Thunyaluk Pojtanabuntoeng, Curtin Corrosion Centre
Introduction
Thermal insulations used for energy conservation and personnel protection can create an environment perfect for corrosion by forming a narrow annular space that can trap water between the metal surface and an insulation material or by acting as a wet sponge in contact with the metal surface. Once water reaches the metal, corrosion under insulation (CUI) can occur. It is one of the serious corrosion threats to many industries, including oil and gas, due to its insidious nature that challenges the effectiveness of inspection and monitoring. The severity and the likelihood of CUI is greatly associated with the presence of water within the insulated system. It is therefore reasonable to keep water away and avoid the ingression of water. Barriers against water ingress may include a weather barrier and/or jacketing and the use of water-repellent insulations.
This work compares the influence of 2 types of insulation with distinctly different water absorption properties on corrosion of carbon steel. The benefit of drain holes or an escape path for water is also demonstrated and discussed.
Insulation materials and CUI
There are various types of insulation materials. For hot applications where operating temperature is greater than the ambient, insulation materials such as mineral fibre, cellular glass, calcium silicate, and pearlite are generally used [1]. In addition to differences in thermal conservation properties, various insulations have different abilities to absorb and repel water depending on their inherent properties and/or additives used. Mineral fibres (rock wool, glass wool, etc.) are one of the most commonly used insulation materials. These are water absorbent insulations whose hydrophobicity can be improved by certain binders and resins. However, hydrophobicity can gradually degrade with time due to aging and/or thermal degradation. Closed cell insulation such as elastomeric foam and cellular glass are designed to prevent water ingress to the metal surface in relation to open-cell insulation materials and are generally used in cold applications.
Nonetheless, it appears that no insulation is fully safe from CUI as failures have been reported for most of them; e.g. mineral wool insulation [2], [3], glass wool [4], fiberglass [5], and polyurethane foam [3]. Failures in mineral wool system is the most reported probably because the majority of existing structures are insulated with mineral wool.
Not always about keeping water away
Conventional barriers such as hydrophobic insulation, jacketing, weather barrier, etc. have been employed as a standard practice to keep water away. However, general consensus is that flaws are often introduced during manufacturing, installing and maintenance to allow water ingress. Furthermore, the integrity of those barriers can eventually deteriorate with time causing the system to become more susceptible to water ingress. Despite the hydrophobicity of an insulation, water can enter from various locations and accumulate; e.g. when there is flooding and the system becomes completely submerged, the top end of the vertical section is not properly sealed, pre-mature degradation or incorrect application of sealant at the overlapping joints, etc. Therefore, systems that do not retain water may be less susceptible to CUI. Not only is it advised to avoid water penetration but also to ensure effective drainage of water.
Laboratory demonstration of reduced cui due to effective water drainage
Experiments were conducted using a conventional mineral fibre insulation (insulation A – water absorbent) and a new generation insulation (insulation B – non-water absorbent) [6], [7]. Both insulation materials met the criteria for leachable anions (particularly Cl–) according to ASTM C871-11. However, the %water uptake differed significantly between these two insulations. Employing the procedure outlined in ASTM C1511-11, Insulation A absorbed water approximately 5 times of its original weight whereas Insulation B absorbed water about 0.5 times of the insulation original weight.
Corrosion experiments were simulated using a CUI test rig and an environmental chamber that Curtin University developed in collaboration with Santos [6]. An enclosed system was created using a polycarbonate sleeve to mimic the function of field jacketing. This simulates an event when water fully floods the system causing the carbon steel to remain in contact with wet insulation for a prolonged period of time. Expectedly, severe corrosion damage was found after 2 week exposure to water absorbent insulation (insulation A) that had been saturated with the test solution. Localised corrosion in a form of pitting was the predominant mode of damage (Figure 1).
Interestingly, significant pitting was also observed on carbon steel exposed to insulation B (non-water absorbent type) despite the significantly less water absorption. This demonstrated that a small volume of water was sufficient for corrosion to initiate and propagate to an unacceptable level regardless of the hydrophobicity of the insulation. The comparison of normalised localised corrosion rates can be seen in Figure 1.
To demonstrate the influence of water drainage on CUI, three 1-cm drain holes were drilled at 8 cm intervals at the bottom part of the jacketing. In all cases, having drain holes showed beneficial effects at reducing the extent of CUI. Particularly for insulation B, the presence of drain holes effectively reduced CUI as localised corrosion rate was minimal. The nature of this insulation allows effective water drainage that promotes the removal of the excess water within the annular space. In addition, the benefits of using non-water absorbent insulation was also demonstrated even when drain holes were not at the lowest point. The breathability of the insulation permits water vapour to pass through; hence promoting drying out of the insulation particularly for hot application when the operating temperature is greater than ambient.
To highlight the correlation between CUI severity and wetness of insulation, Figure 2 demonstrates an in-situ variation in impedance measured on carbon steel electrode placed underneath the insulation during the corrosion experiments. Low impedance indicates that the insulation remained wet due to the presence of water, and vice versa. In this test, the samples placed at the bottom remained wet for a longer period of time due to the accumulation of water; i.e. 3 days versus 0.5 days. As a results, corrosion at the bottom section was more severe as shown in the photographs.
Figure 1: Comparison of pitting rate among tests with different environments. Bars present maximum pitting rate averaged from triplicate samples. Dots present the maximum pitting rate among triplicate samples. Graph adapted from [7].
 |  |  |
| a | b | c |
Figure 2: a) The change in impedance with time at top and bottom locations. Photographs showing corroded samples at the b) top and c) bottom section. Experiment was conducted using non water-absorbent insulation with top drain holes. Graph adapted from [7]
Drain holes increase the influx of oxygen into the insulation which could increase corrosion initially. However, CUI is reduced in the long term exposure as water gradually evaporates. Evidence also suggests that by inhibiting the formation of differential aeration cells and/or the local depletion of oxygen, localised corrosion can be reduced. Figure 3 depicts 3D images of samples exposed to the water absorbent insulation in the absence and presence of drain holes. It is evident that the extent of localised corrosion had decreased. The depletion of oxygen is confirmed by the formation of magnetite (Fe3O4) as the main corrosion product adjacent to the steel surface identified by SEM/EDS and XRD. The presence of magnetite could induce galvanic corrosion by acting as a cathode in relation to carbon steel which would behave as the anode. This could explain the severe localised corrosion observed underneath the corrosion product in the completely enclosed environment.
Figure 3: 3D images of carbon steel samples exposed to insulation A for 2 weeks in left: the absence (where severe localized corrosion was found) and right: the presence of drain holes (where corrosion appeared more uniform).
Conclusions
Direct correlation exists between the insulation wetness and CUI based on literature and experimental results shown in this report. The importance of drain holes or pathways for water to escape was shown particularly when used in conjunction with the non-water absorbent insulation. The escape path for water should be considered to allow excess water to escape at the lowest point or as vapour when the operating temperature is greater than ambient.
Few exceptions should be noted. Firstly, experiments were conducted on un-coated specimens and did not take into account the aging of coating which could vary greatly. This report also focused on hot application whereas the effect of drain holes on cold application has not been investigated.
References
[1] S. Winnik, European Federation of Corrosion Publication number 55: Corrosion Under Insulation (CUI) Guidelines. 2008.
[2] W. Geary, “Analysis of a corrosion under insulation failure in a carbon steel refinery hydrocarbon line,” Case Stud. Eng. Fail. Anal., 2013 (1), p. 249–256.
[3] A. Babakr and S. Al-Subai, “Under insulation stress corrosion cracking of process piping,” CORROSION 2006, paper no. 06500.
[4] M. Suresh Kumar, M. Sujata, M. a. Venkataswamy, and S. K. Bhaumik, “Failure analysis of a stainless steel pipeline,” Eng. Fail. Anal., 2008 (15), p. 497–504.
[5] D. McNaughtan and M.Najami, “Practical considerations for effective corrosion under insulation (CUI) management from a north sea perspective,” CORROSION 2009, paper no. 9135.
[6] T. Pojtanabuntoeng, L. Machuca, M. Salasi, B. Kinsella, and M. Cooper, “Influence of drain holes in jacketing on corrosion under thermal insulation,” Corrosion, 2015 (71), p. 1511–1520.
[7] T. Pojtanabuntoeng, B. Kinsella, H. Ehsani, and M. Brameld, “Comparison of insulation materials and their roles on corrosion under insulation,” CORROSION 2017, paper no. 9287.