 0,99023 0
5 10 20 25 30 3S
-4
-6 4
Langbd&la Pouldon z (in)
Figure 1. Linearized exponential decay function for a given thickness and azimuthal location for alpha heating.
Equations of the form of Eq. (2) were determined for each of the provided thickness and azimuthal locations. These equations were then used as the exponent of the exponential function to return it to the form of Eq. (1). Similar results were obtained for the gamma heating profile.
The A and B coefficient variation with thickness was then determined. Due to the discontinuities occurring at the interfaces of the aluminum cladding and BORAL core, the overall function for volumetric heat generation was defined in a piecewise fashion where the A and B coefficients remained constant with thickness through the cladding on the fuel-side and beryllium-side, while the coefficients varied with thickness through the BORAL core. The A coefficient variation followed a 4th order polynomial for the alpha heating profile and a 2 nd order polynomial for the gamma heating profile. The B coefficient followed a 2 nd order polynomial curve for both the alpha and gamma heating profiles. Examples of the A and B coefficient variation with thickness through the BORAL core can be seen in Figs. (2-5).
A Coefficient vs. Thickness in BORAL Core Alpha Heating 0.0375 in ; t S 0.1375 in 350 300 250 100 50 y = 3E+07x4 - 1E-+07x3 + 1E+06x' - 76060x + 1603.4 fe = 0.97606 0 o 0.02 0.04 0.06 0.03
- Thkknem, t (In) 0.1 0.12 0.14 0 16 Figure 2. A coefficient variation with thickness in BORAL core for alpha heating at left edge location.
B Coeffident vs. Thickness In BORAL Core Alpha Heating 0.0375 in < t S 0.1375 in 0
0.02 0004 0.06 0.06 0.1 0.12 0.14 0.16
-01
-0.2
-0.4
-0.S
-0.6 y = -31.659x2 + 4.5776x - 0.5079 R2 = 0.94256 Thickness, t (in)
Figure 3. B coefficient variation with thickness in BORAL core for alpha heating at left edge location.
A Coefficlent vs. Thickness In BORAL Core Gamma Heating 0.0375 In S t $ 0.1375 In 20.6 20.S 20.4 20.3I01 yz27.37Sx2 - 11.38x + 20.952 R2= 0.97722 a
20 19.9 19.8 0.02 0.04 0
0.06 0.06 Thic s t (in)
- 0.
0.12 0.14 0.16 Figure 4. A coefficient variation with thickness in BORAL core for gamma heating at left edge location.
B Coefficient vs. Thickness in BORAL Core Gamma Heating 0.0375 In S t ; 0.1375 In 4).2035
-0204 41L045
-0.20SS
-0.206
-O.2065 0.02 0.04 0.06 0.0 0.1 0.12 014 0.16 y 40.o56W + o.033Sx - o.207S R2 t.96642 Ilkes P'
Figure 5. B coefficient variation with thickness in BORAL core for gamma heating at left edge location.
The resulting functional form for the heat generation rate within the BORAL resembled Eq. (1) where Aa(t) = C4t 4 + cst 3 + c2t2 + clt + CO and for the alpha heating profile and Ba(t) = d2 t2 + dit + do Ay(t) = C2t 2 + clt + CO By(t) = d2 t 2 + dit + do (3)
(4)
(5)
(6) and for the gamma heating profile. These c and d coefficients varied along the width of the plate and were defined for each of the four quadrants. The resulting functional forms of the alpha heating and gamma heating profiles are summarized in Tables 1 and 2 along with the values of the various coefficients.
Table 1. Functional form of the alpha heat generation in BORAL control blade.
Volumetric Heat Generation Functional Form - Alpha Heating For 0 < t < 0.038:
q"' = f(w, z) = AeBz (Fuel-Side Aluminum)
Quadrant A
B Left Edge 45.78
-0.33 Left Mid 46.91
-0.35 Right Mid 63.86
-0.38 Right Edge 71.89
-0.35 For 0.038 <t<0.138:
q.' = f(t,z) = A(t)eB(t)z (Boral)
A(t) = c4t 4 + c 3t 3 + C2 t 2 + c1t + Co Quadrant C4 C3 C2 Cl CO Left Edge 3.32E+07
-1.11E+07 1.38E+06
-7.61E+04 1.60E+03 Left Mid 3.03E+07
-1.02E+07 1.27E+06
-7.04E+04 1.50E+03 Right Mid 3.54E+07
-1.18E+07 1.46E+06
-8.02E+04 1.69E+03 Right Edge 4.21E+07
-1.43E+07 1.80E+06
-1.OOE+05 2.14E+03 B(t) = d 2 t 2 + d 1t + do Quadrant d2 di do Left Edge
-31.65 4.57
-0.50 Left Mid
-29.46 4.61
-0.53 Right Mid
-30.41 4.58
-0.54 Right Edge
-29.52 4.88
-0.55 For 0 < t < 0.038: q.' = f(z) = AeBz (Beryllium-side Aluminum)
Quadrant A
B Left Edge 105.24
-0.42
Left Mid 69.88
-0.39 Right Mid 121.31
-0.44 Right Edge 82.51
-0.37 Table 2. Functional form for the gamma heat generation in BORAL control blade.
Volumetric Heat Generation Functional Form - Gamma Heating For 0 < t < 0.038:
q.' = f(z) = AeBz (Fuel-side Aluminum)
Quadrant A
B Left Edge 8.91
-0.17 Left Mid 8.84
-0.17 Right Mid 8.76
-0.17 Right Edge 11.22
-0.17 For 0.038 <t < 0.138:
q.' = f(t,z) = A(t)eB(t)z (Boral)
A(t) = c2t2 + clt + Co Quadrant c2 cl co Left Edge 27.37
-11.38 20.95 Left Mid 36.92
-12.52 20.62 Right Mid 27.46
-7.90 20.32 Right Edge 63.84
-18.86 25.79 B(t) = d2 t 2 + d1t + do Quadrant d2 di do Left Edge
-0.05 0.03
-0.20 Left Mid
-0.08 0.03
-0.20 Right Mid
-0.08 0.07
-0.21
Right Edge
-0.11 0.03
-0.21 For 0 < t < 0.038: q"' = f(z) = AeBz (Beryllium-side Aluminum)
Quadrant A
B Left Edge 8.69
-0.16 Left Mid 8.58
-0.16 Right Mid 8.92
-0.16 Right Edge 10.91
-0.17 Comparisons between the fits and the data provided by MURR can be seen in the following figures and tables. Figures 6 and 7 are the alpha heat generation through the BORAL core at the left edge (w-3.115 inch) of the control blade data and curve fit surface plots, respectively. Here width is held constant as thickness, t, and longitudinal position, z, are varied.
Alpha Heat Generation Data through BORAL core at Left Edge (w=.3.115) 1W
>1 1180-200 160%180
- 140-160
- 120-140 S100M120
'80-100 i 60-80
- 40-60 8 20-40
' 0-20
-n&I Pwmac, I (in) 012 0,
0.137S Tmcbmk t (10)
Figure 6. Alpha heat generation data through BORAL core at left edge.
Alpha Heat Generation Fit through BORAL core at Left Edge (w=-3.115) 35o 300 250 2W I SO 35 5.
7,5 9.5
-'I IiI 300-350
- 250-300 0 200-250
- 150-200 9 100-150 R 50-100 n 0-50 Longitudlnal Positon., x 11n) mckoess, t (in)
Figure 7. Alpha heat generation fit through BORAL core at left edge.
Figures 8 and 9 are the alpha heat generation at the fuel-side BORAL face (t = 0.0375 in) for the data and curve fit. Thickness, t, has been held constant while width, w, and longitudinal position, z, have been varied to generate the surface plots.
Alpha Heat Generation Data at fuel-side* BORAL face, (t = 0.037S In) 200
- 1000 400-05
- 180-200 160-180 6 140-160 0120-140
- 100-120
'80-100 E 60-80 8 40-60 8 20-40 8 0-20 25S 3.11471M401 101 Axim" 7
Auiruthal Position, w (in) 23.9 3.114712401 10.2 17 30.6 LPsitional
- posltlo, 2(h)
Figure 8. Alpha heat generation data at the fuel-side BORAL face.
Alpha Heat Generation Fit at fuel-side BORAL face (t = 0.0375 in) 20
- ~ ~
1if 0D o
0.5 1.5 255 3 5.
10'2 15 7
20
- 23.
2s 30 30.6 34 0 200-250 0 150-200 100-150
. 50-100
- 0-50 S2"076474934
_B3 Azimuthal 2 7.
Position. W (in)
-4.152949M6 Longitudinal Position, z (in)
Figure 9. Alpha heat generation fit for the fuel-side BORAL face.
Table 3 compares the data obtained from MURR and the values obtained from the functional form of the heat generation for the alpha heating profile at discrete points within the control blade.
Table 3. Alpha volumetric heat generation data and fit comparison at left edge (w = -3.115 in).
Longitudinal Volumetric Heat Volumetric Heat Error Position, z Generation Data Generation Fit (in)
(W/cm3)
(W/cm 3) 0.50 144.427 144.884 0.32 1.50 94.874 99.006 4.36 2.50 76.926 67.656 12.05 3.50 64.632 46.232 28.47 10.20 2.612 3.606 38.07 17 0.125 0.270 116.11 23.80 0.009 0.020 111.25 30.60 0.002 0.001 34.94 t
0.088 in 0.50 28.419 39.216 37.99 1.50 20.679 27.642 33.67 2.50 17.329 19.484 12.44 3.50 18.241 13.734 24.71 10.20 2.586 1.319 49.01 17 0.115 0.122 6.16 23.80 0.007 0.011 45.30
30.6 0.001 0.001 15.93 t = 0.138 in 0.5 192.955 248.881 28.98 1.5 132.400 154.462 16.66 2.5 103.999 95.863 7.82 3.5 87.631 59.495 32.11 10.2 2.561 2.435 4.92 17 0.105 0.095 9.56 23.8 0.005 0.003 37.99 30.6 0.000 0.000 5.76 Model Setup in Abaqus FEA A numerical simulation in Abaqus FEA typically consists of 3 separate stages namely the pre-processing stage, simulation evaluation stage and the post processing stage. The pre-processing stage involves creation of an input file which contains all the constraints under which the model is to be numerically evaluated. The material properties in Table 4 are input into the model during this stage. In the second stage the numerical analysis is evaluated based on the constraints specified in the pre-processing stage or the Abaqus FEA input file. The post processing stage involves visualizing the results and extracting the data.
Table 4. Material Properties of the BORAL and cladding used in the analysis.
Material Property Aluminum 1100 BORAL Thermal Conductivity (W/mK) 186.4 76.8, 98, 132 Density (Kg/m3) 2713 2481 Elastic Modulus (GPa) 69 62.053 Poisson's Ratio 0.33 0.23 Thermal Expansion Coeff (K-')
2.36E-05 1.97E-05 Specific Heat (J/KgK) 1120 1380 The curvature of the control blade was created first using the dimensions in the drawing provided and then extruded to 39.75" to form the full length of the control blade. This was followed by creating a partition at 34.25" to separate the bolted region from the region occupied by the BORAL (Fig. 10). A cavity equivalent in curvature and length of the BORAL was cut-extruded to the length of the BORAL meat. To help simulate heating profile 'B', the BORAL was partitioned into 4 sections (Fig. 11) along the curvature.
Assembly of the model consisted of placing the BORAL meat in the cavity created in the control blade. The 'translation' option in Abaqus was used to perform the assembly by specifying the
start and end points for translation. A fully coupled thermal stress step was created and perfect contact was assumed to exist between the BORAL meat and the aluminum cladding. Figure 10 illustrates the assembled model of the control blade.
Figure 10. Assembled model of the control blade with the BORAL meat.
Thermal Load and Boundary Conditions It was assumed that all the external surfaces of the aluminum cladding were exposed to coolant at 325 K that provided an effective heat transfer coefficient of 1000 W/m2K. Perfect interfacial contact between the BORAL and aluminum cladding was assumed. The internal heat generation profile obtained based on the curve fit data was used as the loading condition on the BORAL.
A.
Figure 11. BORAL meat split into 4 sections to apply the load.
Mechanical Boundary Conditions The mechanical boundary conditions were applied to the bolt holes to prevent translation of these holes in the radial and longitudinal direction as illustrated below in Fig 12.
Figure 12. Mechanical boundary condition applied to the bolt holes of the control blade.
Results Figures blade.
(13-16) illustrate the thermal and corresponding mechanical behavior of the control
+2.4114ft+0
+3.23.D++Z It Figure 13. Temperature distribution contour of the control blade.
Figure 13 shows that the right side of the blade is hotter than the other side and this is consistent with the curve fits (Figs 6-9) that were put together based on the data provided by MURR. The variation of the temperature along the cladding centerline is presented in Fig 14. The temperature also drops off in the longitudinal direction where the heat generation rate also drops off.
Temperature distribution along the cladding centerline ig 344 No 338 334 332 330 328 326 324
-ta s@
00-oft" Ros 0
10 20 30 Length along the blade (inch) 40 s0 Figure 14. Temperature distribution along the cladding centerline.
The radial deflection of the blade resulting from the temperature is illustrated in Fig 15. A first look at this figure gives the impression that the deflection is symmetrical when Fig 13 shows that asymmetry in temperature exists along the curvature of the blade. To verify that asymmetry exists in the radial deflection along the curvature, the deflection as a function of the circumference was plotted as illustrated in Fig 16. This plot shows that the right edge of the blade (which is the hotter side) deflects slightly more than the other side and the difference in deflection is 1.2 x 10-3 inches.
U. all (-ASM4KY-T-AMlH CIVS-I)
-bOM
.LW06$42
-042 4.9943604 V
Figure 15. Radial deflection contour of the control blade.
Temperature Defiection EaU Magnitude of deflection on the sides of the blade 0.240 0.235
++
0.230 B
Blade Cante
+.-i X
f Asunwd po~en romtao hi tde dockuwheand cmdadockwhe dhction for X astorermIn ti amime.
This mkes It easlecto compare the dflections.
. Left side C
-. 0.225
.2 0.220 0.215
_ 0.210 +
0*0205 i
t t ýT
-.+/-
+"
+ Right side 0.200 -
Blade Edge 0.195 0
5 10 15 20 25 30 35 40 Angle, theta (IDegee)
Figure 16. Deflection variation on the sides of the control blade.
Radial displacementalong the cladding centerline (Profile-B)
E CL 0!
0.27 0.24 0.21 0.18 0.15 0.12 0.09 0.06 0.03 0
N.
oral Region Boted Region k 76.8 WImK
-k98 WImK
...... k 132 WImK Control blade channel gap 0
10 20 30 40 5o Length along the blade (inch)
Figure 17. Deflection variation along the length of the blade with varying BORAL thermal conductivity.
BORAL Thermal Conductivity Study The thermal conductivity of the BORAL used in the current analysis is 76.8 W/mK as illustrated in Table 4. However, experiments at the University of Missouri suggested that the thermal conductivity of the BORAL was 115++/-17.W/mK. A parametric study was performed by varying the thermal conductivity of the BORAL with a lower bound of 98 W/mK and an upper bound of 132 W/mK. The change in maximum radial deflection was found to be negligibly small and of the order of 4x 10-4 inch. Figure 17 illustrates that the maximum radial deflection of the blade is within the control blade channel gap even at the upper bound (132 W/mK) of the BORAL thermal conductivity. It should be noted that the three different thermal conductivity values modeled provide deflections that lie nearly on top of each other making them difficult to distinguish in the plot.
Conclusion The commercial finite element code Abaqus FEA was used to perform a thermal mechanical analysis on a BORAL control blade to help MURR with their relicensing. The task was to establish that the deflection of the blade due to heating will be within the control blade channel gap of 0.26 inch (6.5 mm).
The BORAL meat was split into 4 quadrants and a heat generation profile that varied with thickness and longitudinal position was applied to each quadrant. The magnitude of the resulting thermally induced deflection was within the channel gap limit of 0.26 inch (6.5 mm). The heating profile showed some asymmetry in temperature and the corresponding asymmetric deflection along the blade curvature has been discussed. Finally, thermal conductivity parametric studies on the BORAL suggested that the maximum radial deflection of the blade is still within the control blade channel gap limit even at a BORAL thermal conductivity of 132 W/mK, which was considered to be the upper bound in the thermal conductivity study.
TD-0 33 k~,Ar~
TcAiJ :1D Technical Data Report TDR No.: 0133 Revision No.: 0
Title:
MURR Boral Control Blade Thermal-Mechanical Analysis Originator(s) Signature Date Approval(s) Signature Date External Distribution Approval Date N/A N/A Distribution L. Foyto C. McKibben J. Fruits C. Herbold K. Kutikkad N. Peters R. Butler Abstract:
A coupled thermal-mechanical analysis was completed on the BORAL control blades that are used by the University of Missouri Research Reactor (MURR). Internal heat generation caused by radiation absorption is applied as a thermal load in a finite element model constructed in Abaqus.
A convective heat transfer coefficient is applied to the outer surfaces of the blade to simulate cooling by the pool water. The solution to the thermal model provides the temperature distribution within the composite blade.
The resulting temperature profile is applied as a mechanical load on the control blade.
A rigid boundary condition is applied to the mounting plate region of the control blade while the temperature profile is applied as a mechanical load. The Abaqus finite element solver then provides the estimated blade deflection due to radiative heating. The results indicate that the amount of deflection is smaller than the size of the control blade channel gap.
- - Abstract Only
- i -
TcJAic I t Report Technical Data Report TDR No.: 0133 Revision No.: 0 Rev.
Summary of Change Approval Date 0
Original Issue N/A N/A List of Effective Pages Page Rev.
Page Rev.
Page Rev.
Page Rev.
Page Rev.
Page Rev.
Page Rev.
1 0
14 0
2 0
15 0
3 0
16 0
4 0
17 0
5 0
18 0
6 0
19 0
7 0
20 0
8 0
21 0
9 0
10 0
11 0
12 0
13 0
-ii-
TDR-0133 Revision 0 INTRODUCTION The BORAL control blades used by the University of Missouri Research Reactor (MURRO) will experience a thermally induced deflection during reactor operation. The deflection will arise due to the composite structure of the control blade. The absorbent is BORAL, which is a boron-based material. The BORAL is then clad with aluminum. The interfaces between the BORAL and aluminum are pressed to establish an adherent bond. As the BORAL absorbs both thermal neutrons and gamma rays, there will be volumetric heat generation and a corresponding rise in temperature. Since the BORAL and aluminum materials have different coefficients of thermal expansion, there will be a tendency of the control blade to deform as the blade temperature changes and the materials expand at different rates.
In addition to the composite nature of the control blade, there will be spatial variations in temperature within the control blade caused by non-uniform heat generation within the BORAL meat. The high boron-10 cross-section of the B-10 (n, a) Li-7 thermal neutron reaction produces the vast majority of the heating in the control blade. This reaction primarily occurs within the first 20 mils of the BORAL meat surface and produces about 2.79 MeV of energy, of which 0.84 Mev is the reaction energy of the Li-7 and 1.47 Mev is the alpha particle. The remaining 0.48 MeV is a gamma ray. Therefore, about 80% of the nuclear reaction's energy is deposited within a few mils of the reaction location. Consequently, the major heating is in the outer two surfaces of the BORAL meat, thus making the heat generation through the blade low except for the outer 20 mils of each surface. These combine to produce a variation in the heating profile through the thickness and about the circumferential width of the control blade. The heat generation is also non-uniform along the longitudinal direction because the thermal neutron flux drops off significantly from the leading edge (bottom) of the control blade to the top. With typical mixed bumup cores at cold clean startup each week, only about nine inches of the control blade is inserted past the top section of the fuel meat in the fuel elements.
A mathematical curve-fit is generated for the non-uniform volumetric heat generation profile caused by the B-10 (n, a) Li-7 reaction. Similarly, a curve-fit is also generated for the gamma absorption heating rate. The functions are applied as heating conditions within a finite element model of the control blade.
A convective heat transfer coefficient is applied to the outer boundaries of the control blade and the neutral assembly temperature is assumed to be room temperature.
The finite element model is solved for the temperature distribution.
The temperature distribution is then used as an applied load in a mechanical deflection model. The resulting deflection is compared with the control blade channel gap to determine if there is a significant risk of the blade binding during operation.
CONTROL BLADE GEOMETRY The commercial finite element code Abaqus FEA (version 6.10-2) was used to perform a fully coupled thermal stress analysis on the MURR BORAL control blades to determine the magnitude of thermally induced deflection [1].
The model consists of the BORAL meat sandwiched between aluminum cladding to form a composite structure.
The BORAL core geometry is 34 inches long, 0.100 inches thick and 8.06 inches wide. The radius of curvature of the centerline of the BORAL is 6.555 inches. The aluminum cladding thickness is 0.0375 inches 4
TDR-0133 Revision 0 on the front and back faces of the BORAL with 0.125 inches on both edges. The mounting region of the blade extends to 5.5 inches beyond the BORAL cavity. Thus, the overall profile of the control blade including the mounting region is 8.3 1 inches wide x 0.175 inches thick x 39.75 inches long.
DEVELOPING A HEAT GENERATION RATE PROFILE BORAL Meat Split into 4 Quadrants - Heat Generation Rate as a Function of Thickness and Longitudinal Position The volumetric heat generation function used to simulate the heat generation from the nuclear reaction in the BORAL was curve fit from discrete points obtained from MCNP. The data at the longitudinal locations 10.2, 17, 23.8, and 30.6 inches from the fuel-side (concave) aluminum cladding and beryllium-side (convex) aluminum cladding were used for the heat generation values within the BORAL thickness. Fuel-side data points were considered constant through the first four BORAL thickness locations.
Beryllium-side data points were considered constant through the last four BORAL thickness locations. The data point at the center of the BORAL was averaged from the fuel-side and beryllium-side points.
For a given azimuthal position, the volumetric heat generation data was plotted against thickness and longitudinal position using 3D plotting software to visualize the general functional trends.
Additionally, for a given thickness, the data was plotted against azimuthal and longitudinal position to further aid in visualization. These plots suggested that the volumetric heat generation follows an even polynomial profile in the radial direction, t, and a decaying exponential in the longitudinal direction, z. Thus, a function of the form:
q.. = f(t, z) = A(t)eB(t)z (1) where the functions A(t) and B(t) allow for a different decaying function to be determined at a given radial position. A function of this form was determined for each of the four partitions the blade was divided into, the left edge, left middle, right middle, and right edge quadrants. The equation governing the left edge section was held constant along the azimuthal direction between
-36.3' and -18.75'. The equation governing the left middle section was held constant along the azimuthal direction for -18.15' to 00. Likewise, the right middle section applied to the 0' to 18.15' portion of the blade and the right edge equation applied to the 18.15' to 32.30 partition.
In order to derive the functions A(t) and B(t), the natural log of Eq. (1) was taken to obtain:
Inq'" = lnA(t) + B(t)z = B(t)z + C (2)
From here it is simple to fit a straight line relating In q"' to z at each thickness location. An example of this process can be seen in Figure 1 where In q.' is plotted against z at the BORAL fuel-side surface (t = 0.0375 mils) at the left edge location for alpha heat generation. The slope and intercept are provided on the plot, with the slope being the value of B and intercept the value of C.
5
TDR-0133 Revision 0 ln(q'") vs. Longitudinal Position At Left Edge, Fuel-Side Surface Alpha Heating 6r
-3y
=
%39x
+ 5.1461 RI- 0.99023 2 20 2S 30 35
-8 Lnut ad Figure 1 Linearized Exponential Decay Function for a Given Thickness and Azimuthal Location for Alpha Heating Equations of the form of Eq. (2) were determined for each of the provided thickness and azimuthal locations. These equations were then used as the exponent of the exponential function to return it to the form of Eq. (1). Similar results were obtained for the gamma heating profile.
The A and B coefficient variation with thickness was then determined. Due to the discontinuities occurring at the interfaces of the aluminum cladding and BORAL core, the overall function for volumetric heat generation was defined in a piecewise fashion where the A and B coefficients remained constant with thickness through the cladding on the fuel-side and beryllium-side, while the coefficients varied with thickness through the BORAL core. The A coefficient variation followed a 4h order polynomial for the alpha heating profile and a 2 nd order polynomial for the gamma heating profile. The B coefficient followed a 2nd order polynomial curve for both the alpha and gamma heating profiles. Examples of the A and B coefficient variation with thickness through the BORAL core can be seen in Figures 2 through 5.
6
TDR-0133 Revision 0 A Cofflcient vs. Thickness In BORAL Core Npha Heating 0.0375 in !5 t ! 0.1375 in 350 300 25O 150 y a3E407X4 - IE+IO7x3 + IE+06xZ - 76060x + 1603.4 Rz - 0.97606 so h 0
0 0.02 0.04 0.06 0.1 0.12 0.06 Thictnss, t (in) 0.14 0.16 Figure 2 A Coefficient Variation with Thickness in BORAL Core for Alpha Heating at Left Edge Location B Coffilent vs. Thickness in BORAL Core Alpha Heatin 0.037S in S t 50.137S In
-0.1
-CA
-.0S
-0.6 0.08 0A 0.12 0.14 0.02 0.04 0.06 0.16 y = -31.659x2 + 4.S776x -0.3079 R'2 = 0.942S6 Thkknuss, t (in)
Figure 3 B Coefficient Variation with Thickness in BORAL Core for Alpha Heating at Left Edge Location 7
TDR-0133 Revision 0 A Coefficient vs. Thickness In BORAL Core Gamma Heating 0.0375 In !S t & 0.1375 In 20.6 20 y - 27.375Xz - 11.38x + 20.952 IOA R2 0.97722 203 20.1 20,-
20 19.9 0
0.02 0.04 0.06 0,08 0.1 0.12 0.14 0.16 Thlkaiess t (in)
Figure 4 A Coefficient Variation with Thickness in BORAL Core for Gamma Heating at Left Edge Location B Coefficlent vs. Thickness In BORAL Core Gamma Heating 0.0375 In 5 t S 0.1375 In
-02035S 0.02 0.04 006 0-0o 0.1 0.12 014 0.16
-0 4.0%60 + 0.0335x -0.207S R2 0.9642
-0.206 T~dwon t fMn)
Figure 5 B Coefficient Variation with Thickness in BORAL Core for Gamma Heating at Left Edge Location 8
TDR-0133 Revision 0 The resulting functional form for the heat generation rate within the BORAL resembled Eq. (1) where Aa(t) = c4 t 4 + c3 t 3 + c2 t 2 + c1 t + Co (3)
(4) and Ba(t) = d2 t2 + d1t + do for the alpha heating profile and AV(t) = c2 t2 + clt + co (5) and By(t) = d2t 2 + dlt + do (6) for the gamma heating profile. These c and d coefficients varied along the width of the plate and were defined for each of the four quadrants. The resulting functional forms of the alpha heating and gamma heating profiles are summarized in Tables 1 and 2 along with the values of the various coefficients.
Table 1 Functional Form of the Alpha Heat Generation in a BORAL Control Blade Volumetric Heat Generation Functional Form - Alpha Heating For 0 < t < 0.0375:
q.' = f(w, z) = AeBz (Fuel-Side Aluminum)
Quadrant A
B Left Edge 45.78
-0.33 Left Mid 46.91
-0.35 Right Mid 63.86
-0.38 Right Edge 71.89
-0.35 For 0.0375 <t_<0.1375:
q.' =f(t,z) = A(t)eB(t)z (BORAL)
A(t) = c 4 t 4 + c 3 t 3 + c 2 t 2 + c 1 t + cO Quadrant c4 c3 c2 cl Co Left Edge 3.32E+07
-1.11E+07 1.38E+06
-7.61E+04 1.60E+03 Left Mid 3.03E+07
-1.02E+07 1.27E+06
-7.04E+04 1.50E+03 9
TDR-0133 Revision 0 Right Mid 3.54E+07
-1.18E+07 1.46E+06
-8.02E+04 1.69E+03 Right Edge 4.21E+07
-1.43E+07 1.80E+06
-1.00E+05 2.14E+03 B(t) = d2 t 2 + d1t + do Quadrant d2 d,
do Left Edge
-31.65 4.57
-0.50 Left Mid
-29.46 4.61
-0.53 Right Mid
-30.41 4.58
-0.54 Right Edge
-29.52 4.88
-0.55 For 0 <t < 0.0375: q`' = fl(z) = AeBz (Beryllium-side Aluminum)
Quadrant A
B Left Edge 105.24
-0.42 Left Mid 69.88
-0.39 Right Mid 121.31
-0.44 Right Edge 82.51
-0.37 Table 2 Functional Form for the Gamma Heat Generation in a BORAL Control Blade Volumetric Heat Generation Functional Form - Gamma Heating For 0 <t 5 0.0375:
q.' = f(z) = AeBz (Fuel-side Aluminum)
Quadrant A
B Left Edge 8.91
-0.17 Left Mid 8.84
-0.17 Right Mid 8.76
-0.17 Right Edge 11.22
-0.17 For 0.0375 _<t_< 0.1375:
q.' =f(t,z) = A(t)eB(t)z (BORAL)
A(t) = c2t 2 + clt + co 10
TDR-0133 Revision 0 Quadrant C2 C
CO Left Edge 27.37
-11.38 20.95 Left Mid 36.92
-12.52 20.62 Right Mid 27.46
-7.90 20.32 Right Edge 63.84
-18.86 25.79 B(t) = d2 t2 + dlt + do Quadrant d2 di do Left Edge
-0.05 0.03
-0.20 Left Mid
-0.08 0.03
-0.20 Right Mid
-0.08 0.07
-0.21 Right Edge
-0.11 0.03
-0.21 For 0 : t _< 0.0375: q.' = f(z) = AeBz (Beryllium-side Aluminum)
Quadrant A
B Left Edge 8.69
-0.16 Left Mid 8.58
-0.16 Right Mid 8.92
-0.16 Right Edge 10.91
-0.17 Comparisons between the fits and the data provided by MCNP can be seen in the following figures and tables. Figures 6 and 7 are the alpha heat generation through the BORAL core at the left edge (w = -3.115) of the control blade data and curve fit surface plots, respectively. Here width is held constant as thickness, t, and longitudinal position, z, are varied.
11
-TtOR9-133 Revi~loflO B.A. cote A1~
~p'wo two*
£d wo~w~ew
-t 4 v..31S
-1
.180-200
£ *600
- 120-140 S0-20 Ii "iv'ii lin 10 110 ai Wn 007 m05 0 7 0
id 0.1 2 2 5 0.1 2 1 s O A 3? '
- WPAWO, volow, t (in) figure 6 core at Left ~
neainData BoroAg Alphla katGerel'I g
MLce C'.won "I
XVha6 atLeftEdseý 3S 100 20 10.
- 3200250 81.5o-7 00 60.0-50 L0nwA10 t tll e i gure I BtRA u
L Core at Left Edge IaReat Generation Vit tjfouge -.1s BORAL face (t =
A p
+ p w.id th, I, 0.0315in) for htne and xongttudil Figures 9aft dlata and CA:
PoSitionlý Z, the surface
TDR-0133 Revision 0 Alpha Heat Generation Data at fuel-side BORAL face (t = 0.0375 in)
RIOO 0-5
- 180-200 160-180 0 140-160 8 120-140 0 100-120 8 80-100 4 60-80
- 40-60 8 20-40 8 0-20 15
-~ -~
2.5 3.5 Position, - (in) 102 17 21A 306 Position, a (In)
Figure 8 Alpha Heat Generation Data at the Fuel-Side BORAL Face Alpha Heat Generation Fit at fuel-side BORAL face (t = 0.0375 in) 250 T ISOO 0 L 0 05 15 25 35
.S
.5
.3
.5
.3 9.5 10.2 IS 17 20
.8 1
23 530 30.6 34 6 200-250 0 150-200 I 100-150
.0-50 F..:-=
4O3 Azimuthal Position, w (In)
-4.15294M860 Longitudinal Position, z (i)
Figure 9 Alpha Heat Generation Fit for the Fuel-Side BORAL Face Table 3 compares the data obtained from MCNP and the values obtained from the functional form of the heat generation for the alpha heating profile at discrete points within the control blade.
13
TDR-01 33 Revision 0 Table 3 Alpha Volumetric Heat Generation Data and Fit Comparison at Left Edge (w = -3.115 in)
Longitudinal Volumetric Heat Volumetric Heat Eror Position, z Generation Data Generation Fit (in)
(W/cm 3)
(W/cm 3) 0.5 144.4 144.9 0.32 1.5 94.9 99.0 4.36 2.5 76.9 67.7 12.05 3.5 64.6 46.2 28.47 10.2 2.61 3.61 38.07 17 0.125 0.271 116.11 23.8 0.010 0.020 111.25 30.6 0.0024 0.0015 34.94 t
0.0875 in 0.5 28.4 39.2 37.99 1.5 20.7 27.6 33.67 2.5 17.3 19.5 12.44 3.5 18.2 13.7 24.71 10.2 2.59 1.32 49.01 17 0.115 0.122 6.16 23.8 0.0078 0.0113 45.30 30.6 0.0013 0.0011 15.93 t =0.1375 in 0.5 192.9 248.9 28.98 1.5 132.4 154.5 16.66 2.5 103.9 95.9 7.82 3.5 87.6 59.5 32.11 10.2 2.56 2.44 4.92 17 0.105 0.095 9.56 23.8 0.0060 0.0037 37.99 30.6 0.00015 0.00014 5.76 MODEL SETUP IN ABAQUS FEA A numerical simulation in Abaqus FEA typically consists of 3 separate stages namely the pre-processing stage, simulation evaluation stage and the post processing stage. The pre-processing stage involves the creation of an input file which contains all of the constraints under which the model is to be numerically evaluated. The material properties listed in Table 4 are input into the model during this stage. In the second stage the numerical analysis is evaluated based on the constraints specified in the pre-processing stage or the Abaqus FEA input file.
The post processing stage involves visualizing the results and extracting the data.
14
TDR-0133 Revision 0 Table 4 Material Properties of the BORAL and Cladding Used in the Analysis Material Property Aluminum 1100 BORAL Thermal Conductivity (W/m-K) 186.4 76.8, 98, 132 Density (Kg/mr3) 2713 2481 Elastic Modulus (GPa) 69 62.053 Poisson's Ratio 0.33 0.23 Thermal Expansion Coeff (K-)
2.36E-05 1.97E-05 Specific Heat (J/KgK) 1120 1380 The curvature of the control blade was created first using the dimensions stated in the control blade fabrication drawing and then extruded to 39.75 inches to form the full length of the control blade. This was followed by creating a partition at 34.25 inches to separate the bolted region from the region occupied by the BORAL (Figure 10). A cavity equivalent in curvature and length of the BORAL was cut-extruded to the length of the BORAL meat. To help simulate heating profile 'B,' the BORAL was partitioned into 4 sections (Figure 11) along the curvature.
Assembly of the model consisted of placing the BORAL meat in the cavity created in the control blade. The 'translation' option in Abaqus was used to perform the assembly by specifying the start and end points for translation. A fully coupled thermal stress step was created and perfect contact was assumed to exist between the BORAL meat and the aluminum cladding. Figure 10 illustrates the assembled model of the control blade.
A N
Figure 10 Assembled Model of the Control Blade 15
TDR-0133 Revision 0 THERMAL LOAD AND BOUNDARY CONDITIONS It was assumed that all the external surfaces of the aluminum cladding were exposed to coolant at 325 K that provided an effective heat transfer coefficient of 1000 W/m2-K. Perfect interfacial contact between the BORAL and aluminum cladding was assumed. The internal heat generation profile obtained based on the curve fit data was used as the loading condition on the BORAL.
Figure 11 BORAL Meat Split into 4 Sections to Apply the Load MECHANICAL BOUNDARY CONDITIONS The mechanical boundary conditions were applied to the holes of the control blade mounting plate to prevent translation of these holes in the radial and longitudinal direction as illustrated in Figure 12.
16
TDR-0133 Revision 0 Figure 12 Mechanical Boundary Condition Applied to the Mounting Holes of the Control Blade RESULTS Figures 13 through 16 illustrate the thermal and corresponding mechanical behavior of the control blade.
Kelvin Kir 1I+3.4.4.+l 3-4*ft
- U
+314+
,+3:lm+eIN
- +33Z~ r l
+3.i1+
V Figure 13 Temperature Distribution Contour of the Control Blade Figure 13 illustrates that the right side of the blade is hotter than the other side and this is consistent with the curve fits (Figures 6-9) that were put together based on the data provided by 17
TDR-0133 Revision 0 MCNP. The variation of the temperature along the cladding centerline is presented in Figure 14.
The temperature also drops off in the longitudinal direction where the heat generation rate also drops off.
Temperature distribution along the dadding centerline 344 342 I 336 R
3-Seate' Rqoss" 332 330 328 326 324 0
10 20 30 40 so Length along the blade (Inch)
Figure 14 Temperature Distribution along the Cladding Centerline The radial deflection of the blade resulting from the temperature is shown in Figure 15. A first look at this figure gives the impression that the deflection is symmetrical when Figure 13 shows that asymmetry in temperature exists along the curvature of the blade. To verify that asymmetry exists in the radial deflection along the curvature, the deflection as a function of the circumference was plotted as illustrated in Figure 16. This plot shows that the right edge of the blade (which is the hotter side) deflects slightly more than the other side and the difference in deflection is 1.2 x 10-3 inches.
18
TDR-0133 Revision 0 II. Ui VAISENHUYT-DATUM cvts-z*
+&2330-03
- S.Yof.103
+4.127.42 0.10
+240
- 3.3e42 060
- LSU.4 90.HO V
Figure 15 Radial Deflection Contour of the Control Blade Tem atrm df U
Magnitude of deflection on the sides of the blade 0.240 -
0.+
\\(++~*++*+/-+*
0.235 mad. Center 1+
I U
Ammedpoidw dodcwmmidm
- dkoton for X sibito renuh Thmmakesft euleto
-I.wth dtfi~cttuL
'0225
+
g f C
0.215 0.210 1 0.205 f
t
. Left side
+/--t +
+ Right side 0.200 -
Blade Edge 0.195 0
5 10 15 20 25 30 35 40 Angle, theta (Degree)
Figure 16 Deflection Variation on the Sides of the Control Blade 19
TDR-0133 Revision 0 Radial displacernent along the cladding centerline (Profile-B) k 76.8 WlmK 0.27-
-50.24 O2.7 0.18 0.15 C
0.12 0.09 0.06 i
0.03 Boral Region Iolted, Region N..
k98 W/mK
...... k 132 WImK
-Control blade channel gap V
I i
i 0
10 20 30 40 50 Length along the blade (inch)
Figure 17 Deflection Variation Along the Length of the Control Blade with Varying BORAL Thermal Conductivity BORAL THERMAL CONDUCTIVITY STUDY The thermal conductivity of the BORAL used in the current analysis is 76.8 W/m-K as illustrated in Table 4. This value was obtained from BORAL Composite Standard Specifications, Ceradyne Canada, ULC [2]. However, experiments performed at the University of Missouri on BORAL coupons of the same material specifications of a MURR control blade suggested that the thermal conductivity of the BORAL was 115+17 W/m-K. A parametric study was performed by varying the thermal conductivity of the BORAL with a lower bound of 98 W/m-K and an upper bound of 132 W/m-K. The change in maximum radial deflection was found to be negligibly small and of the order of 4 x 10-4 inches. Figure 17 illustrates that the maximum radial deflection of the blade is within the control blade channel gap even at the upper bound (132 W/m-K) of the BORAL thermal conductivity. It should be noted that the three different thermal conductivity values modeled provide deflections that lie nearly on top of each other making them difficult to distinguish in the plot.
CONCLUSION The commercial finite element code Abaqus FEA was used to perform a thermal mechanical analysis on a MURR BORAL control blade. The task was to establish that the deflection of the blade due to heating will be within the allowable limit of 0.26 inches (6.5 mm).
The BORAL meat was split into 4 quadrants and a heat generation profile that varied with thickness and longitudinal position was applied to each quadrant. The magnitude of the resulting thermally induced deflection was within the allowable deflection limit of 0.26 inches (6.5 mm).
The heating profile showed some asymmetry in temperature and the corresponding asymmetric 20
TDR-0133 Revision 0 deflection along the blade curvature has been discussed.
Finally, thermal conductivity experiments on BORAL coupons of the same material specifications as the MURR control blades suggest even higher thermal conductivity values than what are stated in Reference 2, which would result in even less deflection.
REFERENCES
[1]
Govindarajan, S.G., Moreland, J.A, Solbreken, G.L., BORAL Control Blade Thermal-Mechanical Analysis, Mechanical and Aerospace Engineering Department, University of Missouri, Columbia, Missouri, June 2012.
[2]
BORAL Composite Standard Specifications, SP-BORA-00len, Rev.
10-17-08, Ceradyne Canada, ULC.
21