A new algorithm for identifying abnormal glow curves in thermoluminescence personal dosimetry.

In this paper an algorithm for the investigation of routine curves in thermoluminescence personal dosimetry is presented. In rare cases, the luminescence glow curves do not exhibit the expected form. They have an abnormal shape as a result of, for example, external contamination, hardware problems, and poor heat transfer. So, glow curves from a monthly exposure period are compared with regular glow curves. Each curve is divided into four regions of interest (ROIs) and the relationships between the different ROIs are analysed. There are few criteria combining all four ROIs, which are necessary to distinguish between normal and abnormal glow curves. For that, the numerical value and the channel of the curves maximum also need to be considered. In most cases an additional set of criteria permits the identification of the ROI in which the irregularity occurs.