Mechanical Technology February 2016

⎪ Proactive maintenance, lubrication and contamination management ⎪

was stainless steel to match the internal surface cladding of the lower part of the column. Two areas appeared to have suffered localised damage in the incident, one close to the base and one just below the mid-point, both taking the form of local bulging or wrinkling of the shell. It was noted that both of these areas were on the stainless steel clad portion, raising the fear of possible cracking or disbondment of the cladding. Dimensional examination Once the column had been brought out of the plant and laid horizontally, the bend in the longitudinal direction was clearly

visible (Figure 2). A di- mensional survey showed that the column was about 50 mm out of true at the centre – with the bending being fairly uniform along the length – and presented marked ovality in two dis- tinct areas (Figure 3). Shell thickness is an important factor in the de- sign calculations for any pressure vessel, and thus a detailed survey of the shell thickness was undertaken using a precision ultrasonic

Figure 3: Bending and ovalling results as revealed by dimensional survey.

Figure 2: Bending of the column was clearly visible.

technique. To ensure that accurate readings were obtained, each area measured was cleaned by lightly grinding the surface and removing the roughness due to corrosion, but removing the minimum amount of material. Measurements were taken over a series of regularly spaced positions located in a square grid pattern covering the whole of the cylindrical portion of the column. The results were a fairly uniform thickness with a standard deviation of only 0.34 on a mean thickness of 7.71 mm (Figure 4). No individual thickness measurement fell below the mini- mum allowable thickness of 7.0 mm specified in the design. Special attention was paid to the two wrinkled areas to deter- mine if disbondment of the internal cladding had occurred, but no sign of this was found. Mechanical properties Whilst mechanical properties were of paramount importance, the need for non-destruction of the vessel necessitated an in- direct determination technique. Quite clearly, the vessel could not be moved into the laboratory and neither could material be removed from it, and thus neither conventional tensile testing nor high precision hardness testing using laboratory equipment was possible. A portable hardness testing technique had to be employed, and the Equotip ® system was utilised. This system uses the velocity measurements from a spring-driven hammer as the hammer approaches the surface to be tested and then again as it rebounds. The difference represents the energy absorbed, which can be related to Vickers hardness, HV EQ . The results of a detailed hardness survey on the same posi- tions as the thickness survey gave a mean value of 93.4 HV EQ , with a standard deviation of 6.93 HV EQ , a remarkably uniform result (Figure 5). It was noted that, despite their different ap-

Figure 4: Measurements revealing fairly uniform, but low thickness in the shell.

Figure 5: The hardness survey results showing uniformly low hardness.

pearance, having apparently been locally heated during the fire, no significant reduction in the hardness of the two discoloured zones was found. Relating hardness to ultimate tensile strength in steels is a well-known, though empirical, technique and can be ac- complished with fairly good accuracy using conversion tables published in a number of standards [5, 6]. Relating hardness to yield strength, upon which the mechanical design of the

Mechanical Technology — February 2016

13