Corrosion test improvement by data driven climatic modeling. Issue 12 (5th December 2016)

• Record Type: Journal Article
• Authors: Mayrhofer, R.; Bäck, T.
• Is Part Of: Materials and corrosion. Volume 68:Issue 12(2017)
• Page Start: 1338
• Page End: 1342
• Publication Date: 2016-12-05
• Subjects: corrosion test -- data driven modeling -- field data -- optimization -- predictive analytics; Materials -- Periodicals; Metals -- Periodicals; Corrosion and anti-corrosives -- Periodicals; 620.1122305
• DOI: 10.1002/maco.201609288
• Languages: English
• ISSNs: 0947-5117

Abstract

Abstract : Many corrosion test experiments, using a range of testing methods and car body materials, are done by the automotive industry worldwide to meet the corrosion requirements in the field. In this paper, we use test data and climate field data for data driven modeling of the test results. One contribution reported here is to find the most influencing factors for corrosion test progress and field correlation. Therefore, iron and zinc panels with automotive paintings conducted in several corrosion tests are measured and transformed into suitable data. The corrosion parameters are defined as factors and the result of corrosion testing (i.e., scribe creepage values) as output variable. With this data, a statistical model is generated and validated. The best fit model is analyzed to find sensible testing parameters effecting the corrosion results. The same procedure is also applied for climatic field data, which is first transformed into characteristics such as humidity, duration of wetness, and salt load. The data from corrosion testing experiments are then mapped to the climatic load in the field. A match to the field correlation can be found, as demonstrated by our results. Abstract : Based on corrosion test data for standard automotive substrates, data driven modeling is applied to obtain models for predicting scribe creepage as a function of the corrosion test parameters, based on aggregated test characteristics such as temperature, humidity, salt load, freezing phase, and duration. Key parameters affecting iron‐zinc creepage ratio and iron creepage are identified, using the predictive models and the matching between corrosion tests and field climate zones is quantified.