A Novel Non-parametric Sequential Probability Ratio Test Method for Structural Condition Assessment

In this paper, a novel non-parametric Sequential Probability Ratio Test (SPRT) method based on Mann-Whitney rank sum test was proposed for structural condition assessment by utilizing long-term structural health monitoring data. Compared with the fixed sample size test, the sequential probability ratio test has many advantages and is widely used in hypothesis testing. When using the SPRT method in a process of hypothesis testing, a probability distribution function of the samples is need to be assumed, for examples, as normal or exponential probability distribution function. However, the actual probability distribution function of the samples was unknown or could not be expressed as a simple distribution function on occasion. Assuming that the samples of the normal condition were known, the log-likelihood ratio based on Mann-Whitney rank sum of the samples of undetermined condition was calculated accurately and the decision was made by comparing with the thresholds. This method did not require knowing the probability distribution function of the samples, so it could be applied to the conditions with different probability distribution functions. The method was validated by a numerical example. The results showed that this proposed method was effective to distinguish different conditions and was better than other non-parametric sequential probability ratio test method.