Update: Benchmarking Boiler Tube Failures

Boiler tube failures continue to be the leading cause of downtime for steam power plants. Is your boiler tube failure reduction program showing improvement when compared to programs at peer plants? The EUCG’s recent update of its boiler tube failure study can help you answer that question. The full study is available only to members, but this POWER exclusive presents many of the key results, which could help you improve the operation of your plant.

The call often comes on a weekend or in the middle of the night and goes something like this: "Hello, this is the Operations Shift Supervisor from the plant calling. I hate to tell you this, but I think we have a tube leak on the unit…. Yes, it’s blowing pretty good…. No, I don’t think we can hang on until the weekend or even tomorrow; we’re starting to come off-line now…. OK, see you in a few hours."

The bad news hasn’t changed: Boiler tube failures are still the leading cause of forced outages in coal-fired boilers.

Those who work in power plants understand that boiler tube leaks can ruin a weekend or holiday for the plant staff. Your unit can be running along at full load, seemingly problem-free, and then an operator making a round hears the telltale roaring noise, or sees water in an economizer hopper.

Once the decision has been made to take a forced outage, the unit is taken off-line, and we wait for the furnace to cool before workers can enter. What usually follows are workers finding the leak(s), erecting scaffolding, making the boiler tube repair, inspecting the repair, and performing a hydro before the unit can return to service several days later. Sometimes we piggyback a furnace cleaning when the unit is off-line. The operation and maintenance (O&M) costs plus lost power costs can easily reach a million dollars a day.

The bad news hasn’t changed: Boiler tube failures are still the leading cause of forced outages in coal-fired boilers. The Electric Power Research Institute (EPRI) agrees with that assessment in its Boiler Tube Failure Reduction/Cycle Chemistry improvement program description: "Boiler tube failures have been the number one availability problem for utilities with fossil plants for as long as reliable statistics have been kept in individual utilities and by nations."

The good news is that, despite aging boilers, boiler tube failures are not occurring more frequently in general, and for some categories of boilers, they have actually decreased in frequency over recent years. Lessons learned from EUCG benchmarking studies may well be part of the reason for this improvement.

Benchmark Your Leaks

In a two-part series in 2005, POWER reported on a 2004 EUCG benchmarking survey on boiler tube failures (see box). That article reviewed the frequency and duration of boiler tube leak outages across a range of unit sizes and furnace designs.

In the Spring 2009 EUCG Fossil Productivity Committee workshop, that survey was updated by authors Daryl Von Behren of Kansas City Power & Light and Dave Cook of Constellation, with data compilation by Joyce Cook Jackson of Cook-Jackson Inc. This most recent survey focused on data from the years 2006 through 2008.

Seventeen utilities participated, representing 167 units. Those units ranged in size from very small (8 MW) to very large (1,264 MW), with a median of 205 MW. By age they ranged from two to 68 years old, with a median of 48 years. Boilers in this survey group were mostly of wall-fired design (46%) or were tangential-fired (41%); 10% had cyclone furnaces, and 3% were fluidized bed or other. The original boiler equipment manufacturers represented in this survey were dominated by Babcock & Wilcox and Alstom/CE. Only 12% of the units were supercritical; the remaining were of the more conventional subcritical steam pressure design.

Despite the age of these units, their capacity factors remain fairly high. For smaller units, 300 MW or less (100 units), the 2008 data show that 66% of these units have 70% or higher capacity factors. For units over 300 MW (53 units), 79% reported a 70% or higher capacity factor. In the earlier survey, the 2003 data showed that 71% of units over 300 MW reported capacity factors greater that 70% and only 56% of units less than 300 MW reported that their capacity factor for the year exceeded 70%. The most recent benchmarking study seems to reflect the fact that smaller coal-fired plants tend to operate more frequently these days and that larger plants continue to improve their reliability, regardless of age.

One factor that doesn’t seem to change over time is that the boiler design of the unit plays a significant role in the number of tube leaks experienced (Table 1). Wall-fired units fared the best, with an average of 7.1 tube leaks per year. Tangential design units performed slightly worse over the study’s time frame, with an average of 7.8 tube leaks per year; however, the average for this class was 8.7, driven by 24% of the units experiencing a high number of tube leaks — greater than 10 leaks per year. Cyclone boilers had varied performance, with an average of 18.5 tube leaks, and 24% of these units were plagued with more than 30 tube leaks per year.

Table 1.    Survey respondents’ average tube leaks by boiler design type. Data are averaged over the period 2006 through 2008. Source: EUCG

Which Units Leak the Most, and Where

It’s well known that the North American coal fleet is aging, and with advancing years come maintenance challenges. It’s surprising, therefore, that the data in this study seem to indicate that the oldest units, those more than 54 years old, have a distinct edge over their younger counterparts — those less than 34 years old — when it comes to boiler tube leaks (Figure 1).


1.    Young and stressed. Close examination of data from the oldest and youngest units in the survey sample reveals a perhaps surprising finding: Although older and younger units had somewhat similar numbers of leaks when you consider low incidents of leaks (fewer than six per year), when you look at the worst-performing units—those with the highest number of leaks (more than six)—older plants fared noticeably better, especially in the most recent survey year. The total sample size was 164 units; 39 units were less than 34 years old (25th percentile) and 35 units (75th percentile) were more than 54 years old. Source: EUCG

Looking at units that fared best — with four or fewer tube leaks per year over the study time frame — 83% of the oldest units met that mark, whereas only 64% of the younger units did so. One possible explanation for this result is that the older units that are still operating may have had tube panel or section replacements for areas of the boiler with high failure rates, whereas the younger units are just approaching the time when those panel or section replacements may be necessary. Perhaps the boiler design margins, predating the use of modern computerized design tools, were more robust.

The older units also seem to be experiencing fewer leaks over time in recent years, presumably due to improved boiler tube failure reduction (BTFR) practices. In 2006, 77% of units had four or fewer tube leaks per year; that number jumped to 89% in 2008.

Although more units that are 300 MW and smaller reported zero leaks than did the cohort of units larger than 300 MW, it’s important to understand that nearly twice as many small units participated in the survey as did large ones (Figure 2). The lower capacity factors reported for the smaller units may indicate that this size class of plant is cycled more often than the larger, baseloaded plants.


2.    Number of leaks. Although this chart shows more leaks for smaller than for larger units, it’s important to know that there were more small units in the survey. The total sample size was 164 units; 62 units were greater than 300 MW in capacity. Source: EUCG

The area of the boiler most plagued by tube leaks continues to be the waterwalls, followed by the convection and pendant superheat sections (Figure 3). In addition to leak location, the data were analyzed to determine the most likely mechanisms causing the tube leaks. The answer: fly ash and sootblower erosion are the two most likely causes for failures. Fatigue failures and welding flaws were also significant failure modes (Figure 4).

3.    Location, location, location. Survey respondents reported the locations where most of their boiler tube leaks occurred. The data show the number of units for which a particular problem was rated either the first-, second-, third-, or fourth-leading cause of leaks. For example, for more than 70 units, waterwall leaks were the most frequent source of leaks, and for just under 40 units, waterwall leaks were the second-most-frequent source of leaks. Source: EUCG

4.    Why tubes leak. This chart reflects data reported from 164 units. The survey asked respondents to identify the top four failure mechanisms for each unit, with 1st representing the failure mechanism with the greatest number of incidents. On the chart, the bars indicate the number of units that identified that failure mechanism as first through fourth in tube leak frequency. For example, 22 units indicated that the most tube leaks occurred due to fly ash erosion; 20 units said that fly ash erosion caused the second-highest number of leaks; and so on. Source: EUCG

Good O&M Lowers Leak Numbers

Most of the sub-groupings of units showed either steady performance or improvement in minimizing tube leaks over the three-year span of this study. The reasons for this performance can be inferred from some of the practices that surveyed utilities employ to prevent boiler tube failures.

Most, 79%, have a formal BTFR program in place, and 67% responded that they use some sort of outside tools to track and plan a strategy to prevent tube failures. Tools used include Aware (ATI), UDC Tracker, and EPRI Boiler Workstation.

Control of water chemistry also appears to have played a role in minimizing tube leaks, with 90% of the units reporting that their water treatment histories were either good or had predominantly nonsevere variances. (The more-detailed report received by participating utilities includes more data on specific boiler and feedwater chemical treatment programs and chemical cleaning criteria.)

In addition to preventing boiler tube failures, tight control of water chemistry parameters can improve heat transfer and, therefore, unit heat rate. Even with good control, however, boiler tube internal chemical cleanings are sometimes necessary in the life of a boiler. Roughly two-thirds of utilities, 62%, determine the need for cleaning by sampling tubes and measuring the deposit density on the tubes’ internal surface, 23% clean at a fixed time interval, and the remainder clean based on both factors or when budget is available. For those that clean at a fixed interval, the most frequent intervals named were three years or five years, most likely coinciding with boiler outages.

Outage managers know that chemical cleanings are costly and come with other risks. However, by removing deposits that can cause tube metal overheating and under-deposit corrosion, chemical cleanings can help minimize the incidents of future boiler tube leaks.

Tube Repair Options

When tube leaks do happen, a new section of tube, called a dutchman, replacing a section of failed tube is often the repair of choice, although it can take longer than other methods, and access to the failed area is sometimes problematic. Some areas of some boilers can only be accessed by openings so small that only a few, very thin welders can access the areas. In other areas, welders have to cut through nonfailed tubes to get to the failed ones and then weld repair both the failed and cut tubes.

Even if the area can be accessed, welding often must be done in awkward spots, requiring great skill and concentration to make a successful weld. Window welds sometimes allow a tube leak to be accessed from the exterior of the boiler, minimizing scaffolding time and cost, but they are a more technically difficult weld to perform and are more prone to failure. For those reasons, utilities responding to the survey are almost evenly split on whether they allow window welds as an accepted repair practice: 51% do; 49% do not.

Pad welds, on the other hand, are usually a faster repair than either dutchman or window welds, but they are prone to future failure and are often considered a temporary repair until they are replaced during a future outage. Ninety percent of the survey respondents allow pad welds as part of their repair programs, while 10% do not.

Outage Planning

Perhaps the most difficult decision of all is selecting the best time to make boiler tube repairs. The decision must balance the importance of making expeditious repairs with the costs of bringing a unit off-line for several days.

As shown in Figure 5, in the majority of cases, plant leaders will wait until the following weekend to take a unit with a boiler tube leak off for repair, because weekend power prices are generally lower than weekday prices. Although this strategy minimizes replacement power costs, waiting can increase the probability of secondary tube damage, so some utilities take a unit off-line when a tube leak increases in flow rate; others prefer to take an affected unit off-line immediately.

5.    Wait for the right time. Survey respondents favor waiting for the next weekend to remove a boiler from service to repair a boiler tube leak. Source: EUCG

Finally, the survey looked at the frequency and duration of planned boiler outages. The survey data were parsed into major boiler outages and short boiler outages. For major outages, the most frequently selected interval was every two years (37% of respondents), followed by every three years (24%) and intervals of four or more years (21%). For short boiler outages, the majority (59%) perform outages every year (Figure 6).

6.    Down for repairs. Survey respondents indicated the frequency with which their units dealt with longer and shorter planned outages. Source: EUCG

The duration of an outage is often determined by a number of factors, including the scope of repairs, and the scope is sometimes influenced by the interval between outages. When major boiler outages are performed every two to three years, the predominant duration for the outage is 30 to 39 days.

Well-planned and -executed outages are just one contributor to reducing the number of unplanned outages caused by boiler tube leaks. Other factors noted by this study include organized BTFR programs, cycle chemistry improvement programs, and attention to weld-repair quality.

Daryl Von Behren of Kansas City Power & Light, one of the authors of the EUCG BTFR study, noted that best-practice companies strictly adhere to their BTFR programs. Two common components of a successful program are the identification and elimination of root causes and the replacement of damaged tubing through an aggressive boiler pressure part replacement strategy.

By adopting these programs and methods, some utilities have reduced the frequency of those dreaded phone calls from plant operating personnel reporting tube leaks, as well as the cost of maintaining boilers.

Special thanks to Jim Patrick, former EUCG Fossil Productivity Committee president.

—Sharon Pfeuffer ([email protected]) is plant director for DTE Energy’s River Rouge, Trenton Channel, and Conners Creek Power Plants, Peakers Fleet, and Shops as well as a participant in and contributor to EUCG Fossil Productivity Committee workshops.