Machine learning sintering density prediction model for MOX fuel pellet

Uranium and plutonium mixed oxide (MOX) pellets used as fast reactor fuels have been produced from several raw materials by mechanical blending through various processes, such as ball milling, additive blending, granulation, pressing, and sintering. It is essential to control the pellet density, which is one of the important fuel specifications, but it is difficult to understand the relationships among many parameters in the production of MOX pellets. The database for the production of MOX pellets was prepared from production results in Japan, and input data of eighteen types were chosen from the production process to form a data set. A machine learning model for predicting the sintered density of MOX pellets was derived using a gradient boosting regressor and could represent the sintered density of MOX pellets with R² = 0.996 as a parameter that affects production conditions, such as the type of raw material used and sintering temperature. (author)