Correctly specifying materials in chemical process industries
The rapid failure of several 316 stainless steel pipes in a heat exchanger raised questions as to whether a manufacturing fault or service conditions were to blame. This second in a series of three case studies conducted by Dr. Elayaperumal examines the corrosion of stainless steel tubes in chloride solution and why duplex would have been the best material of choice.
^ Chemical process industries
Article by Dr. Kailasanathan Elayaperumal, Metallurgical Consultant, Chandan Steel Ltd, India
The 45 tubes under discussion meansure 25.4 mm OD x 1.6 mm WT x 6 M long, made of A-213/ Type 316 SS and were cold finished seamless, annealed and white pickled. A shell- and-tube heat exchanger made from these tubes leaked after about three months of service. Seven out of 45 tubes were found to have leaked.
This case study deals with the use of type 316 stainless steel (SS) tubes in an application where a suitable duplex stainless steel (DSS) should have been specified as the appropriate material of choice.
Cracking from the inside
Photographs 1 and 2 below show a sample cut from a leaking tube. Through and through cracks, visible on both outside and inside surfaces and thick black coating on the inside surface can be seen. The crack is wider on the inside than on the outside surface, indicating that crack initiated from the inside.
These features represent a localized corrosion phenomenon and not a general corrosion phenomenon attributable to any quality deficiency of the tubes.
Photographs 3 and 4 show the metallurgical microstructures across the wall of the tube at the cracked position.
The following features are noticeable:
- The general structure is that of properly annealed austenitic stainless steel representative of 316SS.
- There are cracks initiated at multiple locations (not only one location) on the inside surface and propagated towards the outside surface.
- Cracks have propagated in a transgranular manner (cutting across grains and not necessarily through grain boundaries).
- Cracks have propagated in a branched manner (tree-like branches). All the above features are representative of Chloride Stress Corrosion Cracking of Austenitic Stainless Steel (CSCC) initiated on the inside surface of the tubes.
The following are the service conditions experienced by the tubes:
Shell side: Thermic fluid (hot oil) entering at 175°C and leaving at 160°C at a pressure of 4.5 kg/cm2. This fluid is the source of heat, transferring heat across the wall of the tube to the inside fluid.
Tube side: Industrial waste water of 40,000 ppm total dissolved salts (TDS), mostly chlorides, heated from 110°C to 130°C at a pressure of 6 kg/ cm2. Part of the water is converted to steam used as a heat source in down-stream multiple effect heat exchangers.
The black coating seen on the inside surface is due to concentrated waste wet sludge (much higher than 40,000 ppm chloride in the bulk water) remaining in contact with the inside surface after a considerable portion of the water evaporates and goes out as steam.
The thermic fluid is pure oil without any aqueous base dissolved salts. Hence there is no black coating/ deposit on the shell side surface of the tubes.
In industrial waste waters, chlorides are invariably present in appreciable quantities. Chloride even in small ppm level leads to CSCC in stainless steels at temperatures above about 50°C.
The stress required for CSCC to occur in this case is residual stress present even in annealed tubes. The level of stress required for CSCC to occur is quite low. A small percentage (even 2 or 3 %) of the allowable stress as per ASME is sufficient. For a heat exchanger tube, residual stress can occur throughout the length as a result of high pressure hydro testing during manufacture or locally near expanded tube ends or near baffle positions through non-noticeable vibrations.
This level of stress in a heat exchanger varies from tube to tube. Hence only 7 tubes out of 45 failed.
CSCC also requires a time factor to initiate and to completely propagate to leakage depending upon the level of stress present, wall thickness, level of chloride present and also operating temperature. In this case it has taken about three months for the complete propagation of the crack to the point of leakage.
Cracks or any other failure occurring due to any manufacturing defect, including tube quality or tube manufacturing steps, should have occurred at a single location, without multiple locations, and almost immediately within a few days of operation, not after 3 months.