Fast detection method for radionuclides based on Sequential Bayesian approach

Abstract

BackgroundAppropriate fast detection method for radionuclides is necessary for customs radioactive security inspection with high flow rates and low counting rate levels. Compared with the traditional uncertainty analysis method, the Bayesian method and the Sequential Probability Ratio Test can fully utilize all the information of measured physical quantities and reduce the required sample size.PurposeThis study aims to develop a new method to solve the problem of the fast and accurate detection of radionuclides at low radioactive counting rate scenarios.MethodsA new Sequential Bayesian fast detection method for radionuclides was proposed on the basis of binary hypothesis H0 (no radionuclides) and H1 (radionuclides). Based on the principle that the time interval between adjacent two rays was exponential distributed, the decision was caculated by collecting a series of ray time samples in chronological sequence, and decisions were made by comparing the decision function with the preset upper and lower thresholds. Finally, experimental verifications were conducted on the feasibility, detection performance, and universality of the method by placing a set of standard point sources at different distances from the front of a LaBr3(Ce) detection system in both low and natural radiation background environments. The effects of the key parameters of the method on the detection performance were investigated.ResultsUnder the absence of radionuclides in two type of background radiation environments, the average detection time for background radiation by this method is 24.08 s and 10.54 s, with an average detection sample size of 1 427 and 1 742, respectively. Under the presence of radionuclides in two type of background radiation environments, the lower limits of detection sensitivity of experimental measurements are 8.2% and 6.1%, respectively, the corresponding average detection times were only 8.59 s and 6.61 s respectively, and the experimental measurement false negative rates were all zero.ConclusionsResults of this study verify that above proposed method is very suitable for fast detection of low-level radionuclides.

Keywords

• sequential bayesian method
• fast detection for radionuclides
• binary hypothesis
• lower limit of detection sensitivity