ISO 9223:2012 pdf download – Corrosion of metals and alloys一 Corrosivity of atmospheres一 Classification, determination and estimation.
A.1 General
The corrosivity of atmospheric environments shall be ctassified either through determination of the corrosivity category based on the exposure of specimens or by estimation based on environmental parameters and the use of a dose-response function. The use of these two different denominations for the assessment of a corrosivity category implies that two different levels of uncertainty can be expected fr the determination (low uncertainty) and estimation (high uncertainty) procedures. This annex serves to establish those two levels of uncertainty.
The basis for the details given in this annex is a separate statistical analysis performed In connection with the derivation of the dose-response functions for the estimation procedure.
A.2 Distribution of errors
Log-normal distributions. i.e. normal distributions of logarithmic values, are observed for the corrosion rates. If the uncertainty is expressed by a standard deviation of logarithmic values, , then
This means that the uncertainty interval in general is asymmetric and can be expressed as However, if .c is small, the interval becomes approximately sm,metflc. This can be illustrated by the following two examples. If s • 0,7, then w 2 and e-’ 1/2, which corresponds to an interval from 50% to +100%. On the other hand, ifs z 0,01. then e’ 1,01 and e 0.99, which corresponds to an interval from -1 % to +1 % or ii %.
A.3 Levels of uncertainty
Table A. 1 gives the estimated levels of uncertainty. The large difference between the two procedures is clear from the table arid this is what motivates the two different denominations of the assessment. A.4 gives a description of possible sources 01’ en-or and an explanation of those errors is included in the values given in Table A.l.
A.4 Sources of uncertainty
Regarding the determination of corrosivity categories based on exposure of specimens, it should first be clarified that the levels given in Table A. I are expressed for an average value calculated from three individual values and not for an individual corrosion value.
The uncertainty levels given in Table Al. both for determination and estimation, are based on the exposure of materials at many different test sites, but only for one exposure penod. Thus, the values should have a general validity, but the variation In corrosion attack that can occur from year to year, depending on natural variations in climate, is not included in the values in Table A.1.
Regarding the estimation of corrosivity categories based on the dose4esponse functions, the total uncertainty consists of two parts. uncartainty in the dose-response functions and uncertainty in the measurements of environmental parameters. Of these, the uncertainty in the dose-response functions is dominant, Also, the values in Table A.I are based on an average uncertainty over the range of parameters used in the function. As for all regression functions, this uncertainty is lowest in the middle range, corresponding to corrosivity category C3, and higher at the lower and upper ranges, correspondmg to corrosivity categories Cl and C5. Uncertainty for category CX is the highest and it is not covered by this calculation.
It is necessary for the classification approach for informative corrosivity estimation to be simple and use-friendly as regards the choice of parameters to be considered. For the purposes of this Intemational Standard, the key factors of the atmosphere for corrosion of metals and aloys are the temperature-humidity complex, and sulfur dioxide and chloride pollution levels.
For unsheltered positions, the corrosion impact is mentioned in terms of dry and wet deposition. Wet deposition includes transport by means of precipitation and dry deposition denotes transport by any other process.