Compressor Instability with Integral Methods Episode 2 Part 4 docx

Chapter 8 Surface Quality Aspects 8.1 Surface Quality Features Quality features of substrate surfaces determine the performance and properties of applied coating systems. Related to the performance of corrosion protective coat- ing systems, ISO series 8502 states the following: “The performance of protective coatings of paint and related products applied to steel is significantly affected by the state of the steel surface immediately prior to painting. The principal factors to influence this performance are: (1) the presence of rust and mill scale; (2) the presence of surface contaminants, including salts, dust, oil and greases; (3) the surface profile.” Numerous regulative standards are issued to define these factors, and testing methods are available to quantify them. Blast cleaned surfaces show some distinct features, and extensive experimental studies were performed to address this special point, often in direct comparison to other surface preparation methods. Figure 8.1 shows a scanning electron microscope image of a steel substrate after blast cleaning. A general distinction can be made between primary and secondary substrate surface features. Primary surface features may include those properties which are part of coating specifications. They include the following features: r visual cleanliness; r chemical cleanliness; r physical cleanliness; r profile properties; r surface integrity. Secondary surface features may include properties which are determined by, and depend on, primary features. They may include the following features: r surface energy; r wettability; r adhesion of coating to substrate. A. Momber, Blast Cleaning Technology 337 C  Springer 2008 338 8 Surface Quality Aspects Fig. 8.1 Morphology of a blast cleaned low-carbon steel substrate. Parameters: p = 0.475 MPa; d N = 8 mm; ϕ = 90 ; abrasive: aluminium oxide; ˙m p = 19 g/s; d P = 165 μm (Photograph: author) 8.2 Visual Cleanliness 8.2.1 Visual Standards A number of regulatory standards have been developed in order to define and to characterise steel surfaces prepared by blast cleaning. These standards are more or less based on the standard preparation grades given in ISO 8501-1 (uncoated parts of the surface) and ISO 8501-2 (partial surface preparation). Visual blast cleaning standards cover the following two issues: r initial condition (rusty steel or shop primer); r visual surface preparation definition (visible contaminants and cleaning degrees). Table 8.1 provides a general review on current visual surface preparation stan- dards. Visual standards should always be used in conjunction with the written text, and they should not be used as a substitute of a written standard. The standards listed in Table 8.1 are limited to hot-rolled steel surfaces prepared for painting. They are applicable also to steel substrates that show residues of firmly adhering paint or other foreign matter in addition to residual mill scale. Therefore, care must be taken in applying these standards to other substrate materials. 8.2 Visual Cleanliness 339 Table 8.1 Contents of visual blast cleaning standards Standard Surface reference for Rusty steel Coating/primer Flash rust Cleaning degree ISO 8501-1 xx ISO 8501-2 xx x International Slurry blasting Standard a xxx SSPC – NACE xxx a Issued by International Paint, Newcastle 8.2.2 Initial Conditions Initial conditions have a notable effect on the performance of coatings applied to blast cleaned steel substrates. The effects of rust grades and preparation grades on coating lifetime are illustrated in Fig. 8.2. It can be seen that lifetime decreased if rust grade increased and if preparation grade decreased. A high rust grade (“C” or “D”) required a higher degree of surface preparation in order to guarantee equal coating lifetimes. The low surface preparation standard Sa 1 was extremely sen- sitive to changes in the initial conditions. Rust grade “A” (adhering mill scale but little, if any, rust) was least sensitive to changes in surface preparation degree, but 20 rust grade A B C D 15 10 5 0 1 23 preparation grade coating lifetime in years 4 Fig. 8.2 Effects of rust grade and preparation grade on coating lifetime (Bahlmann, 1982). Prepa- ration grades: 1 – brushed; 2 – Sa 1; 3 – Sa 2; 4 – Sa 3 (see Table 8.2 for preparation grades) 340 8 Surface Quality Aspects if rust grade “C” was present, an increase in the surface preparation grade could substantially extend coating lifetime. Initial conditions are designated in several visual standards (see Table 8.1). These conditions can be subdivided into two groups: (1) rusty steel; (2) primers or coatings. The initial steel grades apply to uncoated steel surfaces that are deteriorated due to corrosion (either thermal or atmospheric oxidation). Rust grades are illustrated in Fig 8.3. These rust grades are defined as follows: r Steel grade A: Steel surface largely covered with adhering mill scale but little, if any, rust; r Steel grade B: Steel surface which has begun to rust and from which the mill scale has begun to flake; r Steel grade C: Steel surface on which the mill scale has rusted away from which it can be scraped, but with slight pitting visible under normal vision; r Steel grade D: Steel surface on which the mill scale has rusted away and on which general pitting is visible under normal vision. Previously coated steel surfaces are designated in ISO 8501-2. There are a large number of possible systems and coating conditions. The cleaning results do not depend only on the intensity of cleaning in these cases, but also essentially on type, thickness and adhesion of the coating systems, and on earlier surface preparation steps. For these reasons, only analogous applications to real cases can usually be derived. The coated steel surfaces considered in visual blast cleaning standards in- clude the following coating/primer systems and conditions: r coatings: – iron oxide shop primer; – corrosion protection system; rust grade A (a) (b) (c) (d) rust grade C rust grade D rust grade B mill scale steel rust Fig. 8.3 Initial surface conditions according to ISO 8501-1 (adapted from Kjernsmo et al., 2003). (a) Rust grade “A”; (b) Rust grade “B”; (c) Rust grade “C”; (d) Rust grade “D” 8.2 Visual Cleanliness 341 – a sound coating; – an unsuitable coating; r conditions: – upper side of a hatch cover; – upper side of a steel girder; – new construction work: tubes in a power station. Other designations for coated substrate surfaces apply to coating failure schemes, mainly according to the degree of rusting as defined in ISO 4648-3. These particular cases will be discussed in Sect. 9.1. 8.2.3 Preparation Grades Effects of blast cleaning preparation grades on the performance of a metal-sprayed coating are shown in Fig. 8.4. It can be seen from the graphs that pull-off strength increased if the preparation grade increased. The curves for iron grit and silica sand showed almost equal linear trends, whereas the trend was different if copper slag was used as a blast cleaning medium. One reason for the different trend of copper slag could be the high friability of copper slag. Because the very thorough Fig. 8.4 Effects of blast cleaning preparation grades on pull-off strength of arc-sprayed aluminium (Bardal et al., 1973) 342 8 Surface Quality Aspects preparation grade Sa 3 required a very intense material treatment, slag particles were fractured. The fracture debris did not effectively contribute to the material removal process. The surface texture was changed, leading to a deteriorated bond between substrate and coating, and the initially sharply rising curve started to drop. Preparation grades as defined in ISO 8501-1 indicate the following two designa- tions: r method of surface preparation; r degree of cleaning. The preparation method dry blast cleaning is designated by the letters “Sa” throughout the standard. Degrees of cleaning range from “1” to “3”. These degrees are defined in Table 8.2. The preparation grades are defined by written descriptions of the surface appearance after the blast cleaning operation, which are also provided in Table 8.2, together with representative photographic examples. Cleaning degrees are defined according to the presence of visible contaminants. These visible contaminants include the following substances: r rust; r previously existing coatings; r mill scale; r foreign matter. Table 8.2 Blast cleaning preparation grades and cleaning degrees (ISO 8501-1) Preparation grade Designation Description of surface (when viewed without magnification) Preparation method Cleaning degree Sa 1 Light blast cleaning The surface shall be free from visible oil, grease, dirt, dust, and from poorly adhering a mill scale, rust, paint coatings and foreign matter. Sa 2 Thorough blast cleaning The surface shall be free from visible oil, grease and dirt, and from most of the mill scale, rust paint coatings and foreign matter. Any residual contamination shall be firmly adhering. Sa 2 1 / 2 Very thorough blast cleaning The surface shall be free from visible oil, grease and dirt, and from mill scale, rust, paint coatings and foreign matter. Any remaining traces of contamination shall show only as slight stains in the form of spots or stripes. Sa 3 Blast cleaning to visually clean steel The surface shall be free from mill scale, rust, paint coatings and foreign matter. It shall have a uniform metallic colour. a Mill scale, rust or paint coating is considered to be poorly adhering if it can be removed by lifting with a blunt putty knife 8.2 Visual Cleanliness 343 Table 8.3 Cleaning degrees for different blast cleaning designations; see also Table 8.2 Standard Preparation method Cleaning degree ISO 8501-1 (Sa) Sa 1 2 2 1 / 2 3 ISO 8501-2 (Sa) P Sa – 2 2 1 / 2 3 SSPC – SP 7 SP 6 SP 10 SP 5 NACE – 4 3 2 1 NACE wet blasting WAB – 6 10 – International Slurry Blasting Standard a SB – 2 2 1 / 2 – a Issued by International Paint, Newcastle The highest cleaning degree always requires that the surface shall be free of all these matters, and it shall have a metal finish. The term “foreign matter” may include larger amount of water-soluble salts and welding residue. Comparative degrees of cleaning as defined for other surface preparation methods than dry blast cleaning are listed in Table 8.3. The achievable preparation grade depends on a number of param- eters, namely air pressure, abrasive type and abrasive particle size. Effects of abra- sive types were already investigated in an early study conducted by Nieth (1955). This author related the cleanliness to the capability of impinging abrasive particles to deform the substrate. Heavily deformed substrates, which are characterised by folded and bended surface sections, often contained traces of rust and mill scale after blast cleaning. With respect to the abrasive particle size, a comparative in- vestigation of Snyder and Beuthin (1989) has shown that coarse as well as very fine silica sand generated high surface preparation grades (90–95% cleaned to Sa 3; 5–10% cleaned to Sa 2 1 / 2 ), whereas medium and fine silica sand generated lower preparation grades (75% cleaned to Sa 3; 25% cleaned to Sa 2 1 / 2 ). If copper slag and coal slag were used, the highest preparation grade (Sa 3) could not be achieved. Copper slag delivered 10% cleaned to Sa 2 1 / 2 and 90% cleaned to Sa 2, whereas coal slag delivered 75% cleaned to Sa 2 1 / 2 and 25% cleaned to Sa 2. An early approach to replace the rather imprecise and subjective visual assess- ment through a physically founded parameter was due to Bullett and Dasgupta (1969). The parameter “reflectivity” of a steel surface, measured with an optical method, was applied for the assessment of the surface cleanliness. Results of re- flectivity measurements are shown in Fig. 8.5. Reflectivity increased if treatment time increased, and the progress of the function was very pronounced in the range of short treatment times. A reflectivity value of 360 corresponded to a visually estimated preparation grade Sa 2 1 / 2 . If this preparation grade was achieved, fur- ther blast cleaning action delivered a marginal increase in reflectivity only. The value of maximum reflectivity may vary from steel to steel, or with different abrasive materials, but for each combination there is a rapid initial increase with time of blast cleaning towards the asymptotic value. Bullett and Dasgupta (1969) could also show that reflectivity values dropped if a blast cleaned steel was ex- posed to an open environment. A reflectivity value of 60%, measured immediately after blast cleaning, dropped down to a value of 10% after 24h of exposure. Apps (1969) applied a reflectance meter for substrate cleanliness assessment, and 344 8 Surface Quality Aspects Fig. 8.5 Reflectivity of a blast cleaned steel surface as a function of treatment time (Plaster, 1973) he investigated the effects of different process parameters on reflectivity. He found that reflectivity dropped with an increase in blasting angle. This author could also show that reflectivity depended on abrasive type and condition. Reflectiv- ity values were high for new and for well-worn chilled iron grit, and they were low for worn, dusty grit. Blasting pressure and stand-off distance did not have any significant effect on reflectivity. Bardal (1973) also made an approach to ex- press visual cleanliness in terms of reflectivity. Reflectivity was defined in his study as the percentage of reflectivity of a base sample, which was a light grey tile. Thus, the unit of reflectivity was percent (%). Some results are displayed in Table 8.4. One striking result is that reflectivity was not a feature of cleanliness standard alone, but also depended on abrasive type. For a cleaning standard of Sa 3, for example, reflectivity was only about 60% for copper slag, but it was about 80% for silica sand. The explanation for this result was the dark colour of the almost black copper slag. Slag debris embedded into the substrate surface Table 8.4 Effects of surface preparation grade and abrasive type on reflectivity (Bardal, 1973) Preparation grade Reflectivity in % Copper slag Silica sand Sa 2 40 60 Sa 2 1 / 2 50 75 Sa 3 60 80 8.2 Visual Cleanliness 345 Fig. 8.6 Relationship between reflectivity, abrasive type and pull-off strength for metal-sprayed coatings (Bardal, 1973). Left: Copper slag (d P = 300−2, 500 μm); right: Silica sand (d P = 600− 1, 500 μm) reduced reflectivity (see Fig. 8.7 and Sect. 8.5) for more information on this is- sue). It could, however, be shown that reflectivity had a unique relationship to the pull-off strength values for sprayed metal coatings. Results of these investi- gations are exhibited in Fig. 8.6. The graphs in the figure verify that reflectiv- ity alone is not a sufficient measure of the coating carrying capability of a blast cleaned substrate; abrasive type is another considerable parameter. For equal val- ues of reflectivity of 60%, the substrate prepared with copper slag (left graph) delivered a pull-off strength of about σ A = 13 MPa, whereas the substrate pre- pared with silica sand (right graph) delivered a pull-off strength of about σ A = 9 MPa only. Hochweber (1971) introduced a method for the assessment of surface preparation grades which deployed the differences in the electric resistances between metal sub- strate and a measurement zone. This parameter, denoted transition resistance, was very low if the substrate featured a high degree of cleanliness. The author found the following relationships between surface preparation grade and transition resistance: Sa 1 = 1.0 ⍀;Sa1toSa2= 0.5 ⍀;Sa2= 0.1 ⍀;Sa2 1 / 2 and Sa 3 = 0.05 ⍀. More recently, Terrat and Boissel (1995) developed a method for the assessment of the cleanliness of metal substrates based on the measurements of the electrical 346 8 Surface Quality Aspects surface potential decay. This method proved to be capable to distinguish between clean metal, oxidation products and oil layers. 8.2.4 Special Remarks The following factors can influence the result of the visual assessment (ISO 8501-1): r initial state of the steel surface other than any of the standard rust grades A to D; r colour of the steel itself; r regions of different roughness; r surface irregularities, such as dents; r marks from tools; r uneven lighting; r shadowing of the surface profile caused by angled projection of abrasive; r embedded abrasives. The latter aspect is considered through a supplement to ISO 8501-1, which il- lustrates the differences in surface appearance, including colour, which are obtained when the same surface is prepared by blast cleaning with different abrasive mate- rials to the same preparation grade. The situation is shown in Fig. 8.7. This situa- tion is further illustrated by the results of reflectivity measurements as provided in Table 8.4. For previously coated surfaces, which have been prepared for re-painting, only photographs with rust grades “C” and “D” should be used for visual assessment. The surfaces should be examined either in good diffuse daylight or in equivalent artificial illumination. For rust grades, the worst grade is evident. Some recent developments revealed the use of vision sensing systems (Carew et al., 2001; Chen and Chang, 2006) and image analysing methods (Gupta et al., 2003; Trujillo and Sadki, 2004; Greverath et al., 2005) for the assessment of blast cleaned steel substrates. Two examples are shown in Figs. 8.8. and 8.9. The graph in Fig. 8.8 shows a pixel histogram for a steel substrate, which was prepared according to a surface preparation grade Sa 3 (bare metal). The dimension for the x-axis was the hue-value of the colour. The rather continuous signal characteristics without a dominating peak, characterised a clean substrate without discolouration. The re- sults shown in Fig. 8.9, where differences in peak location and peak height for two close surface preparation grades in a green-colour histogram could be recognised, showed that colour measurements have promise for detecting subtle differences in blast cleaned surfaces. Software solutions are under development for the automatic detection and quality assessment of blast cleaned substrates (Gupta et al., 2003). An example is provided in Fig. 8.10. [...]... methods (Forsgren and Applegren, 20 00) Method Chloride level in μg/cm2 Bresle (10 min) SSM (10 s)a SSM (10 min)a 44 .8 54. 8 15 .2 24. 8 47 .5 72. 8 * * 61.3 96.3 * * Wet blast cleaning 1.6 1.6 0 3 .2 1 .4 0.7 1.7 1.5 2. 7 2. 0 3.1 4. 1 Hydroblasting 1.6 0.8 0 1 .2 2 .4 1 .2 0 15 .2 1.8 2. 4 0.1 4. 8 0 0 – 4 .2 4. 6 2. 1 10.3 1.0 0.8 Wire brush 28 .8 16.0 23 .2 17.6 63.5 32. 6 15 .2 18.1 – 58.9 25 .0 30.3 Needle gun 27 .6 21 .2. .. Na2 SO4 None NaCl Na2 SO4 None NaCl Na2 SO4 0 5 /2 3 /2 0 5 /2 0 0 5 /2 4 /2 0 5 /2 0 2/ 2 5 /2 3 /2 Ri 0 Ri 0 Ri 0 Ri 0 Ri 1 Ri 1 Ri 0 Ri 0 Ri 0 Ri 0 Ri 1 Ri 0 Ri 0 Ri 2- 3 Ri 0 1.3 ± 0.6 1.1 ± 0.6 1 .4 ± 0 .2 3.7 ± 1 .4 3 .2 ± 1.0 4. 0 ± 0.8 1.0 ± 0.6 1.9 ± 0.6 1.9 ± 0.6 4 .2 ± 1.0 3.6 ± 1 .2 4. 3 ± 1.5 2. 1 ± 0 .4 2. 0 ± 0.9 2. 9 ± 0 .4 A/B10, B/C90 A/B100 A/B40, B/C60 A/B50, C50 A/B70, C30 A/B20, C80 B/C100 A/B100 A/B40,... tz, 20 01) u Adhesionc Failure type Coating DFT in μm Rust Blisteringa Rustb degree strength MPa contamination Noverox + 20 4 ± 22 HS epoxy 21 3 ± 17 199 ± 20 Noverox + 21 3 ± 24 DS-mica 22 1 ± 16 20 8 ± 21 Antitrust + 173 ± 15 HS epoxy 1 92 ± 18 1 82 ± 21 Antitrust + 181 ± 17 DS-mica 178 ± 9 169 ± 11 Noxyde 148 ± 15 136 ± 11 151 ± 10 a ISO 4 628 -2 ISO 4 628 -3 c ISO 4 6 24 b None NaCl Na2 SO4 None NaCl Na2 SO4 None... 1960; Doherty, 19 74; Fairfull and Weldon, 20 01); secondary electron mode of SEM (Fairfull and Weldon, 20 01; Momber et al., 20 02a, 20 04) ; back-scattered mode of SEM (Amada et al., 1999; Momber et al., 20 02a, 20 04) ; EDXA plots from SEM imaging (Momber and Wong, 20 05a; Momber et al., 20 02a; Possart et al., 20 02) Two images taken with the secondary electron mode of an SEM are provided in Fig 8 .20 This imaging... through comparative metallographic studies by Kaiser and Sch¨ tz (20 01) on clean and contaminated, respectively, steel samples u (see Fig 8. 12) The authors defined three levels of salt contamination: (1) NaCl: 56 >56 >56 4 >56 >56 >56 >56 10 >56 >56 >56 >56 >56 Epoxy, DFT 23 9 μm 0.6 1 .4 3.9 5.3 7.6 36 3 1.5 1.5 1.5 11 7 3 3 1.5 12 12 7 7 3 >56 >56 >56 >56 5 >56 >56 >56 >56 3 Epoxy novolac, DFT 26 2 μm 0.6 1 .4 3.9 5.3 7.6 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 >56 Epoxy, DFT 25 2 μm 0.6 1 .4 3.9 5.3 7.6 1.5... pipe H2 S scrubber plate Heat exchanger shell Sulphates Phosphates Chlorides Nitrates Sulphates Phosphates Chlorides Nitrates Sulphates Phosphates Chlorides Nitrates Sulphates Phosphates Chlorides Nitrates Sulphates Phosphates Chlorides Nitrates Sulphates Phosphates Chlorides Nitrates Dry blast cleaned 40 0 2 0 5 0 4 0 5 1 28 6 8 0 6 4 39 0 12 0 7 0 17 0 3 0 2 6 5 1 3 11 2 2 32 1 4 2 1 2 7 0 8 1 4 0... Appleman, 20 02) Coating system DFT in μm Salt thresholds in μg/cm2 Chloride (Cl) Sulphate (SO4 ) Unknown 125 22 5 25 –35 7–11 Thin films 100–150 60–190 130–180 One coat Three coats 25 4 – – Two coats 7–30 >1 6 25 7 – 6–30 5–10 1 5 50 . Na 2 SO 4 0Ri 04. 3± 1.5 A/B20, C80 Noxyde 148 ± 15 None 2/ 2 Ri 0 2. 1 ± 0 .4 A/B80, B20 136 ± 11 NaCl 5 /2 Ri 2- 3 2. 0 ± 0.9 A/B80, B20 151 ± 10 Na 2 SO 4 3 /2 Ri 0 2. 9 ± 0 .4 A/B90, B10 a ISO 4 628 -2 b ISO. A/B100 199 ± 20 Na 2 SO 4 3 /2 Ri 0 1 .4 ± 0 .2 A/B40, B/C60 Noverox + 21 3 ± 24 None 0 Ri 0 3.7 ± 1 .4 A/B50, C50 DS-mica 22 1 ± 16 NaCl 5 /2 Ri 1 3 .2 ± 1.0 A/B70, C30 20 8 ± 21 Na 2 SO 4 0Ri 14. 0± 0.8 A/B20,. 1 92 ± 18 NaCl 5 /2 Ri 0 1.9 ± 0.6 A/B100 1 82 ± 21 Na 2 SO 4 4 /2 Ri 0 1.9 ± 0.6 A/B40, B/C60 Antitrust + 181 ± 17 None 0 Ri 0 4 .2 ± 1.0 A/B20, C80 DS-mica 178 ± 9NaCl 5 /2 Ri1 3.6± 1 .2 A/B80, C20 169