Determination of the ultimate strength of corroded stiffened panels

Experimental tests were conducted to determine the buckling load of corroded stiffened panels, which represent parts of a ship's hull. The fabrication and welding of the stiffened panels were done in accordance with accepted ship building practices. The corrosion patterns were divided in four major groups: plate only; stiffeners only; weld area; and opposite stiffeners.; The experimental setting was conceived in such a way that the load is applied in the neutral pane of the stiffened panel. The stiffened panel is simply supported, and instrumented with strain gauges and LVDTs to measure the strains and deflections. Sechler's equation for stiffened panels is used as a basis to determine the buckling load of the stiffened panel. Two analytical variations to Sechler's equation, the average plate thickness (AT) and the effective width thickness (ET) were proposed to compensate for the corrosion.; The ET calculation is better in predicting the reduction in strength for the stiffener and weld as well as for the opposite stiffeners corrosion configurations. The AT is better in predicting the buckling load for the plate only corrosion. Neither the ET or AT approaches provided meaningful results for the weld area corrosion group.; Three ratios are proposed for comparative purposes. The first one RE/A represents the experimental over analytical buckling loads. The second RAnalytical , the analytical buckling load of a corroded stiffened panel over a non-corroded one, is predicting the anticipated reduction in strength. The last one RExp, the experimental buckling load of a corroded over non-corroded stiffened panel, is representing the actual reduction.; A benchmark RE/A for the non-corroded stiffened panels of 70% is found. The overall average RE/A for the ET approach is 68% with a small standard deviation of 5% while for the AT the average is 62% with a large standard deviation of 11%.; The results of the RAnalytical and RExp demonstrate that the ET approach is better in predicting the reduction in strength; the linear regression correlation coefficient for ET is 0.92 versus 0.53 for the AT approach.; A general conclusion on the impact on the buckling load of the stiffened panels is drawn for similar levels of corrosion and their location. The worst case is the stiffener group, followed by the opposite stiffeners, with the last two groups plate only and weld area having equally the least influence.