ML14121A028

NRR E-mail Capture - Testimony of Oregon and Washington Psr - Post-PRB Meeting for Columbia Seismic Concerns 2.206 � Teleconference - 4/30/14

ML14121A028
Person / Time
Site:	Columbia
Issue date:	04/30/2014
From:	Clay Johnson Washington Physicians for Social Responsibility
To:	Lyon F Division of Operating Reactor Licensing
References
G20130776, MF3031
Download: ML14121A028 (20)

Text

1 NRR-PMDAPEm Resource From:

johnsonc20@gmail.com on behalf of Charles K. Johnson [chuck@oregonpsr.org]

Sent:

Wednesday, April 30, 2014 2:18 PM To:

Lyon, Fred

Subject:

Testimony of Oregon and Washington PSR - Post-PRB Meeting for Columbia Seismic Concerns 2.206 - Teleconference - 4/30/14 Attachments:

Stamped Cover letter Volc & flooding 02Jan2014.pdf; Flooding Hazards at CGS vmf Jan 2014-1.docx; Figure 1 CGS topo map.pdf; Figure 2 GC Dam flood map.pdf; Figure 3 GW contour map.pdf; Figure 4 CGS x-sec profile GW levels vf.pdf; 20111003-ucs-brief-bwr-scram-problem-1.pdf April 30, 2014 From: Charles Johnson, Director, Joint Task Force on Nuclear Power, Oregon and Washington Physicians for Social Responsibility To: C. Fred Lyon, Licensing Project Manager for the Columbia Generating Station, U.S. Nuclear Regulatory Commission Re: Post-PRB Meeting for Columbia Seismic Concerns 2.206 - Teleconference Our statement for requesting this meeting was as follows:

Given that an accident caused by a beyond design basis earthquake at the WESF [Waste Encapsulation and Storage Facility] is exactly the scenario we fear for the CGS nuclear plant at Hanford - and that the vulnerable spent fuel pool at the CGS cannot be completely emptied without shutting down the reactor and waiting for several years - we believe that closing the reactor now until appropriate seismic upgrades can be made at the plant is the most prudent course.

http://energy.gov/ig/downloads/audit-report-oas-l-14-04 This report is described in a Tri-City Herald article on April 3, 2014: http://www.tri-cityherald.com/2014/04/02/2908764/inspector-general-says-hanford.html It would seem that the same things that made the US DOE's OIG concerned about the beyond design earthquake hazard at the WESF facility should be of equal concern to the Nuclear Regulatory Commission and Energy Northwest. Degraded concrete and earthquakes that create greater than planned for ground motion are not a good combination for containing a reactor and an elevated pool filled with spent nuclear fuel.

To this statement I would add some links to studies on the behavior of dry concrete (which the CGS spent fuel pool - due to its liner - would be) when bombarded by high doses of radiation.

http://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr5279/#pub-info http://www.claisse.info/2010%20papers/m17.pdf http://przyrbwn.icm.edu.pl/APP/PDF/114/a114z211.pdf

2 http://link.springer.com/article/10.1023%2FA%3A1010971122496#page-1 You will notice that the first document is an NRC document and the fourth is by the same two authors of the NRC report, so this issue of concrete degradation is known. Nevertheless, it is not clear that there has been an analysis of the concrete composition of the CGS spent fuel pool and other critical places in the reactor building where concrete reinforcement is exposed to radiation - and an assessment of the amount of damage that may have been done to the concretes structural integrity.

Given that we know the site has greater seismic potential, and thus more ground motion, than was known when the plant was built, it would make sense for the NRC to show as much caution as the Office of the Inspector General for US Department of Energy did when they stated in their March 26, 2014 memorandum accompanying the Long-Term Storage of Cesium and Strontium at the Hanford Site:

One possible threat is a severe earthquake that may result in loss of power and/or loss of water in the WESF pool. The Department's Office of Environmental Management considers WESF its largest beyond design threat facility, and has identified movement of the capsules to dry storage as a potential interim measure to mitigate the risk posed by these threats.

I want to point out that this is the second time that the US DOE has taken action or recommended taking action based upon the revised earthquake knowledge on the Hanford site, the first being the one year delay in construction of the Waste Treatment Plant in 2005 until it could be reinforced to meet the higher ground motion now known to be possible on the Hanford Reservation.

In contrast, the Nuclear Regulatory Commission has remained content to wait until March 2015, when Energy Northwest completes its seismic review before it requires any modification of the seismic standards the plant needs to meet. We ask that you consider this concrete degradation issue as one more reason why the Columbia Generating Station should be shut down until it can be positively determined it can withstand a worst case earthquake. In the case of potential nuclear plant accidents, the NRC should err on the side of caution. The downside of being wrong is just too heavy a burden to bear.

Another issue I wish to forward to you is not strictly one that is earthquake related, but it does outline a potential pathway for a catastrophic accident.

In the course of organizing an independent study of the CGS nuclear plant, Oregon and Washington PSR hired Licensed Engineering Geologist Terry Tolan to conduct a series of studies - the first of which was an analysis of current knowledge of the potential for seismic activity at the CGS site, which was the basis for our letter to Chairwoman Allison Macfarlane and this 2.206 petition process.

3 Mr. Tolan also completed studies of volcanic activity and flooding. In the volcanic activity study, Tolan found the preparations at the plant to be adequate to meet the potential threat. In his study of potential flooding danger, which is attached along with this testimony (with his cover memo and four maps), Tolan found that the NRCs worst case flooding scenario of a terrorist attack completely destroying the Grand Coulee Dam, did not take into account two major issues:

1) the chaos a huge wall of water - completely inundating the City of Richland - would create in destroying the infrastructure and making access to the site difficult, if not impossible, for additional personnel and equipment;

and,

2) the combination of elevated groundwater levels and a worst case dam burst at Grand Coulee would raise ground water levels four feet higher than the design basis ground water level for the CGS, "and could put structures, systems, and components at risk."

Please note that we are not claiming an earthquake could cause this water inundation scenario, as Mr. Tolan agrees with the NRC that an earthquake would not release the full volume of water that a terrorist detonation could, in that there would likely be remnants of the dam in place that would reduce the size of the flood.

Finally, I want to note of couple of objections to the rationale for rejecting our previous appeal, which you sent to me by email on March 24, 2014. I quote your concluding statement:

[T]he PRBs initial recommendation is to reject the petition, in accordance with MD 8.11 Handbook Part III, paragraph C.2, Criteria for Rejecting Petitions Under 10 CFR 2.206, because the petitioners raise issues that have already been the subject of NRC staff review and evaluation either on that facility, other similar facilities, or on a generic basis, for which a resolution has been achieved, the issues have been resolved, and the resolution is applicable to the facility in question.

We question whether the issue of GE BWRs potential inability to insert control rods during an earthquake has been resolved, as we have been shown no evidence that it has. This is a particularly serious issue and I attach the Union of Concerned Scientists website lay analysis of the issue to refresh the record.

The combination of it not being clearing established that the CGS nuclear power plant can withstand an earthquake with its key cooling and power systems intact and this persistent, unsolved problem of control rods sticking in an earthquake could lead to an enormously damaging catastrophe if the reactor looses coolant with no ability to shut it down.

When you add to all of these issues the fact that the Nuclear Regulatory Commission has now given Energy Northwest until 2017 to complete hardened vents on its containment - vents that are designed to prevent a hydrogen explosion that would most likely occur at the plant in a Fukushima style loss of coolant accident -

4 you have to conclude that the NRC is taking a very lackadaisical approach toward the safety of the workers at the plant, the people of the Tri-Cities, and everyone downstream along the Columbia River. To wait six years after the Fukushima accident to install vents you know are needed at this plant is unconscionable. The Columbia Generating Station should be shut down immediately for this reason alone.

The message is a basic one - shut down the plant until you can make certain that it will not suffer a catastrophic accident. We believe the 2.206 petition gives you the opportunity you need to perform your function of safeguarding the public and we urge you to take it.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Charles K. Johnson Director, Joint Task Force on Nuclear Power Oregon and Washington Physicians for Social Responsibility 812 SW Washington Street, Suite 1050 Portland, OR 97205 (503) 777-2794 cell chuck@oregonpsr.org

Hearing Identifier:

NRR_PMDA Email Number:

1261 Mail Envelope Properties (CAHWtKjek_oyYXkfn266EDcVjC3vVYDYP2+kTJxcQBk=WsBvEWg)

Subject:

Testimony of Oregon and Washington PSR - Post-PRB Meeting for Columbia Seismic Concerns 2.206 - Teleconference - 4/30/14 Sent Date:

4/30/2014 2:17:55 PM Received Date:

4/30/2014 2:20:47 PM From:

johnsonc20@gmail.com Created By:

johnsonc20@gmail.com Recipients:

"Lyon, Fred" <Fred.Lyon@nrc.gov>

Tracking Status: None Post Office:

mail.gmail.com Files Size Date & Time MESSAGE 8847 4/30/2014 2:20:47 PM Stamped Cover letter Volc & flooding 02Jan2014.pdf 234742 Flooding Hazards at CGS vmf Jan 2014-1.docx 29465 Figure 1 CGS topo map.pdf 707312 Figure 2 GC Dam flood map.pdf 466542 Figure 3 GW contour map.pdf 601626 Figure 4 CGS x-sec profile GW levels vf.pdf 394562 20111003-ucs-brief-bwr-scram-problem-1.pdf 487808 Options Priority:

Standard Return Notification:

No Reply Requested:

No Sensitivity:

Normal Expiration Date:

Recipients Received:

Page 1 of 6 REVIEW OF ENERGY NORTHWESTS EVALUATION OF POTENTIAL HAZARDS FROM FLOODING AT THE COLUMBIA GENERATING STATION Introduction As part of the Nuclear Regulatory Commissions licensing process, Energy Northwest evaluated the potential hazards from flooding at the Columbia Generating Station (formerly designated WNP-2) site in the Final Safety Analysis Report (Energy Northwest, 2011) and was reviewed in 2012 (ENERCON, 2012). The Columbia Generating Station is located approximately 3 miles west of the Columbia River at River Mile 352 and is situated on an upland area (Figure 1). The potential flooding scenarios identified and evaluated by Energy Northwest included:

1. Probable maximum flood resulting from local intense precipitation events (i.e., summer thunderstorms).

2. Probable maximum natural flood event on the Columbia River.

3. Probable maximum flood event produced by failure of dam(s) on the Columbia River upstream of the site.

4. Probable maximum rise of the top of the unconfined aquifer (water table aquifer) beneath the site.

For scenarios 1 through 3, Energy Northwest (2011) calculated highest flood water levels for these events and developed design-basis flood criteria for the Columbia Generating Station safety-related structures, systems, and components. In the case of scenario 4, events that could potentially raise the elevation of the top of the unconfined aquifer (water table) beneath the site were assessed. The following sections present a brief review of Energy Northwests analysis of the hazards posed to the Columbia Generating Station by each of these scenarios.

1. Probable Maximum Flood Due to Local Intense Precipitation Events It was determined that intense local thunderstorms have the greatest potential for causing serious flooding (Energy Northwest, 2011, Section 2.4.2.3). Energy Northwest (2011) calculated the probable maximum precipitation event for the Columbia Generating Station site would be 9.2 inches in a 6-hour period using methodology developed by the U.S. Weather Bureau (1966).

Runoff from such an event would drain easterly from the site and collect in the in the low-lying, broad, north-south-trending valley (channel) immediately adjacent to the site (Figure 1). Their calculated elevation of the surface of the flood waters in the valley east of the site as approximately 431.1 feet above mean sea level (Energy Northwest, 2011). An additional 2.2 feet was added to account for possible wind wave action, making the design-basis flood elevation equal to 433.3 feet above mean sea level for a probable maximum precipitation event. The general elevation of the Columbia Generating Station plant site is typically 7 to 9 feet above the design-basis flood elevation for a probable maximum precipitation event, it was concluded that the plant site would not be affected by such a flooding event (Energy Northwest, 2011; ENERCON, 2012).

My review of Energy Northwests (2011) analysis of potential flooding at the Columbia Generating Station from a probable maximum precipitation event appears to be supported by the available data and their conclusions are valid.

Page 2 of 6

2. Probable Maximum Natural Flood Event on the Columbia River Energy Northwest (2011) analyzed hydrographs and reports pertaining to past flood events on the Columbia River, both pre-and post-dam regulation of the Columbia River flow, upstream from the Columbia Generating Station. Based on their analysis of the available data and information, Energy Northwest (2011, p. 2.4-6) calculated that a probable maximum flood event on the Hanford reach of the Columbia River would have an estimated discharge rate of 1,440,000 cubic feet per second and a water level elevation of 390 feet above mean sea level.

Given that the general elevation of the Columbia Generating Station plant site is greater than 440 feet above mean sea level, it was concluded that critical plant site structures, systems, and components would not be affected by a probable maximum flooding event on the Columbia River (Energy Northwest, 2011; ENERCON, 2012). The design-basis flood for the Columbia Generating Station site is that produced by a probable maximum precipitation event and not a natural flooding event of the Columbia River.

My review of Energy Northwests (2011) analysis of potential flooding at the Columbia Generating Station from a probable maximum natural flood event on the Columbia River appears to be supported by the available data and their conclusions are valid.

3. Probable Maximum Flood Event Produced by Failure of Dam(s) on the Columbia River Upstream of the Columbia Generating Station Site Energy Northwest (2011) conducted an analysis of the probable maximum flood at the Columbia Generating Station site produced by a number of upstream potential dam failure scenarios. The primary focus was on the possible seismically (earthquake) induced failure of Grand Coulee Dam which is located approximately 245 river miles upstream of the Columbia Generating Station site. Analysis of potential Grand Coulee Dam seismic failure scenarios by U.S. Bureau of Reclamation (letter reports referenced in USDOE, 1988) suggest that only a portion of the dam might be displaced resulting in a limited breach of the dams structure and would limit the potential release of water. A declassified 1951 study by the Seattle District Corps of Engineers examined the effects of an explosion-induced failure of Grand Coulee Dam was made available to Energy Northwest (2011). The explosion-induced failure, and resulting flood, was determined to represent an upper limit to potential seismically induced failures (worse-case flood scenario) for the Columbia Generating Station site by Energy Northwest (2011).

Other assumptions used by Energy Northwest (2011, p. 2.4.10 -2.4.11) in determining flood water levels at the Columbia Generating Station site from a Grand Coulee Dam failure scenario include:

The Columbia River is at flood stage (570,000 cubic feet per second).

The reservoirs in each downstream dam storage pools are full.

A massive hydraulic failure occurs at Grand Coulee Dam releasing 8,800,000 cubic feet per second. The explosion-induced failure of Grand Coulee Dam represents a more severe failure than any seismic event because of the failure mechanism.

Page 3 of 6 Following failure of Grand Coulee Dam, all downstream dams between the Columbia Generating Station and Grand Coulee Dam suffer some degree of failure and release their storage to the flood.

Approximately 23 hours2.662037e-4 days <br />0.00639 hours <br />3.80291e-5 weeks <br />8.7515e-6 months <br /> after failure of Grand Coulee Dam the initial flood surge would arrive at the Hanford reach of the Columbia River (Columbia Generating Station area) and the flood peak would occur approximately 38 hour4.398148e-4 days <br />0.0106 hours <br />6.283069e-5 weeks <br />1.4459e-5 months <br /> after failure (ENERCON, 2012).

Failure of Arrow and/or Mica Dams, upstream of Grand Coulee Dam in Canada, could result in a greater release of storage volume than Grand Coulee Dam alone. However, the peak flow from the failure of Arrow and Mica Dams is limited to 3,100,000 cubic feet per second due to river canyon restrictions (hydraulic damming) at Trail, British Columbia.

Failure of Arrow and/or Mica Dams in Canada would result in flood waters overtopping Grand Coulee Dam, but the overtopping by flood waters is not considered to be sufficient to cause a failure of Grand Coulee Dam in view of its concrete construction and rock abutments.

Based on the above scenario and assumptions, Energy Northwest (2011) calculated and estimated a maximum elevation of the flood waters in the vicinity of the Columbia Generating Station area to be 422 feet above mean sea level (Figure 2). They also factored in an allowance for simultaneous wind and wave action of 2 feet which increased the maximum elevation of the flood waters to 424 feet above mean sea level. Given that the general ground elevation of the Columbia Generating Station plant site is greater than 440 feet above mean sea level, it was concluded that critical plant site structures, systems, and components would not be directly affected by a probable maximum flooding event (Figure 2) resulting from the failure of Grand Coulee Dam (Energy Northwest, 2011).

However Energy Northwest (2011) does not address the potential indirect impacts of such a flood on the operations of the Columbia Generating Station. While the flood waters would not inundate the Columbia Generating Station site, downstream of the Columbia Generating Station the flood waters would inundate large portions of the cities of Richland, Pasco, and Kennewick where the Columbia Generating Stations work force lives. It is estimated based on the flood model that flood waters would likely be more than 65 feet deep in downtown Richland and would also cut-off highway access through the City of Richland to the Columbia Generating Station. Even with 23 hours2.662037e-4 days <br />0.00639 hours <br />3.80291e-5 weeks <br />8.7515e-6 months <br /> warning, undoubtable there would be significant disorder resulting from having to evaluate much of the areas population from low-lying areas. No plans were found within the available Energy Northwest documents that address how staff critical to the operation of the Columbia Generating Station would be organized and conveyed to the site during, and in the aftermath of, such a major flood event.

My review of Energy Northwests (2011) analysis of potential flooding at the Columbia Generating Station from a failure of Grand Coulee Dam appears to be supported by the available data and their conclusions are valid with one potentially significant oversight with regards to having any plans that address how staff critical to the operation of the Columbia Generating Station would be organized and conveyed to the site during and in the aftermath of such a major flood.

Page 4 of 6

4. Probable maximum rise in the unconfined aquifer (water table aquifer) beneath the site The final potential source of flooding at the Columbia Generating Station site could result from a rise in the elevation of the unconfined groundwater aquifer beneath this area (Energy Northwest, 2011; ENERCON, 2012). A drastic rise in the elevation of the top of the unconfined groundwater aquifer could subject the foundations of critical site structures to force effects of buoyancy (uplift) and lateral hydrostatic pressures (ENERCON, 2012). The top of the unconfined groundwater aquifer beneath the Columbia Generating Station site is below the top of the Ringold Formation at an elevation of approximately 378 + 4 feet above mean sea level (ENERCON, 2012).

Given the high porosity and permeability of the sediments (Hanford and Ringold Formations) beneath the Columbia Generating Station, few credible scenarios exist that would result a significant rise in the elevation of the top of the unconfined aquifer. Newcomb and Brown (1961), Newcomb et al. (1972), Gephart et al. (1979), and USDOE (1988) have described seasonal rise and decline in the elevation of the unconfined aquifer along the Hanford reach of the Columbia River that corresponded with annual flood flow in the river. Newcomb and Brown (1961) noted that the elevation fluctuations in the top of the unconfined aquifer ranged from 2 to 19 feet and the effects of the elevation fluctuations could be documented up to 3 miles inland from the Columbia River on the Hanford Site. However, Energy Northwest (2011, p. 2.4-23) noted that unconfined aquifer wells in the vicinity of the Columbia Generating Station did not detect any seasonal variability in the elevation of the top of the unconfined aquifer. So this natural seasonal event was not selected for determining maximum groundwater elevation (design-basis) beneath the Columbia Generating Station.

Instead the Columbia Generating Stations design-basis maximum groundwater level was based on the possible construction of the Ben Franklin Dam (Figure 3) which could have raised the top of the unconfined aquifer elevation beneath the site to between 405 to 420 feet above mean sea level (Harty, 1979; Energy Northwest, 2011). Plans to construct the Ben Franklin Dam were terminated more than 30 years ago, but the 420 foot above mean sea level elevation was chosen for the design-basis groundwater elevation for the Columbia Generating Station. As noted by ENERCON (2012), uplift and increased lateral hydrostatic pressure are considered in the original design of all Columbia Generating Station Seismic Category 1 structures, safety-related systems, and components to ensure their safety in the event of a rise in the groundwater table to elevation 420 feet above mean sea level.

Critical Columbia Generating Station structures at, or below, 420 foot elevation were designed to resist the increased hydrostatic pressure which would result from a rise in the groundwater level to 420 foot elevation. Columbia Generating Station structures above the design-basis maximum groundwater level of 420 feet above mean sea level (e.g., top of the foundation mat, lowest floor surface in the reactor building at an elevation of 422 feet 3 inches above mean sea level; ENERCON, 2012) are considered unaffected and sealing against groundwater pressure is therefore not required.

Page 5 of 6 My review of Energy Northwests (2011) analysis of the hazards associated with a rise in the elevation of the top of the unconfined aquifer appears to be supported by available data and their conclusions are valid for the scenarios that were covered. However, it appears that Energy Northwest (2011) failed to address the effects that the probable maximum flood resulting from the failure of Grand Coulee Dam would have on the unconfined aquifer beneath the Columbia Generating Station.

As reviewed in the previous section, Energy Northwest (2011) has estimated that flood waters, created by the failure of Grand Coulee Dam, would have a maximum elevation in the vicinity of the Columbia Generating Station area (Figure 2) of 422 to 424 feet above mean sea level. During this event flood waters would enter the banks of the Columbia River and proceed inland beneath the ground surface. The inland subsurface movement of the flood waters would likely experience only minor impedance, especially since this area is underlain by the unconsolidated coarse sands and gravels of the Hanford Formation (Newcomb and Brown, 1961; Newcomb et al., 1972; Gephart et al., 1979). It is possible that the top of the unconfined aquifer beneath the Columbia Generating Station would temporally rise to an elevation of 424 feet above mean sea level during this probable maximum flood event (Figure 4). This would result in the elevation of the top of the unconfined aquifer being up to 4 feet higher than the design-basis maximum groundwater level for the Columbia Generating Station (Figure 4) and could put critical structures, systems, and components at risk.

Summary of Findings Four potential flooding scenarios were identified and evaluated by Energy Northwest included:

1. Probable maximum flood resulting from local intense precipitation events (i.e., summer thunderstorms).

2. Probable maximum natural flood event on the Columbia River.

3. Probable maximum flood event produced by failure of dam(s) on the Columbia River upstream of the site.

4. Probable maximum rise of the top of the unconfined aquifer (water table aquifer) beneath the site.

For flooding scenarios 1 and 2, my review of Energy Northwests (2011) analyses of these potential events appears to the supported by the available data and their conclusions are valid.

For flooding scenario 3, my review of Energy Northwests (2011) analysis of potential flooding at the Columbia Generating Station from a failure of Grand Coulee Dam scenario appears to be supported by the available data and their conclusions are valid with one potentially significant oversight. No plans were found that addresses how staff critical to the operation of the Columbia Generating Station would be organized and conveyed to the site during and in the aftermath of such a major flood scenario.

For flooding scenario 4, my review of Energy Northwests (2011) analysis of the hazards associated with a rise in the elevation of the top of the unconfined aquifer appears to be supported by available data and their conclusions are valid for the initiating events that were addressed. However, it appears that Energy Northwest (2011) failed to address the effects that

Page 6 of 6 the probable maximum flood resulting from the failure of Grand Coulee Dam would have on the unconfined aquifer beneath the Columbia Generating Station. The flood waters from the failure of Grand Coulee Dam (scenario 3) would enter the banks of the Columbia River and proceed inland beneath the ground surface. The inland subsurface movement of the flood waters would likely experience only minor impedance, especially since this area is underlain by the unconsolidated coarse sands and gravels of the Hanford formation. It is possible that the top of the unconfined aquifer beneath the Columbia Generating Station would temporally rise to an elevation of 424 feet above mean sea level during this probable maximum flood event (Figure 4).

This would result in the elevation of the top of the unconfined aquifer being up to 4 feet higher than the design-basis maximum groundwater level for the Columbia Generating Station (Figure

4) and could put critical structures, systems, and components at risk.

References Cited ENERCON, 2012, Flood protection final report - in response to 10 CFR 50.54(f) information request regarding near-term task force recommendations 2.3 - Flooding: Prepared by ENERCON, Kennesaw, Georgia, for Energy Northwest Columbia Generating Station, Richland, Washington, Rev. 0, 13 p.

Energy Northwest, 2011, Columbia Generating Station Final Safety Analysis Report, Amendment 61, Hydrology Engineering: Energy Northwest, Richland, Washington, chp. 2, section 2.4, p. 2.4.1-2.4.34.

ERDA, 1976, Evaluation of the impact of potential flooding criteria on the Hanford Project: U.S.

Energy Research and Development Administration, Richland, Washington, Report RLO-76-4.

Gephart, R.E., Arnett, R.C., Baca, R.G., Leonhart, L.S., and Spane, F.A., Jr., 1979, Hydrologic studies within the Columbia Plateau, Washington - an integration of current knowledge:

Rockwell Hanford Operations, Richland, Washington, RHO-BWI-ST-5, 547 p.

Newcomb, R.C., and Brown, S.G., 1961, Evaluation of bank storage along the Columbia River between Richland and China Bar, Washington: U.S. Geological Survey Water-Supply Paper 1539-I, 13 p.

Newcomb, R.C., Strand, J.R., and Frank, F.J., 1972, Geology and groundwater characteristics of the Hanford Reservation of the U.S. Atomic Energy Commission, Washington: U.S.

Geological Survey Professional Paper 717, 78 p.

USDOE (U.S. Department of Energy), 1988, Site characterization plan - Reference Repository Location, Hanford Site: Office of Civilian Radioactive Waste Management, Washington, DC, DOE/RW-0164, v. 2, chp. 3.

U.S. Weather Bureau, 1966, Probable maximum precipitation - Northwest States:

Hydrometeorological Report no. 43.

Figure 1. Topographic map showing the location of the Columbia Generating Station site in relation to the Columbia River. The elevation of the Columbia Generating Station site ranges from 450 to 420 feet above mean sea level and the elevation of the Columbia River is approximately 350 feet above mean sea level.

Columbia Generating Station Columbia River (flows north to south)

Figure 2. Map showing the approximate extent of flooding at the Hanford Site resulting from a 50% breach of Grand Coulee Dam (discharge 8,000,000 cubic feet per second; ERDA, 1976).

Map reproduced from USDOE (1988, v. 2, chp. 3, p. 3.2-3).

Columbia Generating Station

Figure 3. Map showing the proposed location of the Ben Franklin Dam and resulting elevation of the top of the water table (unconfined aquifer) beneath the Columbia Generating Station.

Reproduced from Energy Northwest (2011).

440 420 400 380 360 Elevation (ft amsl)

CROSS-SECTION A-A EXPLANATION 378 + 4 ft amsl range of the top of the unconfined aquifer beneath the CGS site (ENERCON, 2012) 420 ft amsl design-basis maximum groundwater level at CGS 390 ft amsl probable maximum natural flood event on the Columbia River - impact on groundwater level beneath CGS 424 ft amsl probable maximum flood event produced by Grand Coulee Dam failure - impact on groundwater level beneath CGS Figure 4. Cross-section A-A profile through the Columbia Generating Station (CGS) site showing depth of foundation structures in relation various projected groundwater levels. Cross-section A-A modified from WPPSS (1981, Figure 2.5-67).

Boiling Water Reactor Shut Down System Problem GE Hitachi informed the Nuclear Regulatory Commission (NRC) about a safety problem related to the reactor shut down system at its boiling water reactors (BWRs) via a September 27, 2011 update to NRC Event 46230 dated September 3, 2010:

GE Hitachi (GEH) has determined that the scram capability of the control rod drive mechanism in BWR/2-5 plants may not be sufficient to ensure the control rod will fully insert in a cell with channel-control rod friction at or below the friction limits specified in MFN 08-420 with a concurrent Safe Shutdown Earthquake (SSE). The plant condition for which incomplete control rod insertion might occur is when the reactor is below normal operating pressure (<900 psig) and a scram occurs concurrent with the SSE, for Mark I containment plants, and for the SSE with concurrent Loss-of-Coolant Accident (LOCA) and Safety Relief Valve (SRV) events for Mark II containment plants. In this scenario a Substantial Safety Hazard results because the affected control rods might not fully insert to perform the required safety function.

GEH has determined that when channel-control blade interference is present at reduced reactor pressure and at friction levels considered acceptable in MFN 08-420, a simultaneously occurring Safe Shutdown Earthquake (SSE) may result in control rod friction that inhibits the full insertion of the affected control rods during a reactor scram from these conditions. This scenario was not explicitly considered in MFN 08-420.

This issue brief provides background information about control rods at BWRs and this specific safety problem.

The reactor core of a BWR consists of fuel pellets stacked within hollow metal rods that are formed into fuel assemblies. Each fuel assembly is housed within a hollow metal box called a channel. The BWR core has over 100 fuel cells, each consisting of four fuel assemblies with one control rod in the middle.

The BWR core is powered by the fissioning, or splitting, of atoms in the fuel pellets. Energy, and sub-atomic particles called neutrons, are released when certain atoms split. The energy boils water flowing past the fuel assemblies, hence the name boiling water reactor. The neutrons interact with other atoms, causing them to become unstable and split to sustain the nuclear chain reaction.

October 3, 2011 Page 2 of 4 Control rods function as both the gas pedal and the brake pedal for the reactor core of a BWR. A control rod is X-shaped. Each vane of the X contains vertical metal tubes filled with boron powder. Boron is like neutron glue. When control rods are withdrawn from the reactor core, fewer neutrons released by fissioning atoms get absorbed by the boron inside control rods, leaving them free to cause other atoms to split. This increases the power level of the core. Conversely, when control rods are inserted, their boron absorbs neutrons to slow down, or even stop, the nuclear chain reaction.

Control rods are moved using water pressure acting upon pistons. Water pressure is applied to one side of the piston and vented from the other side to create the force necessary to move the control rod. By swapping which side of the piston receives water and gets vented, the resulting movement can cause the control rod to enter the reactor core or be withdrawn from it. During normal operation, differential pressure of approximately 250 to 300 pounds per square inch causes the control rod to move at a rate of about 3 inches per second, and it takes about 48 seconds for a control rod to travel the entire length from fully inserted to fully withdrawn or vice-versa. In an emergency, differential pressure of over 1,000 pounds per square inch can accelerate movement such that a control rod can move from fully withdrawn to fully inserted within 3 seconds.

The problem that GE Hitachi reported to the NRC in September 2010 involved increased control rod insertion times under emergency conditions because the control rods were slowed down by rubbing against the fuel channels around the fuel assemblies. As the fuel cell schematic on the preceding page illustrates, each fuel channel has two spacer buttons. The spacer buttons on adjacent fuel assemblies are designed to touch to keep space open for the control rod to move freely between them. But the control rods and fuel channels are made of metal that expands when heated. When not properly accommodated, this expansion can cause warping or bowing - the control rod can bend towards the fuel channels and/or the fuel channels can bend towards the control rod. When this occurs, the rubbing of the control rod against the fuel channels can slow it down or even stop it.

The initial problem GE Hitachi reported in September 2010 was limited to a certain type of control rod manufactured by GE Hitachi that was susceptible to rubbing under certain conditions. This initial problem was addressed by requiring owners of reactors equipped with these control rods to time the control rod movements more frequently to check for rubbing. The more frequent testing would continue until the problematic control rods could be replaced.

The September 2011 update to the initial report added a wrinkle that had not been previously considered.

Certain conditions, like an earthquake, could increase the likelihood that control rods would rub against channels. For example, shaking caused by an earthquake could cause fuel assemblies to twist and flex, narrowing or even eliminating the gap in which the control rods move. GE Hitachi reported the worst case scenario to be an event occurring when the pressure inside the reactor is below the normal operating pressure of 1,000 pounds per square inch.

October 3, 2011 Page 3 of 4 The water pressure used to move control rods in an emergency comes from two sources: (1) accumulators, and (2) the reactor vessel housing the reactor core itself. Each control rod is equipped with its own accumulator. An accumulator has two parts: a metal cylinder filled with water and a rounded tank pressurized with nitrogen gas. The picture on the left below shows two complete accumulators and the nitrogen tank from a third. The gauges at bottom indicate the pressures of the nitrogen inside the tanks.

The chart on the right above shows the time it takes a fully withdrawn control rod to insert into the reactor core during an emergency as a function of the reactor pressure. The yellow curve shows the insertion time when the differential pressure to move the control rod comes from only the accumulator. When the reactor pressure is low, the high pressure inside the accumulator rapidly inserts the control rod with little force to oppose it. As the reactor pressure increases (thus increasing the pressure holding the control rod out), it takes longer for the accumulator pressure to insert a control rod. The cyan (blue) curve shows the insertion time when the differential pressure to move the control rod comes from only the reactor vessel.

In this case, the insertion time shortens as the reactor pressure increases - exactly the opposite reaction from the accumulator case.

The accumulators are pressurized to 1,200 to 1,400 pounds per square inch (psi). Water from the accumulators cylinder is applied to one side of the control rod piston while the other side is vented to the atmosphere. The differential pressure across the piston drives the control rod into the reactor core. As the pressure inside the reactor vessel increases, the differential pressure across the piston remains the same (1,200 to 1,400 psi on one side and atmospheric on the other) but the control rod is moving against more force, which slows it down.

When the control rod is inserted using only pressure inside the reactor vessel, the differential pressure across the piston still causes that movement. But as the pressure inside the reactor vessel increases, so does the differential pressure across the piston (from around 500 psi to over 1,200 psi for the cyan curve in the chart). Although the resistance to control rod insertion also rises as the pressure inside the reactor vessel increases, the increase in differential pressure is the larger factor. This is due to the geometry of the piston itself - one side is significantly bigger than the other side. Consequently, equal force applied to

October 3, 2011 Page 4 of 4 both sides of the piston causes it to insert because the force on one side acts upon a larger area and therefore has a greater effect.

In any case, emergency insertions of control rods occur when accumulator pressure and reactor pressure are both available. These two sources produce the red dashed curve in the chart. It shows that at low reactor vessel pressures, the accumulator is the dominant factor affecting control rod insertion times. As the pressure inside the reactor vessel rises, the accumulators influence diminishes until the reactor pressure becomes the dominant factor. The slowest control rod insertion time occurs at a reactor vessel pressure of around 800 psi, when neither source is dominant.

Returning to the safety problem GE Hitachi reported to the NRC on September 27, 2011, the problem is pronounced when the reactor vessel pressure is less than 900 psi.. This corresponds to the peak portion of the red dashed curve, and is when rubbing between control rods and fuel channels poses the greatest risk.

At higher reactor vessel pressures and at very low reactor vessel pressures, the control rod insertion is faster. The high differential pressures producing this speed are also most likely to overcome resistance caused by friction between the control rod and the fuel channels. But in that mid-range region where insertion times increase, friction can further slow or even stop control rod insertion.

Failure of control rods to fully insert can have disastrous consequences. The array of emergency systems that provide makeup water to the reactor vessel when a pipe breaks or other emergency occurs, as well as the massive concrete containment structures, are designed to mitigate an emergency once the control rods shut down the nuclear chain reaction. If the control rods fail to do so and the nuclear engine continues running, it may produce more energy than the emergency makeup and containment systems can handle.

When the problem first surfaced last fall, GE Hitachi recommended that plant owners increase testing of the control rod insertion times and apply a margin to the test results to account for rubbing between the control rod and fuel channels. With the recently expanded dimension of the problem (i.e., the fact that certain conditions can exacerbate the rubbing), GE Hitachi is recommending that plant owners increase the margin applied to the test results.

A better solution would be to design and use fuel bundles and control rods that did not rub against one another. Bump and grind is more suited for the dance floor than in the reactor core.

Prepared by:

David Lochbaum Director, Nuclear Safety Project