For nearly 15 years the environmental test results from long-term exposures have been handled by a computer program. Over 1 14-million bits of information representing data from more than 250 metals and alloys on test at fifteen atmospheric and two seawater locations are included in the program. Calculations are made reflecting aging changes, and changes due to the cleaning procedure and to corrosion. A printout then gives corrosion rates and mechanical property changes for each time period. Data points include 1,2,7, and 20 years for atmospheric specimens and 1, 2, 5, 10, and 20 years for seawater tests. Exposures of new alloys are made each year as phases of the continuing program along with appropriate removals of exposures from earlier series.

Since at least three data points are now on hand from some test metals, it is feasible to predict both corrosion rates and extreme pitting values for future removals. The corrosion rates and pitting data are subjected to curve fitting techniques through a computer operation. Several models were considered. However, based on earlier published studies on aluminum, we decided to begin with a least-squares fit of the cube-root pitting equation, d = kt^1/3. Future work with other models may indicate improved reliability.

It is now possible to predict the 20-year corrosion rate for a marine environment and to estimate the time to perforation for a given alloy system. Actual 20-year data, when available, will be compared with these predicted values. Also included in the program is the capability to predict 10-year results when shorter-term data (1,2, and 5 years) are on hand.