1、2010INTERNAL COATING OF MULTIPHASE PIPELINES REQUIREMENTS FOR THE COATING Ole ystein Knudsen SINTEF Materials and Chemistry Richard Birkelandsvei 2B N-7465 Trondheim, Norway Astrid Bjrgum SINTEF Materials and Chemistry Richard Birkelandsvei 2B N-7465 Trondheim, Norway Ann Karin Kvernbrten SINTEF Mat
2、erials and Chemistry Richard Birkelandsvei 2B N-7465 Trondheim, Norway ABSTRACT Pressure drop along the pipeline is the main obstacle to transportation of unprocessed or partly processed multiphase fluids over long distances. Several parameters contribute to pressure drop in multiphase flow, e.g. li
3、quid hold-up, precipitations, gas-liquid surface drag forces, liquid wetting of pipe wall and surface roughness. Application of coatings inside the pipeline can reduce pressure drop by preventing corrosion, preventing precipitations on the pipe wall and modify the pipe wall wetting properties. Meso
4、scale pressure drop tests have shown that pressure drop is significantly affected by moderate corrosion, even in multiphase flow, demonstrating that application of internal coatings is beneficial. However, the coating must have the same lifetime as the pipeline. In this work we have tried to identif
5、y the most important coating degradation mechanisms and to find relevant test methods for evaluation and qualification of coatings. 2010 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole, must be in writing to NACEInternational, Publications D
6、ivision, 1440 South Creek Drive, Houston, Texas 77084. The material presented and the views expressed in this paper aresolely those of the author(s) and are not necessarily endorsed by the Association.1Paper No.10004INTRODUCTION Pressure drop along the pipeline is the main obstacle to transportation
7、 of unprocessed or partly processed wellstream over long distances. Pressure drop in single phase pipelines is fairly well understood, and available models are able to predict pressure drop with reasonable accuracy. Surface roughness is the most important material parameter in this respect. Pressure
8、 drop in multiphase flow is more complicated and less well understood. Several parameters contribute to pressure drop in multiphase flow, e.g. liquid hold-up, precipitations, gas-liquid surface drag forces, liquid wetting of pipe wall and surface roughness. Surface roughness in multiphase pipelines
9、mainly depends on corrosion and deposits and precipitates. Internal corrosion in low alloy steel pipelines due to CO2is a well studied phenomenon1. Today such corrosion is primarily controlled by use of corrosion inhibitors. However, they are not 100% effective and some corrosion will always take pl
10、ace. In order to decrease corrosion, hydrocarbon wetting of the pipe wall is important. Top of line corrosion due to condensation of water, insufficient delivery of corrosion inhibitor and insufficient hydrocarbon wetting has in some cases been a significant problem2. Application of a coating inside
11、 the pipeline may provide an efficient solution to both the corrosion and the precipitation and deposition problem. The application of internal coatings in order to decrease pressure drop in multiphase pipelines have been studied in other projects, e.g. the Deepstar project. A wide range of surface
12、coating materials has been tested with respect to wax deposition, including plastics, PTFE etc. In the oil the abrasive silica slurry is carried up, out of a bath of slurry. The abrasive particles are fed into the gap between the wheel and specimen and thus abrade it. The ASTM D4060 Taber abrader ca
13、n also be used in order to evaluate the wear resistance of the coatings13. In this test the organic coating is applied on a flat surface, and after curing the surface is abraded by rotating the panel under weighted abrasive wheels. Abrasion resistance is calculated as loss in weight at specified num
14、ber of abrasion cycles. Corrosion at coating damages Few of the standards found describe specific corrosion tests for internal pipeline coatings, probably because they are made for testing coatings for dry gas pipelines where corrosion should be no problem. The only pipeline coating standard that de
15、scribes a corrosion test is API RP 5L2, which is using the ASTM B117 continuous salt spray test. For atmospheric coatings, the salt spray test is heavily criticized for having low correlation to field exposure. Whether this is the case for internal pipeline coatings as well has not been investigated
16、, but due to the very different environments in the test and the pipeline this can be expected. Jelinek tested the corrosion protection properties of an internal epoxy coating with the following test method 3: Electrolyte: 8% NaCl, pressure: 100 kPa CO2, temperature: 60C. The test duration was 60 da
17、ys. After the test corrosion rate was determined by electrochemical impedance spectroscopy (EIS) and atomic absorption spectroscopy analysis of iron in the water phase. Testing was performed in the pure electrolyte and the electrolyte with 30% crude oil. After the test the coating was visually inspe
18、cted for undercutting. He concluded that the corrosion rate at coating damages was comparable to uncoated steel. EXPERIMENTAL Test panels and coatings Flat steel panels about 60 mm x 100 mm x 3 mm were cut from carbon steel panels (DIN 17100/UNS G10150). Prior to coating application, the panels were
19、 pre-treated by blast cleaning to Sa 2 (NACE No. 2). Table 2. Coating systems. Code Number of coats Generic type A 1 coat Solventless epoxy B 1 coat Solventless epoxy C 1 coat Solvent borne epoxy D 1 coat Solventless epoxy E 2 coat Epoxy + epoxy powder F 2 coat Epoxy + epoxy powder The reverse side
20、and bare edges of the different panels were sealed with an epoxy mastic coating. Prior to corrosion testing, 2 mm scribes were made in the coating, parallel to the longest edge of the samples, using a horizontal milling machine. Coating C, D, E and F were applied by the coating supplier. 5Corrosion
21、testing Some of the panels were pre-corroded by salt spray in order to create a crevice between the coating and the substrate. Our aim was to obtain 3 mm scribe creep corrosion before the autoclave test, but this was not achieved for all coatings. The purpose with the pre-corrosion was to study whet
22、her corrosion inhibitors were able to penetrate the crevice under the coatings and prevent corrosion. Corrosion testing was carried out using two identical autoclaves. The panels, three parallels of each coating system, were placed in each autoclave as shown in Figure 1. A volume of 7.5 - 8 litres e
23、lectrolyte was added, resulting in halfway submersion of the scribed panels. The electrolyte was a 5% NaCl solution, stabilized at pH 4.0 with an acetate buffer. A film forming inhibitor in a concentration of 50 ppm was added to one of the autoclaves. 100 % CO2was purged through the electrolyte whil
24、e the temperature was increased to 60C and the pressure was increased to 1000 kPa. The electrolyte was purged with CO2for one day. pH measurements and CO2purging were repeated approximately once a week. After 24 days of exposure, the pH had increased to 4.4 and 4.7 in autoclaves with and without inh
25、ibitor, respectively. The panels were examined visually after 27, 40 and 56 days of exposure. Due to severe blistering in submerged areas samples A, C, and D were removed after 56 days. The test was terminated after 81 days for system B, E and F. The electrolytes in the two autoclaves were changed t
26、hree times, after 27, 49, and 62 days of exposure. After testing, the panels were rinsed using a nylon brush and running tap water. The exposed panels were photographed. Surface topography of corroded panels of coating systems A, C, and D were also documented by 3D images. Delaminated coating along
27、the scribe was then removed using a scalpel. Figure 1. Autoclave with coated panels halfway submerged in the chloride electrolyte. Photographed after the corrosion test was finished. 6Decompression blistering The coated samples were exposed in an autoclave for 24 hours halfway immersed into the same
28、 electrolyte as used for the corrosion testing. The temperature was 100C and the autoclave was pressurized with nitrogen at 15 000 kPa. After 24 hours the temperature in the autoclave was taken down to 90C. The pressure was then released at a rate of 2000-4000 kPa/min. RESULTS AND DISCUSSION Corrosi
29、on Due to severe blistering in submerged areas, samples A, C, and D were removed from the test after 56 days. The remaining coating systems were exposed for 81 days. After 81 days some blisters were observed in coating B, while the two-layer coating systems E and F showed no blistering. Increased co
30、ating film thickness and application of two or more coats are known to improve the barrier properties, and thus increase resistance against osmotic blistering. This obviously was the case in this test as well. Table 3 shows that the three coatings that developed blisters in the test also were the th
31、ree thinnest coatings. Figure 2A shows blistering of coating A after 56 days exposure. After the test the panels showed deep corrosion attacks in and around the scribe in submerged areas. In the part of the sample exposed in the gas phase corrosion had not spread outside the original scribe. On two
32、of the panels the corrosion attack had even penetrated the 3 mm steel panel. Removing the coating along the scribe showed that the steel substrates were severely attacked by uniform corrosion. Corroded areas were generally deeper on panels exposed in the test solution without inhibitor, indicating t
33、hat the inhibitor had some effect on the corrosion rate. However, the selected inhibitor was not very effective under the conditions used, which somewhat limited the value of this test. After removing delaminated coating, the width of the corrosion attacks along the scribe was measured. Table 3 show
34、s the maximum length of corrosion from scribe for all the samples tested. The results show that the inhibitor had no effect on the width of the corroded areas. This is primarily explained by the fact that the inhibitor had a limited effectiveness under the test conditions used. The question whether
35、the inhibitor is able to penetrate into the crevice under a delaminated coating can therefore not be answered after this test. For coating system A and B both pre-corroded and fresh samples were exposed in the autoclaves. There was no difference in depth of attack or width of affected area between p
36、re-corroded and fresh samples. Also, the corrosion width was independent on type of coating. Coating A, C and D had less corrosion than Coating B, E and F, but this was due to the fact that B, E and F were exposed for a longer period of time than A, C and D. The corrosion attack under the coatings p
37、ropagated somewhat differently from what is observed on coatings in corrosive atmosphere. In corrosive atmosphere the corrosion creeps under the coating, usually giving rather shallow attacks, and the corrosion products lift the coating from the substrate. In this test the corrosion seemed rather to
38、 propagate by an anodic undermining mechanism, like illustrated in Figure 3. Looking at the surface of the sample after the test before removal of loose paint, one could not see how far the corrosion had propagated, because the corrosion products were not pushing the coating up. The fact that the wi
39、dth of the corrosion attack was independent of type of coating also suggests that the corrosion mechanism is different from what we find in atmosphere. 7Corrosion creep in atmosphere is normally considered to propagate by a mechanism where the coating in front of the attack looses adhesion by cathod
40、ic disbonding, and that corrosion creeps after when adhesion is lost14,15. In the autoclave test the samples were exposed in an oxygen free environment, and earlier studies have shown that cathodic disbonding is very slow in oxygen free environments16. This may explain why the corrosion seems to pro
41、pagate by a different mechanism. The rather deep attacks of general corrosion in and around the scribe also indicate that corrosion inhibitors should have no problems entering the corrosion front under the coatings, and that we perhaps does not need to fear that the inhibitor will be ineffective und
42、er a delaminating coating. This will be investigated further. Table 3. Corrosion from scribe after corrosion testing in autoclave. No inhibitor Inhibitor DFT m Exposure days As painted mm Pre-rusted mm As painted mm Pre-rusted mm A 171 9 56 7.0 7.2 5.2 5.8 B 311 11 81 10.7 11.2 10.8 10.2 C 69 1 56 7
43、.1 6.1 D 91 1 56 8.1 8.4 E 247 2 81 10.7 11.8 F 228 1 81 10.2 11.7 Figure 2. Coating degradation after corrosion testing. A: Blistering due to low film thickness. B: Corrosion in coating damage. A B 10 mm 10 mm8Figure 3. Propagation of corrosion under the coating when submerged in formation water. D
44、ecompression blistering The results from the decompression test are shown in Table 4. The two powder coatings gave the best performance in the test. The blisters that appeared on coating E were all superficial and could not be regarded as a degradation of the coating. Coating B also showed rather go
45、od performance in the test. Coating A had lots of blisters in the submerged part of the sample, but no blisters in the gas phase. This indicates that the blisters may have been formed during exposure in the electrolyte prior to decompression. As shown in Figure 2A, coating A was susceptible to blist
46、ering when exposed in the electrolyte. Investigation of cross sections after the test confirmed this, since all the blisters were located at the steel/coating interface. Hence, coating A gave no decompression blistering. Coating C and D blistered both in the immersed part and the part exposed in the
47、 gas phase. Again the blisters in the immersed part of the samples were located at the metal coating/interface and most likely appeared before decompression. However, the blisters above the liquid phase must be due to the decompression. Figure 4 shows a cross section of a blister in the gas phase of
48、 sample B. The picture shows a crack in the coating between several pores in the coating, about 50-100 m in diameter. Hence, the blistering seems to be associated with pores in the coating. During compression the gas may enter these pores, and when the pressured decreases the gas expands and creates
49、 cracks and blisters in the coating. The amount and density of pores in coating B indicates that the coating was not applied in an optimal manner. The solvent less epoxies are rather viscous and require high pressure spraying equipment, and for some products it is also recommended to heat the wet paint in order to decrease the viscosity. The amount of pores in this film indicates that the paint was too viscous during application. The decompression rate used in this test is quite high, compared to the decompression rate observed in a pipeline. During a co
