W. Stephen tait, ph.D. Chief Science Officer & principal Consultant, pair O Docs professionals, LLC Corrosion Corner Spray Package Corrosion Risk—Part II similar information is not available for your specific productpackage systems. So how does one read Figure 1? Let’s start with the curve for storage corrosion tests. I’ll provide examples for a) no data (zero test length), b) after three months of storage testing and c) after one year of storage testing. The estimated risks for these three times are: • 62% risk when the time = zero (the no data risk) • 30% risk (or 60% confidence) after three months of testing • 7% risk (or 93% confidence) after one year of testing Thus, Figure 1 illustrates that the corrosion risk decreases as the storage test length increases. Figure 1 also illustrates that an approximately 7% risk is obtained after one year of storage testing. A word of caution about Figure 1 for storage tests: the risk curve is only valid for a correctly designed storage test. For example, using Hello, everyone. Last month I began a detailed discussion on a figure that contains the risk for conventional storage corrosion tests as a function of test time and risks associated with electrochemical corrosion tests. This discussion included: • How the storage test risk-curve was generated • How the correlations (risks) were generated for electrochemical corrosion tests • The types of containers from which the data were obtained • The number of data used to generate the risk for each type of test This month completes the discussion of the risks associated with corrosion tests. Figure 1 has three types of risk data: a corrosion storage test risk-versustime curve for traditional metal aerosol containers and the two correlations between corrosion predictions from electrochemical data and the corresponding actual commercial spray package data for all types of spray packages. The correlations are the square and diamond shapes in Figure 1. The data in Figure 1 are empirical probabilities calculated from actual corrosion data. Empirical probabilities are different from the corresponding statistical probabilities calculated from the theoretical normal distribution. Empirical corrosion-correlations generated with a large number of data are often closer estimations for actual corrosion than the corresponding statistical probabilities. The curve and correlations in Figure 1 are aggregates that only provide the generic risks for spray packaging localized corrosion (e.g., metal pitting and coating blisters). Consequently, the information in Figure 1 is useful for estimating corrosion risks for new products and derivative products (line extensions) when 22 Spray February 2017 Figure 1: The estimated risk of localized corrosion as a function of test length.
Spray February 2017
To see the actual publication please follow the link above