中国物理B ›› 2007, Vol. 16 ›› Issue (12): 3738-3741.doi: 10.1088/1009-1963/16/12/030
Sanuki H1, Itoh K1, 王 灏2, 王爱科2, 杨青巍2, 丁玄同2, 董家齐2
Wang Hao(王灏)a)† , Wang Ai-Ke(王爱科)a), Yang Qing-Wei(杨青巍)a), Ding Xuan-Tong(丁玄同)a), Dong Jia-Qi(董家齐)a), Sanuki Hb), and Itoh Kb)
摘要: Artificial neural networks are trained to forecast the plasma disruption in HL-2A tokamak. Optimized network architecture is obtained. Saliency analysis is made to assess the relative importance of different diagnostic signals as network input. The trained networks can successfully detect the disruptive pulses of HL-2A tokamak. The results obtained show the possibility of developing a neural network predictor that intervenes well in advance for avoiding plasma disruption or mitigating its effects.
中图分类号: (Tokamaks, spherical tokamaks)