Vous êtes sur la page 1sur 9

Combined-Cycle

Water/Steam Monitoring
11/14/2013
Email
Print
0
Facebook
1
Twitter
0
LinkedIn

1
Share

By Brad Buecker, Contributing Editor


Earlier in this series, I wrote about modern techniques to minimize impurity ingress,
scaling, and corrosion in heat recovery steam generators (HRSG). But without accurate
sampling and good data collection, the conditions within an HRSG may be or become
harmful without the knowledge of operators or other technical personnel.
This is another of those cases where an "ounce of prevention is worth a pound of cure."
This article outlines important samples required for successful HRSG operation.

Monitoring the Entire System


Organizations such as the Electric Power Research Institute (EPRI) have written entire
manuals regarding HRSG sampling [2], so obviously a complete overview is not possible
in a single technical article. However, it is possible to discuss the most important
samples and the benefits derived thereby.
HRSGs come in many different designs, but a common general arrangement is shown in
Figure 1. We will use it for this discussion.

Let's begin at the condenser and condensate pump discharge (CPD), with the note that
many HRSGs are not equipped with condensate polishers. For units with water-cooled
condensers, the condenser is the primary source within the system for impurity ingress.
Even a minor leak from one condenser tube can introduce enough impurities to cause
significant damage in the evaporator circuits. In the opinion of many, the CPD sample
point is the most important of any within the entire water/steam network.
Where the condensate is treated by a polisher, the effects of a condenser tube leak are
dampened. However, it is still important that any leak be detected as quickly as
possible to prevent premature exhaustion of the polisher and subsequent carryover of
contaminants to the boiler.
Recommended on-line analyses include:
Cation Conductivity (or becoming popular, degassed cation conductivity)
Sodium
Dissolved Oxygen
The name for cation conductivity has undergone an evolution, and research groups
have now begun referring to the technique as conductivity after cation exchange
(CACE). For this article, we will continue to use cation conductivity. The technique has
been adopted at most plants for detecting impurity in-leakage. If a condenser tube fails
or impurities enter from another source (a makeup water treatment system failure is the
next most likely source), the sodium, calcium, and magnesium salts in the water are

converted to their respective acids, primarily dilute sulfuric and hydrochloric acids, by
the cation exchange column.

The dilute acids that emerge (HCl and H2SO4, primarily) are more conductive than their
respective salts and have an immediate influence on conductivity, thus providing a
quick indication of upsets. The cation column also removes ammonium ions (NH4+),
which are formed by conditioning chemicals added to the feedwater. If ammonium ions
are not removed from the sample, they can mask impurities. In general, the cation
conductivity of a clean condensate sample should be less than 0.2 micromhos
(microsiemens) per centimeter (S). This limit is mandatory for systems that operate on
all-volatile treatment oxidizing [AVT(O)], which has been developed to minimize flowaccelerated corrosion in economizers and evaporator tubes. [3]
So, how does degassed cation conductivity improve upon the method? Air that leaks
into condensers of course contains a small percentage of carbon dioxide. At significant
air in-leakage rates, the CO2 that enters can increase the condensate conductivity and
mask other impurity ingress. Degassed cation conductivity utilizes either a reboiler or
purge vessel (with nitrogen gas feed) to remove carbon dioxide.
Direct sodium monitoring of the CPD is also very effective for detecting condenser leaks
and other impurity ingress. With a tight condenser, sodium levels in the condensate
should be very low (<3 parts-per-billion [ppb]), and in many cases less than 1 ppb. Online sodium analyses in conjunction with cation conductivity are very useful. A
combination of the two instruments provides backup readings in the event of an
instrument malfunction.
Dissolved oxygen (D.O.) analyses are quite important for monitoring air in-leakage to
the condenser from the steam-side. Ideally, if the condenser air removal system is
operating at maximum efficiency, dissolved oxygen levels should be below 10 ppb. A
sudden increase in dissolved oxygen indicates a problem below the waterline in the

condenser. Leaks above the waterline, if severe, may cause significantly decreased
condenser performance. [4]

Deaerator Outlet
The deaerator (D.A.) in multi-pressure combined-cycle units may be integral or separate
to the LP drum, but in either case the D.A. should maintain D.O. concentrations below
10 ppb. Deaerator outlet sampling for dissolved oxygen (grab sampling is most
common) helps the plant operator or chemist monitor deaerator performance. An
increase in D.O. levels over normal concentrations could indicate problems with the
internal deaerator components. Trays may become misaligned. Or, the deaerator vents
may not be adjusted properly.

Feedwater/ Economizer Inlet


This sample is very important, as it is the last checkpoint before the evaporator circuits.
Feedwater chemistry can have a significant impact on boiler operation for several
reasons. First, excessive feedwater contamination will reduce the boiler cycles of
concentration and require increased blowdown. Second, improper control of feedwater
chemistry may cause corrosion of feedwater piping and heat exchanger tubes, which
will introduce corrosion products to the boiler. Third, feedwater is often utilized for main
and reheat steam attemperation. Contaminated feedwater introduces impurities directly
to the steam.
Recommended on-line analyses include:
pH
Dissolved Oxygen
Specific Conductivity
Cation Conductivity
Sodium
Suggested grab sample analysis:
Ammonia
Silica
Ammonia or in some cases an amine (see final paragraph of this article) is typically
utilized to adjust feedwater pH to minimize general corrosion. The optimum feedwater
pH for systems with all-ferrous metallurgy is 9.2 to 9.6. Ammonia monitoring via grab
sample analyses can be quite useful, but most common is to monitor the dosage based

on continuous specific conductivity. Cation conductivity and sodium analyses perform


the same function as for the condensate pump discharge, but also indicate if
contaminated feedwater is being sent to the attemperators.
Continuous dissolved oxygen readings are valuable to confirm that that the deaerator is
performing properly, but except in the rare case where copper alloys exist somewhere in
the feedwater network, in no circumstance should the feedwater D.O. concentration be
zero. As much research has shown, single-phase flow-accelerated corrosion (FAC) is
generated by reducing environments where oxygen has been scavenged by chemicals
such as hydrazine or one of its organic alternatives. [5] In general, single-phase FAC is
most pronounced around 300oF, which in HRSGs corresponds to the LP economizer and
evaporator.
Some recommendations call for continuous sodium monitoring of feedwater, but in this
author's opinion such sampling may be redundant if continuous monitoring is provided
at the condensate pump discharge or after the condensate polisher, if one has been
installed. Grab sample analyses for silica (SiO2) indicate the concentration of this
contaminant that is being transported to the LP drum and beyond, and the
attemperators.
An additional technology, particulate monitoring, is becoming more popular for
condensate and feedwater chemistry evaluation, and it can be particularly useful for
systems operating with air-cooled condensers (ACC). An ACC eliminates the potentially
catastrophic impurity ingress that can occur from water-cooled condensers, but ACCs
are so very large that a vast surface area of carbon steel is exposed to flowing steam
and condensate. Typically, the bulk of the corrosion products that emerge in the
condensate/feedwater consist of iron oxide particles. Particulate monitoring provides a
much better analysis of carbon steel corrosion than dissolved iron analysis.

Boiler Water
Along with condensate pump discharge, the boiler water sample is the most critical. For
drum-type HRSGs as outlined in this article, the high temperatures and the
concentrating effect caused by recirculation of the boiler water require close monitoring.
Furthermore, improperly controlled concentrations of dissolved solids in the boiler water
can introduce excessive contaminants to the steam, where they may form deposits
and/or corrode superheater tubes and turbine components.
Either the drum blowdown line or a downcomer should be utilized for evaporator
sampling, although EPRI recommends the downcomer if the unit is on oxygenated
treatment. [3]
Recommended on-line boiler water analyses include:
pH
Specific Conductivity

Cation Conductivity
Sodium
Phosphate (for those units on phosphate treatment)
Important grab samples include:
Ammonia
Silica
Boiler water pH is one of the few samples where an immediate unit shutdown is
required if the chemistry goes bad. EPRI calls for immediate shutdown if the pH drops
below 8.0 although I am aware of at least one HRSG manufacturer that has set a limit of
7.0. Personally, I feel more comfortable with the 8.0 limit, as if a condenser tube leak or
other serious issue is causing the upset, time is of the essence in taking action. From a
treatment standpoint, many drum units are operated with EPRI's phosphate continuum
program, in which only tri-sodium phosphate (Na3PO4) is utilized for feed with perhaps a
small amount of caustic (NaOH) at start-up. The free caustic concentration should never
be higher than 1 ppm to minimize under-deposit caustic gouging. It must be noted that
for HRSGs as outlined in Figure 1, where the LP circuit provides feedwater to the IP and
HP circuits, the LP circuit must rely on the all-volatile feedwater treatment only with no
phosphate addition. Phosphate can cause difficulties in the IP and HP economizers that
receive feed from the LP circuit.
Specific conductivity is important for monitoring total dissolved solids (TDS)
concentration. Excessive solids will mechanically carry over to the steam, as drum
steam-separators are never 100 percent efficient. This effect is influenced by boiler
pressure, thus the allowable dissolved solids concentration decreases with increasing
pressure. Charts are available to determine TDS limits, but every steam generator is
unique and only actual operation provides the exact carryover data. Cation conductivity
provides a method to evaluate the chloride and sulfate concentration in the evaporator
circuits. Research has shown that these contaminants can cause significant corrosion if
boiler water chemistry is not properly maintained.
With regard to silica, carryover is the primary concern where transfer of silica from
boiler water to steam is primarily by vaporous carryover, not mechanical. The
mechanism is influenced by pressure, where the effect becomes dramatic as pressure
increases. For example, in a 900 psi boiler the recommended maximum drum water
silica concentration is 2.8 ppm to keep silica below 10 ppb in the steam. In a 2,400 psi
boiler the recommended maximum is 0.2 ppm! Again, each boiler will exhibit individual
properties in this regard.
Ammonia sampling is valuable, as at the low phosphate concentration often maintained
in boiler circuits nowadays, ammonia can impart a false pH indication. The operators
may think that chemistry is adequate to prevent under-deposit chloride and sulfate

corrosion, and its progeny, hydrogen damage, when in fact the chemistry is not doing
the job. Grab sample ammonia analyses are typically recommended, where the analyst
can utilize charts to evaluate the chemistry effects. Excel programs are also available
that allow these calculations to be made quickly. [6]

Saturated Steam
In the past, main/reheat steam sampling was considered more important than saturated
steam sampling, and indeed the importance of main/reheat sampling has not
diminished. However, the recommended maximum concentrations of sodium, chloride,
and silica in steam have all been reduced to the extremely low limit of 2 ppb. These
impurities are now known to cause turbine corrosion at very low levels. This issue has
increased the importance of periodic saturated steam sampling for mechanical
carryover, as that is the primary mechanism for Na, Cl, and SO4 to enter the steam.
Main and Reheat Steam
These samples are quite critical, as they provide a direct indication of steam impurities
generated by drum carryover or by feedwater attemperation. Some substances such as
silica exhibit decreasing solubility with decreasing steam pressure, and will precipitate
on turbine blades as pressure lowers upon steam passage through the turbine. Other
contaminants, including sodium, chloride, and sulfate will come out of steam in the
early condensate that begins to form in the LP turbine. These salts can initiate stress
corrosion cracking and corrosion fatigue of blades and rotors.
Recommended continuous analyses include:
Cation Conductivity
Sodium
An important grab sample analysis is,
Silica
The primary measurement is cation conductivity, and it is typically used as the criteria
for performance evaluation and warranty requirement by turbine manufacturers. The
common upper limit is 0.2 S. As in the earlier discussion, degassed cation conductivity
is an alternative to eliminate the influence of CO2 upon the readings. In fact, in some
cases degassed cation conductivity may be particularly useful if amines are used for pH
control in the feedwater. The high-temperatures in superheaters and reheaters induce
decomposition of organic compounds to small-chain organic acids and CO 2. The
compounds naturally will influence conductivity.
Sodium is another good tool for monitoring steam chemistry, as these instruments are
reliable and provide quick results. Sodium readings above 2 ppb suggest impurity

leakage into the steam. Grab sample silica analyses ensure that the limit of 10 ppb is
being maintained in the steam.

Makeup Water Treatment


Space limitations prevent much additional discussion of sampling issues, but briefly,
makeup water treatment system sampling is also important. An increasingly common
makeup water treatment scenario is micro- or ultrafiltration, followed by reverse
osmosis, in turn followed by portable ion exchange or continuous deionization (CEDI) for
polishing. A properly designed and operated system should produce water with a
specific conductivity less than 0.1 S, silica less than 10 ppb, and sodium, chloride, and
sulfate less than 3 ppb. Research has also shown that a good makeup system should
limit total organic carbon (TOC) in the effluent to less than 300 ppb. Organic carbon that
enters the condensate system will break down at the high temperatures in the boiler to
generate small-chain organic acids that potentially may influence turbine corrosion. This
issue is still being researched intensively, but is the primary reason why organizations
such as EPRI do not recommend the feed of organic pH control chemicals, or, if
necessary, organic metal passivator/oxygen scavengers, to condensate and feedwater.

An Additional Monitoring Tool


In my previous articles and in other literature cited in this article and elsewhere, the
severe corrosion that even minor amounts of chlorides and sulfates, particularly
chlorides, can cause in steam generators, especially if they concentrate under deposits.
Chloride and sulfate monitoring is more difficult than other parameters, and these
constituents are often not monitored at all. However, a technique that can be utilized to
detect even part-per-billion (ppb) concentrations of chloride and sulfate is ion
chromatography (IC). The equipment is not inexpensive and it requires a knowledgeable
operator, but for anyone who wishes to closely monitor these harmful constituents, IC is
a good method.

Conclusion
Installation of a reliable sampling system and proper monitoring of the data requires
effort and commitment on the part of plant management. This may sometimes be
problematic given the somewhat minimal staffing at combined-cycle plants. However,
conscientious monitoring and prompt action during upset periods can be worth their
weight in gold if they prevent system corrosion, boiler tube failures and unit shutdowns.
The ultimate price is loss of life, which has occurred due to flow-accelerated corrosion
tube and piping failures at a number of utilities during the last 25 years.
A final note about sampling is that sample extraction and conditioning are critical for
ensuring accurate analyses. I will report on these topics in an upcoming issue of Power
Engineering.

References
B. Buecker, "Corrosion Control in HRSGs"; Power Engineering, July 2011.
B. Buecker, "Combined-Cycle Water/Steam: The Lifeblood of the Plant"; Power
Engineering, June 2012.
Cycle Chemistry Guidelines for Combined Cycle/Heat Recovery Steam Generators
(HRSGs), EPRI, Palo Alto, CA: 2006, 1010438.
B. Buecker, "The Impact of Condenser Performance, Feedwater Heating, and Steam
Reheat on Steam Generator Efficiency"; presentation at the 25th Annual Electric Utility
Chemistry Workshop, May 10-12, 2005, Champaign, IL.
Guidelines for Controlling Flow-Accelerated Corrosion in Fossil and Combined Cycle
Plants, EPRI, Palo Alto, CA: 2005, 1008082.
Buecker B. and S. Shulder, "The Basics of Power Plant Cycle Chemistry"; Pre-Conference
Seminar of the 27th Annual Electric Utility Chemistry Workshop, May 15-17, 2007,
Champaign, IL.

Author
Brad Buecker is a process specialist with Kiewit Power Engineers in Lenexa, Kan., and
a contributing editor for Power Engineering.
"Boiler water pH is one of the few samples where an immediate unit shutdown is
required if the chemistry goes bad."

Vous aimerez peut-être aussi