A preliminary study on object recognition inside the primary containment vessel at Fukushima Daiichi Nuclear Power Plant

Abstract

Currently, decommissioning preparations have been underway at the Fukushima Daini Nuclear Power Plant. It is expected to take a long time because the situation inside the reactor has not been accurately assessed. An investigation inside the Unit 2 in 2018 showed that nuclear fuel debris had melted out over a wide area inside the reactor. It was also found that the debris can be classified into several types according to shape, surface texture, and other factors. Therefore, this paper attempts to develop an AI system that can identify sediments inside PCV applying an object detection algorithm.