On Condenser Tube Corrosion and Turbine Blade Copper Deposition

Harpster, J.W.
“On Condenser Tube Corrosion and Turbine Blade Copper Deposition” 11th Southwest Chemistry Workshop, Park City, Utah, August 13-15, 2002.


(This is a summary of a presentation by Dr. Harpster)

Corrosion of copper bearing tubes on the shell side of condensers(1) should be recognized as a contributor to copper deposition on turbine blades. The resulting copper deposits cause loss of load due to a lowering of turbine energy conversion efficiency. The decrease results from an effective change in the blade shape, surface finish and the opening between blades. Both rotating and stationary blades are affected and the copper thickness can be up to 1/16″.  Load loss or excess fuel usage to maintain load can exceed 4%.

Much discussion of this problem is centered around high throttle pressure units, 2400 psi or greater, where it is mostly, but not always observed. Daniels and Latcovich(2) point to an observation that units operating below 1800 psi have not been observed to have a copper deposition problem. This leads logically, perhaps, to the conclusion that copper bearing tubes in feedwater heaters may be principally responsible particularly if excess hydrazine is fed to oxygenated condensate. These authors point out however, that if these heater tubes are replaced by all ferrous tubes the problem is not relieved; believing then that the feedwater system is contaminated and not easily undone.

Although a listing has been provided(2) of risk criteria that plants may have, leading potentially to turbine blade copper deposition, condenser shell side corrosion of tubes due to air in-leakage and treatment was not included.  Since air in-leakage control has a wide variability between operating plants resulting in a consequential wide range in measured dissolved oxygen and ammonia in condensate, it could be another reason two similar plants differ from no copper deposition to requiring regular cleanings.

Corrosion of condenser shell side tube surfaces is known to occur even though temperatures are low. Harpster(3) has shown by analysis how air bound (AB) zones are generated even at low air in-leakage and how stagnant (S) zones develop for large air in-leakage in condensers. Both zones are subcooled and contain a high concentration of noncondensable gases. These gases consist of components from air, principally oxygen, CO2 and nitrogen, CO2 formed from make-up water contaminants in the boiler and from breakdown of bicarbonates and neutralizing amines, and the generation of ammonia from decomposition of hydrazine, or from amines.

With only 6°F subcooling in either AB or S zones, it has been shown(4) that condensate within these zones can reach 90 ppb for dissolved oxygen (DO) from air in-leakage. Depending on the scavenged relative concentration of CO2 and ammonia, very high concentrations of these gases, along with the DO, can exist within these zones.  Also shown was the concentration of DO to be 380 ppb at an extreme but practical level of 25°F subcooling.(4) In the referenced subcooled condition, the steam temperature range was from 108°F to 115°F at different air in-leakage values for the same condenser.

An example of shell side tube corrosion at AB, or S zones is given by Tucci.(5) The subcooled regions containing the corrosive gases were near the center of the tube bundle where falling condensate is cooled in the presence of noncondensables.(3) These regions are predicted using the comprehensive condenser model, first presented by Harpster,(6) where turbine exhaust steam and noncondensable gas dynamics are described. The model is developed from simple assumptions supported by or derived from condenser measurements or observations. Measurements include temperature rise of circulating water for an array of tubes in condenser as shown by Bell,(7) heat transfer coefficient percentage loss versus mole ratio of air in steam on single tube experiments by Henderson and Marchello,(8) and measurement of water vapor to air mass ratio in condenser vent lines by Harpster(9) using the RheoVac® (12) condenser air in-leakage and diagnostic instrument. The assumptions made are, for ideal condenser operations, all tubes in the tube bundle are responsible for condensing the same amount of steam, the pressure drop across the condenser tube bundle can be neglected, there is no hotwell subcooling for most modern condensers (although this subcooling can be described in any condenser exhibiting hotwell subcooling) and within the S zone all tubes are lost from the condensation process.

As a consequence, all observations made on condensers can be correctly described by the model. These include AB or air blanketed zones discussed by Moore and Sieverding(10) located at presumed “pressure sinks” somewhat similar to AB zones of Bell(7) and “stagnation” regions where the gas concentration can build up and retard heat transfer as presented by Marto.(11) The correctness of any explanation of AB or S zones must be supported by a mathematically descriptive method that is compatible with other system related phenomenon. This is the case when the comprehensive model is employed.

Further, since it is shown that the source of copper can be both condensers and feedwater heaters, it suggests something other than the higher pressure and associated temperature in feedwater heaters are responsible. The conditions of temperature and pressure at the deposition site in the turbine should be examined for its role in copper deposition in the higher pressure units.

Return to Technical Library