By Brad Buecker, Contributing Editor

Due to the uncertainty over future carbon dioxide emissions regulations, many power companies are turning to combined-cycle power technology for future generation. Of course, the combustion turbine is the key component of a combined-cycle plant, but so are the heat recovery steam generator (HRSG) and steam turbine. Minimizing corrosion in HRSGs can sometimes be a daunting task, particularly if the plant cycles frequently. This article examines a number of the most important issues regarding corrosion mechanisms and prevention.

HRSG Design Issues

One factor that influences some of the corrosion issues in HRSGs is the steam generator design. Most HRSGs are of the horizontal gas passage configuration with two or three drum-type pressure circuits. A common three-pressure HRSG outline is shown in Figure 1.

The water/steam flow path is from the low-pressure circuit to the intermediate- and high-pressure circuits, with attendant economizer, superheater and reheater loops. The important concept for this discussion is that the HRSG contains many multiple-tube networks with short-radius elbows at header inlets and outlets. Another important factor with regard to flow-accelerated corrosion, is that most HRSGs generally do not have any feedwater heaters between the condenser and the low-pressure (LP) drum, with the common exception of a deaerator.. Temperature is a critical influence towards flow-accelerated-corrosion (FAC), identified by the Electric Power Research Institute (EPRI) as the most prevalent corrosion mechanism in HRSGs1.

Flow Accelerated Corrosion

I wrote on FAC in Power Engineering magazine2 several years ago, after an FAC-induced failure caused two fatalities in a coal-fired unit that was a sister at the plant where I was working at the time. In short, the condensate/feedwater treatment program that was once popular for steam generators of all kinds, all-volatile reducing AVT(R), is now known to cause dissolution of the protective magnetite (Fe3O4) layer at a temperature range and chemical conditions common to the feedwater network of conventional boilers and the LP circuits of HRSGs. Areas particularly susceptible are flow disturbances such as elbows or other fittings. The graph in Figure 2 shows the potential magnetite dissolution rates in solutions that have been treated with an oxygen scavenger to remove all oxygen and which establishes a reducing environment.

Note that pH in high-purity water is a direct function of the ammonia concentration. It is the lower pH, at low ammonia concentrations in a reducing environment, which is responsible for magnetite dissolution. This explains why corrosion can be much higher at an NH3 concentration of 0.1 ppm than in any other case. The ammonia does not attack the magnetite directly.

The effects of this phenomenon, known as single-phase FAC, are shown in Figure 3. FAC continually erodes tube or pipe wall structure and eventually reduces pipe strength to the point of sudden failure.

Figure 3 FEEDWATER PIPE THINNED BY FAC.

Single-phase FAC has been a particular problem in HRSG waterwall tubes that have many tight-radius elbows. The low-pressure circuits of HRSGs often operate near the temperature of highest corrosion potential shown in Figure 2, which exacerbates FAC mechanism.

FAC Prevention

The preferred method, from a chemistry standpoint, to address single-phase FAC issues in HRSGs is to implement a program developed by EPRI known as all-volatile treatment (oxidizing), or its acronym AVT(O). The concept is that AVT(O) establishes a different protective oxide layer (ferric oxide hydrate, FeOOH) on evaporator tubes and other components. FeOOH is not only tougher than magnetite, but the oxidizing environment does not cause iron dissolution.

AVT(O) is what might be called a “natural” program. Where condenser air in-leakage is minor and condensate dissolved oxygen levels stay at or below 10 ppb, the FeOOH protective layer forms naturally. Ammonia feed to the condensate is used to maintain LP economizer inlet pH within a range of 9.2 to 9.6. For the program to work properly and not influence corrosion, the feedwater cation conductivity should remain at or below 0.2 mS/cm. Excursions in dissolved oxygen concentration and cation conductivity, particularly the former, indicate excess air in-leakage within the condenser. Increased air in-leakage also introduces excess carbon dioxide, which influences corrosion. Thus, for a unit on AVT(O) any air in-leakage difficulties that raise condensate dissolved oxygen concentrations significantly above 10 ppb should be investigated and corrected as quickly as possible. One point needs to be noted, however. Recent research suggests some areas in the condensate/feedwater system, such as boiler feed pump discharge nossles, may suffer from single-phase FAC even when AVT(O) is maintained within the current guidelines.3 Further data hopefully will be forthcoming with regard to this issue.

One issue that has become much better known regarding flow-accelerated corrosion is that of two-phase FAC. At a number of points in the steam generating system, zones of physical separation between water and steam will develop; deaerators in particular come to mind. With a program such as AVT(O), the oxygen will separate with the steam and leave a less protective fluid that contacts water-touched components. Particularly susceptible location are LP economizer and evaporator elbows, and, generally to a much lesser extent, IP elbows. Thus, FAC can occur when chemistry parameters are seemingly in acceptable ranges. Two-phase FAC is difficult to control chemically, but the issue can be addressed mechanically in the design phase by using 1.25-chrome steel in affected areas, particularly LP elbows.

The Rest of the HRSG

What about corrosion prevention in the intermediate- and high-pressure circuits? For the multi-pressure drum HRSG examined here, boiler water treatment for the IP and HP evaporators has largely evolved from the chemistry developed for conventional steam generators. The critical issue in these circuits is operation at a suitably alkaline pH with chemistry that can prevent hydrogen damage by even trace levels of chloride (Cl–) or sulfate (SO4-2).

For decades, some form of phosphate treatment has been most popular for steam generation chemistry control. The backbone of any program was feed of tri-sodium phosphate (Na3PO4) to generate an alkaline pH.

Na3PO4 + H2O ⇔ Na2HPO4 + NaOH

For many years, chemists would blend di-sodium phosphate (Na2HPO4) and sometimes even mono-sodium phosphate (Na2HPO4) to produce what were known as coordinated or congruent phosphate programs. Evidence has since shown that the propensity for phosphates to precipitate at high temperatures would set up direct corrosion of tube metal by these phosphate blends.

Phosphate treatment has evolved in almost all cases to EPRI’s phosphate continuum (PC), where tri-sodium phosphate is the only chemical used, except for perhaps a small amount of caustic (NaOH) at startup. The program is defined as either PC(L) for phosphate continuum (low) and PC(H) for phosphate continuum (high), where the approximate demarcation point between the two is a 3.0 ppm phosphate concentration. The lower control limit for PC(L) or (H) is a sodium-to-phosphate ratio of 3.0 to 1, with an upper control limit of 1 ppm free sodium hydroxide.

Many utility personnel have opted for the PC(L) program to minimize phosphate hideout. Careful control of the 1 ppm free hydroxide limit is important to prevent caustic gouging. However, the low phosphate/alkalinity concentrations provide only minor protection against contaminant ingress. Also, at these low concentrations, ammonia introduced to the feedwater for pH control in the LP evaporator circuit can influence pH readings, such that protection may not be adequate to prevent chloride-induced hydrogen damage. Conversely, excess hydroxide can cause direct caustic damage of evaporator tubes.

Alternatives to phosphate programs include caustic treatment or reliance simply on the AVT(O) utilized in the LP economizer inlet and evaporator circuit. Caustic treatment is a consideration for units where condenser leaks or other impurity ingress is problematic, as sulfates and particualrly chlorides can lead to corrosion and hydrogen damage of waterwall tubes. Caustic is also an alternative for units that suffer from phosphate carryover to superheater and reheater tubes.

Additional HRSG Corrosion Factors

Although combined-cycle units can now achieve 60 precent efficiency, they are still often used for cycling operations. Naturally, this imposes thermal stress on the HRSG. Simple fatigue will reduce component life. Furthermore, even with reasonable chemistry, corrosion fatigue often develops in various locations. In this mechanism, areas that are stressed become more vulnerable to attack by impurities in the water. These may include chloride, sulfate or even the hydroxide ions generated by chemical treatment programs.

Steam chemistry should not be overlooked. Research over the last few decades has shown that even minor carryover (>2 parts per billion) of chloride, sulfate and caustic can induce pitting, corrosion fatigue and stress corrosion cracking in turbine components. Of course, chemistry control in the evaporators can influence impurity transport to the turbine. However, other factors including equipment design and operation can cause difficulties. Excessive ramp rates can cause fluctuations in boiler drum levels that mechanically push impurities into the steam.

The upshot is that good water/steam sampling and monitoring is critical towards maintaining HRSG reliability and longevity. A second part of this article, to be published in a future issue, will cover sampling and monitoring guidelines.

References

1. Cycle Chemistry Guidelines for Combined Cycle/Heat Recovery Steam Generators (HRSGs). The Electric Power Research Institute, Palo Alto, California, 2006.

2. B. Buecker, “Flow Accelerated Corrosion: A Critical Issue Revisited”; Power Engineering, July 2007

3. Information presented at the ASME Research Committee on Power Pnant & Environmental Chemistry Committee spring meeting, April 2008, Austin, Texas.

Author: Brad Buecker is a contributing editor for Power Engineering. He serves as a Process Specialist for Kiewit Power Engineers of Lenexa, Kan. Buecker has over 30 years of experience in or affiliated with the power industry, much of it in chemistry, water treatment, air quality control, and results engineering positions with City Water, Light & Power (Springfield, Ill.) and Kansas City Power & Light Co.’s La Cygne, Kansas station. He has an A.A. in pre-engineering from Springfield College in Illinois and a B.S. in chemistry from Iowa State University. He has written many articles and three books for PennWell Publishing on steam generation topics. Buecker is a member of the ACS, AIChE, ASME and NACE.
Power Engineerng Issue Archives
View Power Generation Articles on PennEnergy.com