Optimizing the sensitivity of the small-disc creep test to damage and test conditions

Abstract

The small-creep disc test is seen as a promising solution to the problem of sampling from in-service components for remanent life estimation. However, experimental studies have revealed substantial scatter in failure times resulting from variations in test and apparatus geometries. These studies therefore suggest that there exists a set of conditions that both minimizes the scatter and also maximizes the sensitivity of the disc test to the determination of remanent life, so enabling reliable estimates of the remaining life to be made. The objective of this paper is to identify such an optimum. The large scatter present in small-disc test data would make the identification of this optimum problematic and inconclusive using an experimental approach. Instead this paper uses a numerical model of the disc test to predict failure times over a wide range of test conditions. The resulting response surface is then approximated using a polynomial and from this model the optimum set of test conditions is identified. Full verification of the model (and the optimum) can then be more successfully accomplished by carrying out experiments close to the optimum conditions where scatter is the smallest. The model was shown to be capable of predicting actual experimental creep test results. Using this approach, failure times were found to be most sensitive to disc thickness and hole diameter but least sensitive to disc diameter. Under clamping, the optimum test geometry was insensitive to the applied load but sensitive to the level of damage. Without clamping, the optimum geometry was insensitive to both of these. The paper gives optimum conditions for various levels of damage under both conditions.