IMRP Abstract: BVN - Improving the method for process release

Reference point routine process monitoring in radiation processing relies on an average reference ratio estimator1 to predict a minimum and/or maximum product dose value. The average reference ratio estimator is a point estimate and is applied as multiplicative correction to the observed value of the reference point location. This method of estimating a product minimum and/or maximum dose has been shown to be inaccurate and will randomly be positively biased and negatively biased2. The random nature of bias and the actual magnitude of the bias even within a 1-sigma level renders this method untenable for assessing conformance of product dose to minimum and maximum product dose specifications with surety. The following presentation shall discuss the bivariate model that provides forecasting of product minimum and/or maximum dose from reference point monitoring data that mitigates the random bias of current industry practice.

1  ISO/ASTM 523023-2015, Guide for absorbed-dose mapping in radiation processing facilities

2  A Statistical Model of Routine Process Monitoring Variates and Conformance Assessment for Radiation Processing, B. Lundahl, Radiation Physics and Chemistry, Volume 205, 2023, 110640, ISSN 0969-806X