Prediction of unusual plasma discharge by using Support Vector Machine

•It is shown that unusual visible light emission inside the plasma vessel can be predicted by using Support Vector Machine (SVM), a machine learning method.•The probability of the unusual emission is obtained by taking mean values of the probability values of several frames in a video.•199 unusual emission videos and 254 videos without unusual emissions are prepared.•The prediction accuracy rate attains to 96.4%.•The unusual visible light emission may be able to be predicted around 0.3 s before the beginning of an unusual emission. This paper proposes a method for predicting an unusual emission of visible light inside the plasma vessel by using a Support Vector Machine (SVM) because the unusual emission of visible light can be caused by unexpected heating on the vessel surface. This emission must be predicted to avoid unexpected situations in which it causes some damage to the vessel. The light reflected from the divertor tiles is used as the unusual emission light. This study aims to predict such unusual emission through pictures before the start of the unusual emission, regardless of the plasma physics. This study experimentally confirms that the unusual emission of visible light inside the plasma vessel can be predicted with an accuracy rate of 96.4%, and approximately 0.3 s before the start of an unusual emission.