Small sample test method for S-N and P-S-N curves

Based on the heteroscedastic regression analysis theory, a small sample test method for S-N and P-S-N curves is established, which can obtain S-N and P-S-N curves of metal materials by scattered test. All the test data of a S-N curve in different stress levels can be analyzed as an integration, so its information quantity is much bigger than traditional group test method. To the group test data, the presented method precision is also higher than the group test method. Therefore, by the presented method, P-S-N curves can be obtained without complementary test through treating the existed test data of S-N curves. Monte Carlo simulations and engineering applications show that comparing with traditional method in the same precision upwards of 60% specimens can be saved by this method.