RECOVERY BOILER SCREEN TUBE FAILURES

The black liquor recovery boiler is an essential component of the Kraft pulp and paper mill. Its two critical functions are the recovery of process chemicals used in the pulping process, and the generation of steam, which is used throughout the mill for a variety of processes.

The recovered chemicals, which consist primarily of sodium carbonate and sodium sulfide, are extracted as a molten "smelt" from the bottom of the boiler. Essential to the Kraft pulping process, their recovery is economically mandatory. However, their presence in the bottom of an operating boiler poses a challenge to boiler owners, designers, and operators throughout the industry because of the potential for catastrophic water/smelt explosions should any of the tubes develop a water leak.

Over the years, we have evaluated hundreds of tube failures from dozens of boilers, involving many different failure mechanisms. Recently, two different clients submitted screen tubes from their boilers, for laboratory analysis to determine the cause of failure, as described in the following two case histories.

Overheating Failure of Screen Tube

During an annual inspection of a 25-year old recovery boiler, one of the screen tubes was found to have a bulge and transverse cracks on its outside surface. Fortunately, timely detection and removal of the tube avoided a potentially catastrophic water leak into the furnace smelt bed.

The tubes were electric resistance welded, 2 1/2 inch diameter by 0.145 inch thick, and were composed of low carbon steel per ASME SA-226. The affected screen tube was the lowest horizontal tube in the screen tube platen and first in line in the flue gas passage. Sections of the affected tube and an unaffected tube from an adjacent platen were submitted to our laboratory for examination.

The tube exhibited closely spaced, transverse cracks localized to a ten-inch band along the bottom half of the outside surface (shown above,left); this was also the region of the bulge. The tube surfaces were otherwise unaffected by pitting corrosion or general corrosion. At the bulge, the tube thickness had been reduced to as low as 0.100 inch. Waterside deposits in the bulge were four times greater than in the other areas of the affected tube or in the adjacent tube

Microstructural examination at various cross sections of the tube indicated it had been overheated at the bulge. The transverse cracks (shown above, right) were singular, oxide covered, and were characteristic of thermal fatigue cracking. The thermal fatigue cracks were found only at the bulge and were up to 0.040 inch deep, which reduced the effective tube wall thickness at that location to as low as 0.060 inch.

The tube failure was attributed to localized tube overheating. Although the exact cause of the overheating was not certain, it was likely due to partial blockage of the particular screen tube or an otherwise induced inadequate boiler water flow, possibly during boiler start-up. The loss of the tube strength at the overheated temperature likely caused the localized bulge. During subsequent boiler operation, transverse thermal fatigue cracks had occurred due to high thermal cyclic stresses in the bulged area. The localized nature of the bulging and cracking on the affected tube, and the absence of damage in the adjacent tube, indicated the damage probably represented an isolated problem, rather than a widespread condition.

Weld Cracking Failure of Screen Tube

Hydrostatic pressure testing during restart of a 20-year old recovery boiler revealed a leak at a superheater screen tube. The leak was from a transverse through-wall crack, several inches long at the end of the tube space-bar weld. The tube was 2 inches diameter by 0.203 inch thick, and was composed of ASME SA-178, Grade A material. A tube section containing the crack was submitted for our metallurgical examination to identify the cause of the failure.

Examination of the crack surfaces revealed that they were covered with oxides and contained numerous fine striations, characteristic of fatigue cracking. The fatigue crack had initiated at the weld toe and propagated progressively across most of the wall thickness before the remaining ligament finally burst, causing the leak. The crack was primarily singular and transgranular, with some limited branching (shown above, right). No evidence of tube overheating or otherwise abnormal boiler service was found.

The tube failure was attributed to long term fatigue crack propagation. The primary contributory factor was the sharp transition at the space-bar weld toe, which acted as a notch for fatigue crack initiation. The weld profile was unacceptable, and in the absence of the defective weld the failure would most likely not have occurred. We recommended the remaining superheater screen tube and space-bar weld joints be inspected and any abnormal weld profiles be ground to provide a smooth transition radius to reduce the stress concentration.

OXIDE SURFACE DAMAGE IN A PLASMA ENVIRONMENT

We have seen the effects of high temperatures on many materials through our inspections of engine parts, kilns, boilers, pressure vessels, etc. In the high-tech world, temperature-related damage is sometimes the result of plasmas and electron/ion beams, which are used for depositing layers and etching surfaces. The damage includes etching, melting, slagging, particle agglomeration, and sintering.

In a typical plasma-related failure, we evaluate the damage on chip surfaces by inspecting with either an optical or electron microscope. Additionally, because the components are so small, any chemical analyses must also be conducted using electron or optical microscope techniques, rather than by bulk chemical methods.

A recent case involved an oxide film which had cracked, etched around the crack and left unwanted oxide deposits at other locations. The figure at left shows a cross section of the oxide layer. The plasma environment etched a shallow, flat-walled, V-shaped trough above each crack. When manufacturers are making chips, they try very hard to get conditions just right so that these flat walls will form.

By changing the carrier gas, active gas concentrations, vacuum pressure, electrical environment (voltage and wave forms), and physical dimensions of the chamber and electrodes, one can get just about any shaped trench one likes.

CHEMICAL REACTOR LEAK

The hexagonal nut was of Type 316 stainless steel and had fractured in two halves (shown at right), resulting in the failure of the connection and the subsequent leak.  The fracture was brittle and had several adjacent secondary cracks.

Metallographic examination indicated that the cracks (shown below) were transgranular, multiple and branching, characteristic of stress-corrosion cracking in austenetic stainless steels.

Stress corrosion cracking is a time-dependent cracking phenomenon which occurs to susceptible materials under the combined influence of tensile stresses and an aggressive environment. The cracks initiated at highly stressed areas of the fittings and propagated to fracture. An energy dispersive spectroscopic analysis in the scanning electron microscope revealed a significant presence of chlorides in the caustic residues on the fracture surfaces and the deposits within the secondary cracks.

The 300-series austenetic stainless steels are highly susceptible to chloride-induced, stress corrosion cracking. No significant corrosion problems were found elsewhere in the reactor vessels or the piping system components.

Corrective measures included replacement of the highly stressed fitting with welded flange connections, and the use of duplex stainless steels, which are more resistant to chloride-induced stress corrosion cracking than the 300-series austenetic stainless steels.

Bob Williams, our Senior Engineering Technician, is retiring after 36 years of exemplary service at MEI-C. Bob joined us back in 1961, fresh out of school at Oregon Technical Institute (Oregon Institute of Technology at that time), with an Associate of Engineering degree in Mechanical Engineering Technology, and an Associate degree in Auto Machine Technology. Over the years, his expertise in strain gaging, failure analysis, fire and arson investigations, trucks/automobiles, and machining has been invaluable to us.

Among his many other hobbies and activities, Bob is a Studebaker enthusiast. Also, he and wife Carolyn are active volunteers with the Red Cross at the local and national levels. With their avid interests in helping others and numerous other activities, we are sure that they will be busier than ever. We wish you well in your retirement!

Shara and Ross Sundberg became the proud parents of a baby boy, Bradley Orion Sundberg, on April 13, 1997. Bradley weighed 7 lbs, 9 oz, and was 21 1/2 inches long. Our congratulations to Shara and Ross!

MEI-C HOSTS HIGH SCHOOL APPRENTICESHIP STUDENT

MEI-C, in cooperation with the City of Portland and Saturday Academy, is sponsoring Nicole Salvus, a junior at Benson High School, in the Apprenticeships for Science and Engineering program. The program offers outstanding students from Oregon and Washington the opportunity to spend the summer working at various academic and industrial centers.

Nicole is working at MEI-C with Dr. Andrew M. Held, studying permeation of chemicals through disposable gloves.

Her work involves measuring the rates at which chemicals pass through gloves when they are present as mixtures with other chemicals. Many previous studies of glove permeability have neglected synergistic and inhibitory effects in mixtures of solvents and chemicals.

A second project will be studying the reactions of vitamin C, and will involve simulating physiologic conditions under which Vitamin C is active in the body, and analyzing the end-products of its reactions