Evaluation of the radiological risks associated with the routine transport of radioactive material within Michigan.

Abstract

Radioactive materials play an important role in modern society. In addition to providing electrical power and supporting national defense, radioisotopes play significant roles in the fields of medicine, research, manufacturing, and industry. Since most of these materials are not manufactured or disposed of at the site where they are used, they must be transported between various processing, use, storage, and disposal facilities. This dissertation examines the mathematical model used to predict the collective dose to the population that resides along a potential transport route, commonly called the off-link dose. The currently accepted RADTRAN and RISKIND transient dose models are reviewed. Then three new individual transient dose models are derived by assuming that a point, line, or surface cylinder can approximate the actual transport package. Groundscatter effects were investigated using a Monte Carlo simulation of the surface cylinder model and found to contribute no more than 12% to the total individual dose from a passing shipment of radioactive material, thus not warranting explicit inclusion in the newly derived transient dose models. All five of the individual transient dose models were evaluated for representative shipments of spent nuclear fuel and low-level waste within the State of Michigan and compared to experimentally measured doses. The individual dose for the Michigan shipment scenarios was found to be on the order of 1 murem. Comparison to the experimental measurements revealed that RISKIND consistently predicts the best estimate of the measured dose, followed closely by the surface cylinder model. RADTRAN consistently over predicted the measured dose by at least a factor of two. Finally, the line dose model is integrated over strips of uniform population along the transport route to arrive at the collective off-link population dose. This off-link dose model was incorporated into an ArcView application using the Avenue scripting language. Then this script was used to investigate the off-link dose to Michigan residents for the previously mentioned representative transport scenarios. The off-link dose was found to be less than 3 person-rem for all of the scenarios investigated.