A novel fault detection system taking into account uncertainties in the reconstructed signals

Controlled Morris method: A new factor screening approach empowered by a distribution-free sequential multiple testing procedure

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.

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

A Study of SPRT Algorithm and New-GUARD for Radiation Detection by Venkatachalam Palaniappan Dr. Shahram Latifi, Examination Committee Chair Professor o f Electrical and Computer Engineering University o f Nevada, Las Vegas A novel and efficient radiation detection algorithm combined with a measuring unit will produce an ideal detector to battle field radiation measurement problems. Studies of Sequential Probability Ratio Test (SPRT) for radiation detection are essential towards developing efficient and accurate radiation detection algorithms. In this study, the performance o f the classical Single-Threshold-Test (STT) and the SPRT First-In-FirstOut (FIFO) algorithms is considered. Next, improvements made by the Last-In-FirstElected-Last-Out (LIFELO) algorithm are analyzed. Further, enhancements to the LIFELO algoritlrm, using the Dynamic Background Updating and Maximum Likelihood Estimator (MLE), are perfoimed. The thesis also provides detailed requirements for an innovative hand-held radiation detection system and underlines additional features available on a New Generation User Adaptable Radiation Detector (New-GUARD) to help the field survey processes. Currently available technologies are studied to rationalize the need for the New-GUARD prototype. The New-GUARD is compared to similar products that are already available

/pdf/a-study-of-sprt-algorithm-and-new-guard-for-radiation-4tcpykejdf.pdf

A study of SPRT algorithm and New-Guard for radiation detection

An elevator fault diagnosis method based on Sequential Probability Ratio Test (SPRT) is proposed in this paper. In order to verify the effectiveness of the method, this paper designed the fault diagnosis experiment for elevator mechanical system. Firstly, the wavelet transformation is used to filter the noise of the vibration signal collected in the experiment. Then the kurtosis value of the filtered signal is extracted as the index to represent the practical status of the elevator. Finally, the SPRT algorithm is used to diagnose the faults of the elevator mechanical system. Experimental results show that this method has high accuracy in practical applications. It is important to improve the performance of the fault diagnosis for the elevator mechanical system.

/pdf/the-elevator-fault-diagnosis-method-based-on-sequential-5fdh9s8igw.pdf

Chenggang Yu

Papers

A non-parametric sequential rank-sum probability ratio test method for binary hypothesis testing