By Stuart Cain, Ph.D. and Ludwig Haber, Ph.D., Alden Laboratories, Inc.
As nuclear engineers well understand, the water in nuclear reactor containment vessels moderates the speed of the neutrons released by the uranium in order to sustain the chain reaction, and carries the heat away from the reactor core, ultimately generating the steam to drive the turbines. Failure of the water system can have dire consequences; hence, nuclear power plants are designed with an emergency core cooling system (ECCS) to provide the cooling needed for extended time required to cool the reactor fuel. There is one potential issue, however, which can cause the ECCS to fail: a loss-of-coolant accident (LOCA) coupled with debris from the LOCA clogging the ECCS suction screens.
For the past two decades, Alden Research Laboratory, Inc. has been working with the U.S. Nuclear Regulatory Commission and the nuclear power industry to design and validate equipment that will eliminate the risk of debris-caused inadequate core cooling after a LOCA.
Clogging the Cooling System
The pipes within the containment vessel of a nuclear reactor contain water at temperatures up to 600 F and pressure of 2,300 psi. (Note that significant differences exist in the design and operating parameters of boiling water reactors (BWRs) and pressurized water reactors (PWRs). This article deals with general issues that affect both, unless otherwise specified.) If one of the pipes ruptures, the explosive force of the water/steam produces pressure waves that blast the insulation off the piping and can even strip the paint from the walls.
After the initial blast, the water collects in the sump. When the water reaches a predetermined level in the sump, ECCS pumps are aligned to bring that water back into the cooling system, with some of it diverted up to the spray nozzles used to cool off the containment. The water will continue to flow out of the ruptured pipe and collect back in the sump so the pumps keep recirculating the cooling water. All these systems are designed to keep the core cooled and the reaction controlled while plant personnel take the steps necessary to assure the safety of the public.
The problem is that it is not just water that collects in the sump. As mentioned, the high pressure water and steam from the ruptured pipe will also destroy insulation, other components and paint from structures inside the containment. Some of this material will be transported to the sump. As the pumps start, the material in the water starts moving to the pump inlet screens. This material can be small particulates and individual fibers, all the way up to large pieces that fell down from the piping. This material gets pulled toward the strainer screens and starts interacting with them.
In the beginning, when there is no fiberglass on the screens, the particulates pass right through the holes. But, as fibers start to accumulate, particulate gets trapped between the fibers. A debris bed builds up and starts acting as a filter, capturing more and more of the material in the water, increasing head-loss, and possibly causing cavitation to develop in the pump, impairing its ability to continue recirculating the cooling water.
The NRC first started looking at this issue in relation to BWRs in the late 1980s with work continuing through the following decade. (See NRC Bulletin (NRCB) 95-02, “Unexpected Clogging of a Residual Heat Removal Pump Strainer While Operating in Suppression Pool Cooling Mode,” Oct. 17, 1995, and NRCB 96-03, “Potential Plugging of Emergency Core Cooling Suction Strainers by Debris in Boiling-Water Reactors,” May 6, 1996) During those years, Alden performed research for the NRC and individual plants, testing the performance of different screens with actual debris. Most BWR operators installed larger strainers and established new procedures to keep the suppression pool free of debris. In October 2001, the NRC concluded that all BWR plants had strainers with sufficient margin to permit the ECCS to operate despite the debris buildup during a LOCA.
Once the issue with BWRs was progressing toward a solution, the NRC turned its attention to PWRs. In 1996, the NRC issued Generic Safety Issue (GSI) 191, “Assessment of Debris Accumulation on PWR Sump Performance.” This time, Alden worked primarily with industry, partnering with Areva NP Inc. and strainer manufacturer Performance Contracting Inc. (PCI). Areva did the initial debris generation modeling to determine what types and amounts of debris would be liberated in a LOCA and Alden used a computational fluid dynamics (CFD) based model together with empirical settling and transport data to determine how much of that debris would wind up in the vicinity of the sump screens.
But even the most advanced CFD modeling is not sufficient to predict the interaction between debris and fluid near the strainer, particularly when the concentration of debris is relatively high and the forms and shapes of the debris are difficult to describe mathematically. A physical test of a section of the sump screen array must be conducted to measure the actual head loss as the debris builds up.
This is where Alden’s expertise in physical modeling and the earlier work with the NRC was taken advantage of. Taking the data from the generation and transport models showing how much debris was likely to be generated in a LOCA, Alden created a full-scale model of a single strainer module (generally 300 to 800 gpm) and conducted tests using different types of fibrous and particulate materials to determine the debris build-up on the screens and to measure the resulting head loss across the suction strainer. Initially, generic testing was performed to validate the generic design of the PCI strainers, which employ PCI’s Sure Flow technology so that each square foot of screen draws the same amount of flow. This was followed up by testing individual strainer configurations for individual power plants to validate the plant’s solution for the NRC. (See sidebar, An Example of Strainer Modeling)
Back to the Beginning
Based on the GSI-191 research and testing, PWR operators have made significant changes to their sump screens. They now typically have several thousand square feet of screen area, to allow the buildup of significant amounts of debris without impeding water flow. About 2/3 of all U.S. PWRs now have a plan in place that will satisfy the NRC requirements for closure of this issue, with the remainder expected to finish in the next few years. But, while that will signify compliance with GS-191, further research and testing is still needed. One matter that was discovered in the GS-191 testing, for example, is that chemical reactions between the debris and certain metals causes chemical precipitants that can build up in the strainers. The possibility of particulates and fibers not caught in the pump screens clogging the screens protecting the fuel rods also continues to be evaluated. Alden is assisting with both these issues.
The NRC is also working with the BWR Owners Group to see how the data uncovered about the PWRs affects BWRs. In October 2009, the NRC Executive Director for Operations, R.W. Borchardt, said “Currently, BWR ECCS operability is not in question and continued operation of BWRs does not pose an undue risk to public health and safety,” but that “the efforts to address GSI-191 have led to new insights that could be relevant to previous assessments of ECCS suction strainer performance in BWRs.” (SECY-09-0156 Status of Applying Lessons Learned from Pressurized Water Reactors to Emergency Core Cooling System Suction Strainer Performance for Boiling Water Reactors.)
By the end of this year, the NRC is expected to release at least preliminary guidance on how BWRs can use the information learned under GSI-191, which may set off a whole new series of tests. In the meantime, Alden is continuing to work with nuclear and other power plants on investigating areas such as pump performance, the clogging of smaller strainers, air entrapment and entrainment, the impact of hurricanes on intake structures and other issues that can affect safety or performance.
An Example of Strainer Modeling
In 2008 Alden tested strainer configurations for several licensees including Comanche Peak. The test was conducted using a flume that is about 45 ft. long and 10 ft. wide, with walls that are positioned to vary the flow velocity to simulate the conditions of a particular plant. Figure 1 shows the test strainer installation before test start. For the Comanche Peak test, the test flume walls were designed to carry a flow of 363 gpm at a peak velocity just above 0.5 ft./sec. As shown in Figure 1 the PCI strainer module is directly connected to the outlet piping. The strainer module for the test has 110 sq. ft. of screen, about 2.9 percent of the full screen size at Comanche Peak.
Prior to beginning the design basis head-loss testing, a number of transport test are conducted to determine whether or not various types of miscellaneous debris can be transported to the strainer. “Miscellaneous debris” usually consists of a variety of tags and labels but also other items that could be found inadvertently left in containment. For Comanche Peak more than a dozen different items were investigated. All of them were found to be unable to overcome the installed debris interceptor curb that surrounds the strainer installation. Debris interceptor curbs are effective at preventing heavier debris whose transport likely occurs by tumbling along the floor from gaining access to the strainer. Fine particulates and fine fiberglass pieces can stay suspended in the water column and represent the greatest threat to the strainer installation.
The water was brought up to the desired temperature (120 F) prior to the test and then kept at that temperature using immersion heaters. Debris was loaded into the flow stream, most of it at about 30 ft. upstream of the strainer. Initially particulates and fine fibrous debris were added to the flume, resulting in a head loss of 0.6 ft. The particulate and fiber debris was followed by the addition of a chemical precipitate surrogate – aluminum oxyhydroxide. The head loss only increased marginally and when the head loss leveled off, the test was terminated. Draining down the water after the test reveals the fine composite coating of fiber and particulate responsible for the measured head-loss. Drain-down also showed that some particulate debris was also partially retained by the debris interceptor. The test successfully demonstrated the operability of the Comanche Peak strainer system under the postulated Post-LOCA conditions.
Nuclear Power International Issue Archives
View Power Generation Articles on PennEnergy.com