Understand the importance of water/steam chemistry
Advent of reliable membrane technology changed the landscape for makeup water treatment
By Brad Buecker
|Hydrogen damage of a boiler tube. Illustration courtesy of ChemTreat. Notice the thick-lipped failure.|
Many thousands of industrial plants around the world generate high-pressure steam for process applications and power generation. Yet, water/steam monitoring and control often take a back seat to process operations, even though corrosion, scaling and other problems caused by poor water/steam chemistry can cost a plant millions of dollars. Plant operators, engineers and technical personnel should be alert to critical issues regarding water/steam quality.
The author’s own experience in the process and power industries provides clear examples of the difficulties that occur without proper understanding of modern water/steam chemistry control methods.
One organic chemicals plant in the Midwest United States had four 550-psig package boilers with super-heaters. Plant personnel replaced the super-heaters about every two years due to heavy tube fouling. Deposits between 0.125 to 0.25 inches in depth within discarded super-heater tubes were observed. The boiler saturated steam samples during inspection exhibited considerable foam. Subsequent investigation revealed that total-organic-carbon (TOC) levels in the condensate return to the boilers had been known to reach 200 parts-per-million (ppm). Guidelines from the American Society of Mechanical Engineers (ASME) call for a maximum TOC concentration of 0.5 ppm in boilers of this pressure. It was easy to understand why impurities carried over to the super-heaters.
In fact, steam generators often are left in service during periods of significant impurity in-leakage via a steam surface condenser. Such leaks, which transfer raw cooling water directly to the condensate, introduce many impurities, including such deleterious compounds as chlorides, sulfates, hardness, and silica. In one case in the early 1980s, operation of an 80 MW, 1,250 psig utility-boiler for three weeks with a condenser leak resulted in severe boiler-tube hydrogen damage. Within a month, tubes began to fail at such a frequent rate that the unit had to be shut down and entirely re-tubed.
In another example, a petrochemical plant in the southern U.S. experienced short runs and poor performance of a makeup demineralizer. In particular, the anion resin of the unit regularly lost capacity. Analysis revealed that that the raw water, which came from an area where rice was grown, contained high concentrations of natural organics. The makeup water pretreatment system was incapable of removing many of these complex organic compounds, and thus they continually fouled the anion resins.
For decades, the conventional treatment program for steam generator condensate/feed-water called for feed of an oxygen scavenger, more accurately known as a reducing agent. This chemistry is now known to cause flow-accelerated corrosion (FAC) of carbon steel. FAC-inducted failures have killed a number of plant personnel in the last three decades, and have caused damage in many other plants. Yet, the mindset of oxygen scavenger chemistry in all-ferrous condensate/feed-water systems simply will not die.
With this background in mind, let’s look at some modern methods for optimizing steam generation chemistry and operation.
Makeup water treatment
In the mid-20th century, the most common makeup water scheme was clarification followed by ion exchange (IX). It was reliable and produced high-purity water sufficient for virtually all drum boilers. However, clarification does not remove dissolved solids from water, and thus the loading on downstream IX units could be quite significant. It was common that only a few days of IX operation was followed by resin regeneration. This proved expensive and induced significant wear-and-tear on the equipment.
The advent of reliable membrane technology has changed the landscape of makeup water treatment. Most common nowadays is reverse osmosis (RO) for primary dissolved solids removal, followed by portable mixed-bed de-mineralizers or electro-deionization (EDI) for final polishing.
A key concept for reliable RO operation is upstream removal of suspended solids, where again membrane processes lead the way. Micro- or ultrafiltration (MF and UF) have emerged as techniques for RO pretreatment. However, clarifiers have by no means disappeared. Increasingly, plants are facing requirements to use alternatives to fresh surface water, and these alternate supplies may include reclaim water or groundwater with high dissolved solids.
High suspended solids can potentially overwhelm an MF or UF unit, while groundwaters with high-hardness may be effectively treated in lime-softening clarifiers before downstream processing. While many still think of clarifiers in terms of the older, large-diameter units with rise rates of perhaps one gpm/ft2, modern clarifiers such as a ballast-sand unit operate at rise rates of about 25 gpm/ft2. This improves operating efficiency, particularly with variable supplies, and also reduces unit size.
As with conventional clarifiers, lime and soda ash can be fed to these units to remove calcium, magnesium, and some silica.
Steam generation chemistry control
For condensate/feed-water systems that contain no copper alloys, oxygen scavenger feed is no longer recommended. The regime will initiate flow-accelerated corrosion (FAC) at flow disturbances such as feed-water and economizer elbows, valve exits and such.
Results from the Electric Power Research Institute have shown that iron dissolution is greatly influenced by, not only reducing conditions, but also by solution pH and temperature.
As shown in figure 3, corrosion reaches a maximum at 300 F. Thus, feed-water systems and heat recovery steam generator (HRSG) low-pressure evaporators are particularly susceptible locations. Also note the influence of pH, as reflected by ammonia concentration, on the corrosion characteristics.
The modern method for chemistry control in all-ferrous condensate/feed-water systems is ammonia (or perhaps an amine) feed to maintain the pH within a range of 9.2 to 9.6, with no oxygen scavenger feed. With very pure condensate (≤0.2 µS/cm cation conductivity), the small amount of oxygen that leaks into the system, perhaps supplemented with a small amount of pure oxygen feed, will cause the steel piping to become overlaid with ferric oxide hydrate (FeOOH), which resists FAC. FAC can be combated in the design phase of a steam generator by fabricating susceptible components out of 1¼” or 2¼” chrome steel. Just a small amount of chromium works wonders.
Like feed-water chemistry, boiler water chemistry has undergone a considerable evolution over the past two to three decades. For decades, phosphate programs have been the workhorse of many treatment programs. Tri-sodium phosphate provides alkalinity beyond that of ammonia.
Na3PO4 + H2O → NaOH + NaH2PO4
For many years, programs known as coordinated and congruent phosphate were employed, where tri-sodium phosphate was blended with smaller amounts of di-sodium phosphate (Na2HPO4) and sometimes even a bit of mono-sodium phosphate (NaH2PO4) to generate bulk solutions with a sodium-to-phosphate ratio of 2.3 to 2.6 (congruent treatment). These programs were developed given that at temperatures above 300 F, phosphate solubility decreases rapidly with rising temperature, and that much of the phosphate precipitates on steam generator internals in a mechanism commonly known as “hideout.”
Researchers were under the mistaken impression at the time that the sodium-to-phosphate ratio in the deposits generally remained within the 2.3 to 2.6 range, and thus by maintaining a sodium-to-phosphate ratio within this range in the bulk-boiler water, phosphate would precipitate “congruently” with the residual remaining in the bulk solution. It is now known that the phosphate does not precipitate congruently, and that deposits can have sodium-to-phosphate ratios of less than 2 to 1, making them acidic and leading to direct metal corrosion.
Phosphate chemistry has evolved such that by far the most common program today is EPRI’s phosphate continuum, where only tri-sodium phosphate is used, with perhaps a bit of caustic (NaOH) addition if the pH is low at startup. Most common is to maintain the bulk phosphate below 3 ppm, and often 1 ppm is the norm. A pH range of 9.0 to 10.0 is standard with this program. It is quite important to maintain free caustic alkalinity below 1 ppm to prevent under-deposit caustic gouging.
Due to continued concerns with phosphate hideout and also, via mechanical carryover, phosphate transport to steam, some plant personnel have switched to alternative boiler water treatment programs including straight caustic, again maintaining the caustic alkalinity below 1 ppm.
Monitoring impurity ingress
Even when a plant has the programs outlined above, ignored or undetected impurity ingress can cause significant to catastrophic problems. Impurities can enter a steam generator at a number of locations. By far the most common, and usually most troublesome, source is in-leakage through water-cooled condensers or via contaminated condensate return. Impurities can cause hydrogen damage, corrosion fatigue, scaling caused by hardness compounds or silicates, and carryover and fouling of super-heaters, re-heaters and turbines.
Consider a cooling water leak, or even an upset from a makeup water system. Among the many impurities that enter the condensate is chloride. Even minor ingress, if it occurs often or over a long period of time, allows corrosion underneath deposits. A well-known reaction is shown below.
MgCl2 + 2H2O + heat → Mg(OH)2↓+ 2HCl
While hydrochloric acid (HCl) may cause general corrosion in and of itself, it also will concentrate under deposits, where the reaction of the acid with iron generates elemental hydrogen, which in turn can lead to hydrogen damage of the tubes. In this mechanism, hydrogen gas molecules, which are very small, penetrate into the metal wall where they then react with carbon atoms in the steel to generate methane.
Formation of gaseous methane and hydrogen molecules causes cracking in the steel, greatly weakening its strength. Hydrogen damage is very troublesome because it cannot be easily detected. After hydrogen damage has occurred, plant staff may replace tubes only to find that other tubes continue to rupture.
To prevent such failures, prompt detection of impurity in-leakage and follow-up actions are required. On-line instrumentation (with data feed to a distributed control system or network to alert plant personnel) should properly monitor steam generation chemistry via the following:
- Water-system effluent sodium, silica and specific conductivity
- Condensate pump discharge cation conductivity, sodium and dissolved oxygen
- Feed-water specific and cation conductivity, dissolved oxygen and pH
- Boiler water pH, cation conductivity, specific conductivity, phosphate, sodium and silica
- Main steam cation conductivity, silica and sodium
This article briefly touched upon important concepts for control of water/steam chemistry in high-pressure steam generators. Proper monitoring and treatment is critical for ensuring reliability of these systems. Otherwise, equipment failures are likely, with potentially catastrophic consequences as the result.
Brad Buecker is a process specialist in the Environmental Services group of Kiewit Power Engineers. Buecker has more than 30 years experience in the power and chemical process industries, and has a B.S. in chemistry from Iowa State University, with additional course work in heat and materials balances, advanced inorganic chemistry, and fluid mechanics. He has written many articles and three books on steam generation topics. He is a member of the ACS, AIChE, ASME, CTI, and NACE. He is also a member of the ASME Research Committee on Power Plant & Environmental Chemistry and the program planning committee for the Electric Utility Chemistry Workshop. He can be reached at firstname.lastname@example.org.
Kiewit Power Engineers, Lenexa, Kansas, provides a complete array of design and engineering services for major power plants.