Steam generator replacement overview
By Harry Chernoff, Science Applications International Corp., and Kenneth Chuck Wade, Energy Information Administration
Since nuclear power began to be widely used for commercial purposes in the 1960s, unit operators have experienced a variety of problems with major components. Although many of the problems have diminished considerably, those associated with pressurized water reactor (PWR) steam generators (Figure 1) persist. Steam generator problems rank second, behind refueling outages, as the most significant contributor to lost electricity generation.
As of Dec. 31, 1995, 38 steam generators had been replaced in 13 of the 72 operating PWRs, and three units had been shut down prematurely, due primarily (or partially) to degradation of their steam generators: Portland General Electric`s Trojan unit, located in Prescott, Ore., in 1992; Southern California Edison`s San Onofre 1, located in San Clemente, Calif., in 1992; and Sacramento Municipal Utility District`s Rancho Seco unit in 1989.
In the coming years, operators of PWRs in the United States with degraded steam generators will have to decide whether to make annual repairs (with eventual derating likely), replace the generators or shut the plants down prematurely. To understand the issues and decisions utility managers face, this article examines problems encountered at steam generators over the past few decades and identifies some of the remedies that utility operators and the nuclear community have employed, including operational changes, maintenance, repairs and steam generator replacement.
Failure and degradation
In the 1970`s, tube wastage was the earliest problem many utilities reported at a number of units. The Electric Power Research Institute (EPRI) formed the first of two Steam Generator Owner Groups to address the wastage problem and an emerging problem: widespread tube denting. By the end of 1982, the cause and remedies for denting were much better understood, and that problem had dramatically decreased. By 1979, stress corrosion cracking and apparent fatigue cracking had begun to be reported at a number of operating units.
The issues associated with steam generator degradation have had a significant impact on nuclear power plant operation. As a result, utilities with degrading steam generators must make a tradeoff between either continued operation with high operation and maintenance costs, high worker radiation exposures, increased risks of forced outage from tube ruptures, derating the plant or replacement.
Common failure modes
Currently, the most common form of failure is intergranular attack/stress-corrosion cracking. This form of failure now accounts for 60 to 80 percent of all tube defects requiring plugging. Fretting and pitting combine to account for another 15 to 20 percent of all tube defects. The remaining failures are attributed to mechanical damage, wastage, denting and fatigue cracking.
Initially, the problems were thought to be isolated incidents resulting mainly from defects in manufacturing, poor operations, poor water chemistry and other factors. Over time, however, a pattern of failures began to emerge, suggesting common factors and common failure modes (see "Common failure modes").
The once-through steam generators have experienced fewer problems than the U-bend design. The reason once-through steam generators have been able to control the degradation phenomenon is that Babcock & Wilcox (B&W), the manufacturer of the once-through design, incorporated flow openings around the tube support plate (a known corrosion area) and fabricated its tubes differently. Instead of Inconel 600 mill-annealed, B&W tubes were Inconel 600 sensitized. Currently, the design and technological improvements have been incorporated in the U-bend design steam generators.
Maintenance and repair
Several strategies have been developed to minimize degradation problems and prolong steam generator life. Tube defects and failures occur for various reasons, particularly when the secondary-loop water contains impurities or particles that lodge in crevices or create sludge or when the water is excessively basic/acidic or excessively oxidizing/reducing. Changes in secondary water chemistry over the years have included substituting all-volatile treatment for phosphate treatment to reduce sludge. Improved water chemistry has helped somewhat but has not arrested the widespread degradation of Inconel 600 mill-annealed tubes.
Accumulations of sludge and corrosion products on the outside of the tubes, especially at the connections with the tubesheet and the tube support plates, are responsible for several types of tube degradation, including stress corrosion cracking and intergranular attack. Mechanical cleaning methods, such as water lancing, are used to reduce deposits and slow tube degradation. Steam generators have been cleaned at six U.S. units (Table 1).
As of December 1993, approximately 38,000 tubes, or 0.9 percent of all the tubes in operating steam generators in the United States, have been plugged. In general, 15 to 20 percent of the tubes may be plugged before replacement or derating is required. Excessive steam generator plugging hinders coolant flow, which may require significant power reduction. In general, plugging is an operator`s initial response to degrading tubes. Steam generators are designed to have an excess number of tubes; therefore, tubes are generally plugged when they degrade. Once a number of tubes degrade, the operator may decide to sleeve including those that were initially plugged. New steam generators have an even higher excess of tubes. From 1987 to 1991, in units throughout the world, the location of the defects requiring plugging varied considerably (Table 2).
Sleeving is used only for steam generator tubes with cracks penetrating no more than 40 per cent of the tube wall; more serious cracking requires the tube to be plugged. A short tube, or "sleeve," is inserted into the base tube to bridge the degraded area. The sleeve is then welded inside the tube to isolate the degraded section of the tube. The sleeve effectively seals the leak from the secondary-loop water. This technique is usually limited to the portion of the tubes near the tubesheet. Although sleeving is more expensive than plugging, and the water flowing through the tube is slightly affected, the tube remains in service. In the United States, sleeving has been done in almost two dozen operating PWRs.
Leaks or degraded condenser tubes can contaminate the secondary-loop water that circulates through the steam generator and lead to steam generator tube degradation. Improved condenser materials (e.g., titanium tubes), better leak detection devices and better water chemistry minimize condenser-related problems and associated steam generator problems.
Even with condenser improvements, water chemistry improvements, inspection and cleaning programs, operational changes and other actions, problems at steam generators are continuing. Recently, there have been reports of circumferential cracks near the tube sheet that went undetected in standard inspections, but were found using more sophisticated tube inspection equipment. Although circumferential cracks are not a new phenomenon, new tube inspection devices have shown that the cracks may be more numerous than initially thought. EPRI reports that 28 plants have reported finding circumferential cracks near the top of the tube sheet since 1987.
In 1994, circumferential cracks were discovered in more than half the tubes at the top of the tube sheet in the steam generators at the Maine Yankee nuclear plant. The utility, Maine Yankee Atomic Power Co., is currently sleeving all 17,109 tubes in the three-loop reactor. The repair is estimated to cost $64 million, not including the cost of replacement power. Due to the industry`s latest findings, the Nuclear Regulatory Commission is asking each PWR operator to prove that, like Maine Yankee, it is adequately inspecting its steam generators for these cracks. Working to address these problems are the individual utilities and vendors and several industry groups such as the Steam Gen erator Replacement Group, the EPRI Steam Generator Strategic Management Project (successor to the EPRI Steam Generator Owners Group), the Westinghouse Owners Group and the Combustion Engineering Owners Group.
Replacement of a steam generator is an economic decision. A steam generator with excessive tube degradation creates extra costs for reasons such as tube inspections and leakage monitoring, maintenance and repair (e.g., plugging and sleeving), and water chemistry control. Condenser inspection, maintenance and monitoring, occupational radiation exposure, power derating due to plugging and the potential for forced outages due to tube leaks or ruptures complete the list of reasons why it may be economically feasible to replace steam generators.
An analysis of one case showed that, compared to continuing with the existing equipment, installing a new steam generator would reduce annual steam generator repair costs by $3.4 million. As maintenance costs increase and derating becomes more likely, the economics of steam generator replacement becomes more attractive.
The cost of a steam generator is $12 million to $20 million. The cost to replace a steam generator is substantially more. Complete replacement at a three-loop PWR in the United States over the past two years cost between $125 million and $153 million, or approximately $139/kW to $170/kW for a typical 900-MWe unit. South Carolina Electric & Gas Co. (SCE&G) spent an estimated $153 million to replace three steam generators at the 885-MWe Summer unit. A total of 13 U.S. units have replaced steam generators, all of which are of the U-bend design.
The most recent replacement, at SCE&G`s Summer unit, took 38 days from the time the reactor coolant system piping was severed until the secondary-side piping was pressurized to 1,500 pounds for testing. The world record for a steam generator replacement, set in France in 1994 at Gravelines Unit 1, is 37 days.
When a utility decides to replace its steam generators, it must go through extensive planning efforts that include examining the extent of damage to the steam generators, estimating the length of time required to replace the steam generators, deciding whether a partial or complete steam generator replacement is needed, and determining the cost associated with replacement, including weighing the cost of purchasing replacement power during the scheduled steam generator outage.
Nine U.S. units are planning to replace steam generators, according to formal announcements or reports concerning placement of steam generator orders. Florida Power & Light expects to spend close to $170 million, excluding replacement power costs, to replace two steam generators at St. Lucie 1 in 1997. Duke Power expects to spend $437 million, excluding replacement power costs, to replace steam generators at three four-loop units (McGuire 1 and 2 and Catawba 1) between 1995 and 1997.
In general, there are four benefits to replacing a nuclear steam generator. The first benefit is avoiding, or at least substantially reducing, the problems associated with tube degradation outlined earlier. The savings from avoiding a forced outage due to a tube rupture are difficult to quantify but could certainly amount to tens of millions of dollars.
The second benefit is that the increased heat transfer surface may allow an uprating in the electric output of the unit. At the Summer station, SCE&G plans an uprate of approximately 50 MWe (approximately 5 percent) following the 1996 refueling outage. The increased number of tubes and the increased heat transfer surface also expand the margin for future plugging, if necessary. The value of an uprating depends on the remaining expected life of the generating unit. For example, the value of a 50-MWe uprate of a good-performance plant, that has 20 years of remaining life, is probably worth tens of millions of dollars.
The third benefit is reduced occupational exposure after replacement. Prolonged operation with degraded steam generators will ultimately increase radiation exposure and extend refueling outages due to the increasing need for extensive tube inspection and repair.
The fourth benefit is deferred decommissioning. Premature shutdown creates two major decommissioning problems. First, the decommissioning trust will not have had the time to accumulate the full amount needed to pay for decommissioning. If a plant is shut down 10 years prematurely, the decommissioning trust is likely to lack at least three-quarters of its decommissioning total. Second, decommissioning requires extensive planning many years in advance of actual decommissioning activity. Planning includes onsite activities, waste disposal preparations, licensing, settlement of state regulatory issues, replacement power planning and the like. Deferring these activities and conducting the planning on a non-emergency basis has significant value to a utility.
Finally, owners considering steam generator replacement will find the job easier to justify if they are also considering license renewal, as a long license term provides a lower per-kilowatthour cost for the replacement.
Units with original steam generators incorporating the Inconel 600 mill-annealed alloy tubing are almost certain to face degradation problems. In 1993, the NRC found "no end in sight" to steam generator tube cracking problems at plants operating with original steam generators. There are 23 U.S. units that could be candidates for steam generator replacement in the future. These 23 units are those units whose percentage of plugged tubes range from 2 to 16 percent, suggesting the unit has some degree of degradation in its steam generators. Utilities are continuing to make necessary adjustments to their systems to prolong steam generator life. For example, the Arizona Public Service Co., the operator of the Palo Verde nuclear plant in Wintersburg, Arizona, has made several adjustments to its plant and believes that its steam generator may, in fact, last the full 40-year license period.
In the final analysis, utility managers must decide whether to maintain existing steam generators or replace them. This is a difficult decision, one that must be based on technical and cost analyses and license terms. Overall, the prospect for continued operation of PWR`s in the United States is good, but the prospect for long-term operation of original steam generators with Inconel 600 mill-annealed tubing is poor. The only exceptions are likely to be those reactors that recently began operation, where the lessons learned in such areas as water chemistry, tubing material, tube support plate material, and tube support plate design and attachment were incorporated from the very beginning of unit operation. z
Harry Chernoff is a senior economist at Science Applications International Corp. in McClean, Va., with 18 years experience in nuclear power economics.
Kenneth Chuck Wade has worked at the Department of Energy`s Energy Information Administration since 1988. He has worked primarily in U.S. and foreign nuclear generating capacity projections. In addition, he has co-authored an article on the recent rise in U.S. nuclear power performance.
Nuclear power at a glance
Nuclear power is the second largest source for electricity generation in the United States, accounting for more than one-fifth of total utility-generated electricity in 1994. Currently, 109 nuclear units are licensed in the United States, representing a total capacity of 99 GWe. Of the 109 units, 72 are pressurized light-water reactors and 37 are boiling-water reactors.
Common failure modes
The physical factors most often responsible for failures, and the typical corrective actions, are as follows:
The most common factor in tube defects has been the tube alloy most widely used in original steam generators both in the United States and throughout the world, Inconel 600 mill-annealed, a thin nickel alloy material that has proven susceptible to many forms of cracking, pitting, denting and other types of degradation.
Tube sheet design and alloys
The tube sheet separates the primary-loop water from the secondary-loop water. Both the tube sheet connection and the exterior of the tubes at the connection tend to accumulate sludge, crack from vibration and show excessive fatigue cracking.
Tube support plate designs and alloys
Tube support connections tend to accumulate corrosive sludge, crack and fret.
During manufacture and operation, small-radius U-bends are subject to greater stress than large-radius U-bends or the unbent portion of the tubes.
This article was excerpted from the August 1995 issue of the EIA`s Electric Power Monthly. For a copy of the full report with references contact: Kenneth Chuck Wade, EIA Analyst, EIA, Forrestal Bldg., Rm. 1F-048, Washington, D.C. 20585 or phone (202) 586-8800