News, O&M, Retrofits & Upgrades, Water Treatment

Breakdown: Is flow-accelerated corrosion a concern in co-generation steam generators?

As previously reported in Power Engineering and other sources (but where regular reminders are extremely important), the phenomenon of flow-accelerated corrosion (FAC) has for years been a serious issue in both conventional power steam systems and heat recovery steam generators (HRSGs).  But what about lower-pressure (LP) systems at co-gen facilities and industrial plants?  This article provides observations from the West Co-Gen complex at the University of Illinois at Chicago. 

A Brief History of FAC Discovery and Its Consequences

Prior to the mid-1980s, researchers had firmly concluded that dissolved oxygen (D.O.) ingress into utility steam generator condensate/feedwater was a prime factor for carbon steel corrosion, and, unarguably, copper alloy corrosion, during normal operation.  (Oxygen can cause severe steel corrosion during boiler shutdowns.  Look for an article on this subject in a future issue of Power Engineering.)  Consequently, both mechanical deaeration and chemical methods, i.e., oxygen scavenger feed, were routinely employed to reduce feedwater D.O. concentrations to near zero.  Oxygen scavenger (the more precise term is reducing agent) treatment combined with ammonia or a neutralizing amine for pH control came to be known as all-volatile treatment reducing [AVT(R)].  The chemistry induces the formation of a gray-black magnetite (Fe3O4) layer on carbon steel surfaces. 

In 1986, the foundations of this chemistry received a severe jolt, as “On December 9 of that year, an elbow in the condensate system ruptured at the Surry Nuclear Power Station [near Rushmere, Virginia.]  The failure caused four fatalities and tens of millions of dollars in repair costs and lost revenues.” [1]  Researchers learned from that accident and others since then that the reducing environment produced by oxygen scavengers is the prime ingredient for single-phase FAC of carbon steel.  The attack occurs at flow disturbances such as elbows in feedwater piping and economizers; feedwater heater drains; locations downstream of valves and reducing fittings; attemperator piping; and, most notably for combined-cycle HRSGs, low-pressure evaporators, where the waterwall tubes, aka harps, have many short-radius elbows.  In fact, FAC is typically the leading on-line corrosion mechanism in HRSGs.

Figure 1.  Sudden failure induced by FAC.  Photo courtesy of Steve Shulder, Electric Power Research Institute (EPRI).

FAC induces gradual metal loss, which can lead to sudden failure when the pipe wall can no longer withstand the pressure.

Several other factors influence FAC, most notably pH and temperature.

Figure 2.  Influence of temperature and pH on iron dissolution from carbon steel.  The temperature aspect is why FAC is typically most prevalent in the feedwater system and LP evaporator of HRSGs.  Source:  Sturla, P., Proc., Fifth National Feedwater Conference, 1973, Prague, Czechoslovakia.

Based on the method of oxygenated treatment (OT) implemented in Europe in the early 1970s, EPRI developed a program to replace AVT(R) for utility drum units, known as all-volatile treatment oxidizing [AVT(O)].  If the condensate/feedwater system contains no copper alloys, which is true for virtually all HRSGs, then AVT(O) is recommended instead of AVT(R).With AVT(O) chemistry, oxygen scavenger feed is eliminated, and a small residual concentration [5 to 10 parts-per-billion (ppb)] of dissolved oxygen is maintained at the economizer inlet.  Ammonia or an ammonia/neutralizing amine blend is still utilized for pH control.  High-purity condensate (cation conductivity ≤0.2 µS/cm) is a requirement for AVT(O), but when proper conditions are established, magnetite becomes overlaid and interspersed with a tighter-bonding oxide, known variously as a-hematite or ferric oxide hydrate.  It is noticeable for its distinct red color.

Figure 3.  Properly passivated surfaces in a unit with AVT(O).  Photo courtesy of Dan Dixon, Lincoln Electric System.

The Potential for FAC in Lower-Pressure Steam Generators

Absolutely critical to the application of AVT(O) is high-purity makeup, which should have a cation conductivity ≤0.2 µS/cm.  Water of lesser quality will induce carbon steel oxygen attack, but this raises a question: what is the potential for FAC in lower-pressure steam generators, with typically less sophisticated makeup water treatment systems and/or significant condensate return from external heat exchangers?    These systems exist at thousands of plants around the world.  AVT(O) and OT cannot be employed because of unacceptable impurity concentrations in the boiler feedwater. 

Such questions arose for the UIC West steam plant.  Recent testing has uncovered some feedwater piping FAC, but in the broad scheme, the current treatment program has kept the corrosion to a minimum.  Unfortunately, at other plants where chemistry programs may be of lower priority, the probability of FAC may be greater.  The case study below will hopefully provide some insight on these issues.

Case Study: The University of Illinois at Chicago

The University of Illinois at Chicago, founded in the late 19th century, has an enrollment of nearly 30,000 students and operates the largest medical school in the country.  The extensive campus and medical facilities require substantial quantities of steam for heating and chilled water for cooling.  For many years, steam at the West plant was produced by six, 1950s vintage Babcock &Wilcox boilers. 

Figure 4.  UIC West steam plant.  This plant services the west side of the UIC campus and the two hospitals.  It also provides chilled water for the West campus and UIC hospital.

Three of the boilers have been retired in place, but the other three are still operational.  The plant also has a five-year-old Victory boiler (180,000 lbs./hr steaming rate maximum) that helps supply steam capacity and backup capabilities during the cold winter months.  Two decades ago, the plant added co-generation with the installation of three Solar gas turbines, each capable of seven-MW power production.  (Co-generation offers much higher net efficiencies than standalone steam production and is becoming increasingly popular for many applications. [2])  The turbine exhaust gases power three HRSGs.  During normal operation, each HRSG can produce 30,000 lbs./hr of steam, but the steaming rate can be increased to 120,000 lbs./hr with duct firing.  Superheated steam (430–450°F) at 143–145 psig is provided to an inlet header, which then feeds the distribution lines to the campus.  Much of the condensed steam is recovered, with a maximum condensate return of around 85 percent.

Chiller equipment consists of seven York Chillers, with three 2,500-ton-capacity units and four 2,000-ton machines.  System volume is 700,000 gallons, and the chillers feed the entire campus except for Rush Hospital.

Three, five-MW Wärtsilä natural gas-fired reciprocating engines provide additional power, with five diesel generators (four for the hospital, and one for the co-gen plant) as emergency backups.

Much work has been done to improve water/steam chemistry and reliability at this plant, some details of which have been previously reported. [3]  Most recently, the steam generator systems have been investigated for potential FAC.

UIC FAC Initial Test Results

The operations staff at the plant have paid particular attention in recent years to FAC reports from EPRI and updates at conferences such as the Annual Electric Utility Chemistry Workshop.  Even though water/steam pressures are much lower at this facility than utility boilers, a sudden failure caused by FAC could still have very serious and potentially life-threatening consequences.  A feedwater header failure at the plant served as a catalyst for FAC testing.  With ChemTreat’s assistance, UIC personnel brought in a firm that specializes in non-destructive evaluations (NDE) for first-phase testing of those locations anticipated to be particularly susceptible to FAC and to measure the wall thicknesses of these spots.

Given that FAC occurs at spots of flow change in piping systems (another form of the mechanism, two-phase FAC, can occur in boiler drums, deaerators, and other locations with water/steam mixtures, but that is a subject for another article), the initial focus was on suction and discharge piping of the six boiler feed pumps (designated BFP 6, 7, 8, 9, 10, and 11) and blowdown piping from the system deaerator.  In all, 73 locations were examined in this first round of testing.  The pulsed eddy current (PEC) technique was utilized for 63 tests, and ultrasonic thickness (UT) testing was performed on the remaining 10.

Out of the 73 tests, several indicated FAC wall-thinning.  These were:

  • Two 2” and two 3” elbows on the discharge of BFP No. 6
  • A 4” i.d., elbow on the discharge of BFP No. 7
  • A 6” i.d., elbow on the suction side of BFP No. 8
  • A nozzle and spool piece and a weld location on the discharge of BFP No. 9
  • A 6” i.d., elbow on the suction side of BFP No. 10

In two locations on the BFP No. 6 discharge, the wall thickness was below the 87.5% Manufacturer’s Minimum Wall Thickness (MMWT) and in two other instances had approached this value.  The remaining worn spots were below nominal wall thickness (NWT), but above MMWT.  None of the components are in imminent danger of failure, so in all cases the examiners recommended reinspection in three years.  The discovery of these FAC-affected areas shows the tests were very timely.  The plant staff intends to be diligent in operating and controlling the chemistry of the water/steam systems in the future and will incorporate additional testing into the overall program.  The data also shows the need for a more formal risk assessment based on such parameters as geometry, chemistry, and flow rates.  Although elbows have suffered the brunt of the damage, FAC was located at other flow disturbances including a nozzle and weld seam. 

The fact that these initial tests confirmed good pipe integrity after years of operation is undoubtedly a result of the well-managed feedwater chemistry program that has been employed by plant personnel.  Severe FAC has been known to occur in other HRSGs and steam generators in very short periods of time, often caused by poor understanding of the chemistry behind the corrosion mechanism.  The following list outlines some of the key parameters of the UIC program.

  1. Reverse osmosis permeate serves as the steam generator makeup water.
  2. Neutralizing amines are utilized for feedwater pH conditioning, with feedwater pH always maintained above 9.4 and usually within a range of 9.6–10.0.
  3. Sodium sulfite is the reducing agent/oxygen scavenger.  Residual concentrations are maintained within a range of 30–60 ppm.
  4. Dissolved oxygen is measured at the deaerator outlet, equivalent to the boiler feed pump inlets. The typical D.O. range is 2–4 ppb.

Among these items, one clearly stands out regarding FAC control.  Referring to Figure 2, we again note the strong influence of pH on iron dissolution and the curve of minimum corrosion occurring at pH 9.6.  This is a lesson the UIC operations staff adheres to rigorously.  In contrast, the authors have seen numerous reports over the years of plants operating condensate/feedwater systems in a pH range well below 9.0, sometimes with excursions down into acidic ranges.  This is a recipe not only for FAC but for other corrosion mechanisms as well.  Low-pH excursions within boilers proper can initiate underdeposit corrosion and hydrogen damage that may be very rapid, very costly, and, in some cases, may require unit shutdown and lengthy repairs.

Future Work and Recommendations

Additional FAC testing is planned for the UIC West steam plant.  Because FAC is a localized corrosion mechanism, it can occur in isolated spots while the remaining pipe material remains in solid condition. 

A recommendation for this plant and for readers of this article at other facilities is to install instrumentation that allows for regular measurements of condensate, feedwater, and even boiler water iron concentrations.  Iron levels are often a direct indicator of FAC or lack thereof.  With a properly-controlled chemical treatment program, iron concentrations should be in a low part-per-billion range. 

A critical point to remember is that, typically, 90% or more of steel corrosion products exist as iron oxide particulates.  This factor should be considered when selecting any monitoring program. [4]  Tests that only measure dissolved iron will grossly underestimate the total concentration.

Acknowledgement

The authors would like to acknowledge Nigel Mohammed of ChemTreat for his work in helping to arrange the FAC tests at the UIC facility and his support of the chemistry program.   

References

  1. Guidelines for Controlling Flow-Accelerated Corrosion in Fossil and Combined-Cycle Plants, EPRI Report 1008082, Final Report, March 2005, the Electric Power Research Institute, Palo Alto, California. 
  2. B. Buecker, “Basic Thermodynamics:  Why Combined Cycle and Co-Generation Are Much More Efficient Than Conventional Steam Generated Power Production”; Energy-Tech, February 2017.
  3. Buecker, B., Murphy, F., and N. Mohammed, “Steam Chemistry Advancements at UIC”; Industrial WaterWorld (now Water Technology), July/August 2019.
  4. Buecker, B., Kuruc, K., and L. Johnson, “The Integral Benefits of Iron Monitoring for Steam Generation Chemistry Control”; Power Engineering, January 2019.

About the author: Brad Buecker is senior technical publicist with ChemTreat.  He has four decades of experience in or affiliated with the power industry, much of it in steam generation chemistry, water treatment, air quality control, and results engineering positions with City Water, Light & Power (Springfield, Illinois) and Kansas City Power & Light Company’s La Cygne, Kansas station.  He is a graduate of Iowa State University. Buecker has authored many articles and three books on power plant and water/steam chemistry topics.  He may be reached at [email protected].

Frank Murphy is an Assistant Chief Engineer at UIC’s West CoGen Steam Plant. He has over 40 years of experience in steam plant operation beginning as a boiler technician in the U.S. Navy and including more than a dozen years in the power industry at Boyle Energy Services and Technology, Inc.

His email is [email protected].