Short-Term High Temperature Failures
David N. French
President, David N. French Inc, Mettalurgists
Summary: The following article is a part of the National Board Technical Series. This article was originally published in the April 1991 National Board BULLETIN. (4 printed pages)
The operation of a boiler is a dynamic balance between heat flow from the combustion of a suitable fuel and either steam formation within the furnace or steam super heating within the superheater or reheater. In effect, the steel tube is "heated" by the flame or hot flue gas and simultaneously "cooled" by the fluid (steam, water or a mixture of steam and water) flow. When this balance is maintained within the design limits, metal temperatures are also maintained within design parameters.
TABLE I -- SHORT TIME ELEVATED TEMPERATURE TENSILE STRENGTH
|TEST TEMPERATURE, oF
||TENSILE STRENGTH, PSI
The expected life under these conditions can be measured in a score or more years. However, when the balance is upset, metal temperatures rise and failures occur sooner that expected. Depending on the relative temperature rise, failures can occur either very quickly, that is, in a matter of minutes; or over a much longer time period, that is, in matter of many months. For convenience these two regimes are defined as "long term" and "short-term" overheating. This article will discuss "short-term" high-temperature failures.
For all materials used in boiler construction, the strength decreases as temperature increases. Table I lists the short-term tensile strength for SA192 and SA213 TP321H that illustrates this point.
The simplest explanation for all "short-term" overheating failures is: when the tube metal temperature rises so that the hoop stress from the internal steam pressure equals the tensile strength at elevated temperature, rupture occurs. For example, in a super-heater of SA192 tubes, with a designed metal temperature of 800oF, the ASME Boiler and Pressure Vessel Code gives the allowable stress at 800oF as 9,000 psi. If the tube-metal temperature should rise to a temperature of around 1300oF, the hoop stress would be equal to or slightly greater than the tensile strength at 1300oF, and failure would occur in a few minutes.
The balance between heat flow and fluid flow can be upset from either side; too much heat flow or too little fluid flow. In a waterwall tube, steam forms as discrete bubbles, nucleate boiling. When the bubble is large enough, the bubble is swept away by the moving fluid, and the cycle repeats. At too high a heat flux or too low a fluid flow, steam-bubble formation is too fast for removal by the moving fluid. Several bubbles join to form a steam blanket, a departure from nucleate boiling, DNB. Heat transfer through the steam blanket is poor, steam is an excellent insulator, and tube-metal temperatures rapidly rise and failure occurs quickly.
In a superheater or reheater, DNB cannot occur as only steam super heating takes place, no boiling. However, short-term overheating failures do occur but usually during start-up. Boiler operational problems that can lead to these short-term high-temperature failures include, among others:
- Flame impingement from misaligned or worn burners that leads to the formation of a steam blanket, as the local heat flux is too great for the fluid flow through the tube.
- Blockage of a superheater tube with condensate or foreign material that prevents steam flow. These problems are more frequent during start-up.
- Reduced flow in either a water or steam circuit that leads to inadequate cooling. Pinhole leaks, especially at poor welds or slag falls, severe dents from slag falls or ruptured tubes, and partial blockage from debris or other foreign matter are some of the more obvious causes.
- Foreign objects, broken attemperation- spray nozzles, for example, in headers that partially block a superheater or reheater tube.
Regardless of the location within the boiler that these failures occur, the appearance is similar. There is a wide-open burst with the failure edge drawn to a near knife-edge condition, and the length of the opening four or five tube diameters. These failures display considerable ductility; the thinning at the failure lip may be more than 90% of the original wall at the instant of rupture. The microstructures throughout the failure will usually indicate, in the case of ferritic steel, the peak temperature at the time of failure. For ferritic steels there is a transformation from ferrite and iron carbide or pearlite, to ferrite and austenite. This temperature is referred to as the lower-critical transformation temperature and occurs at 1340oF or higher, depending on the exact alloy composition.
For failures that occur below the lower critical transformation temperature, the microstructure through the failure lip shows considerable distortion and elongation of the ferrite and pearlite. If the peak temperature at the instant of failure is in the austenite and ferrite temperature regime when failure occurs, the escaping steam will rapidly cool the metal to the steam temperature. The resulting microstructure will contain ferrite which undergoes no transformation on rapid cooling, and bainite which is the transformation product of austenite on rapid cooling. The failure lip may also show considerable distortion to the ferrite and bainite, again, indicative of a rapid metal deformation at the time of failure. At the end of the failure where there has been much less distortion and swelling of the tube prior to failure, the microstructure is ferrite and bainite, but without the elongation to the grains, indicative of tube swelling.
Under rare conditions, and usually in low-pressure boilers, the peak temperature at the moment of failure can be in excess of 1600oF, a microstructure at the time of failure that is all austenite. When these failures occur, the microstructure does not show the distortion to the austenite, as the transformation to bainite effectively wipes out the evidence. When the cooling rate is slightly slower, ferrite will form preferentially at the austenite grain boundaries and along certain planes of atoms in the austenite. The remaining austenite then bans forms to bainite. The resultant microstructure is a mixture of ferrite and bainite and is sometimes referred to as a "Widmanstatten structure."
One final comment, ductile failures can also occur at normal operating conditions but are not, strictly speaking, high-temperature failures. Wastage of a tube from corrosion or erosion can reduce the wall thickness, which, in turn, raises the hoop stress. Such failures occur in waterwall tubes, for example, where sootblower erosion has reduced the wall thickness, or in the convection pass from fly-ash erosion. These failures can occur at normal operating temperatures if the wall thickness reduction is sufficient.
While these microstructures and the estimated peak temperature at the time of failure cannot predict the sole cause of the failure, the metallurgical analysis can suggest the kind of boiler-operational problem that is likely to be the cause of the rupture.
Editor's note: Some ASME Boiler and Pressure Vessel Code requirements may have changed because of advances in material technology and/or actual experience. The reader is cautioned to refer to the latest edition of the ASME Boiler and Pressure Vessel Code for current requirements.