Takahiro Nishigaki, Takushoku University, Japan

This paper examines a method for estimating operation stoppage time of crushing operations in a recycling plant for agricultural plastics in Japan.In light of the growing environmental problems caused by waste plastics around the world, resource recycling of plastics is being promoted.Agricultural plastics are plastics discharged from the agricultural sector. Agricultural plastics have a high recycling rate because the main materials are often the same, even if the manufacturers are different.Companies scattered in rural areas are mainly engaged in the recycling and processing of agricultural plastics, but many of them face a social problem in rural areas: a shortage of workers.In order to address the above issues, there is a need for automation of work in recycling plants, such that machines can replace workers.In this study, we aimed to automate the operation of crushing work in a recycling plant for agricultural plastics.The crushing work is the crushing of agricultural plastics fed into a crusher by rotating blades inside the crusher.Work stoppage is determined by monitoring a current meter that indicates the resistance value of the rotating blades and a camera inside the crusher.In crushing operations, there is a demand for automation of this operation stoppage judgment. In a related study, as a preliminary step to estimating the work stoppage time, we estimated the time period when the current value decreases and the variability decreases, which is indicated in the current value before the operation stoppage.The time period before operation stoppage was defined as the time period from 150 seconds before to the work stoppage time of the crushing operation, and the estimation of the operation stoppage phase was conducted with the name “end of operation phase”.

Automation, Outlier Detection, Anomaly Detection, Predictive Failure.