作者: Kiyoshi Tamayama , Kenshiu Watanabe
DOI:
关键词:
摘要: A plant malfunction diagnostic method is characterized by determining simulation a change in state variable, forming pattern among variables obtained autoregressive analysis of the inserting formed neural network, performing learning until preset precision obtained, and identifying cause inserting, which indicates data gathered from plant. This makes possible early identification malfunction. Plant rate operation safety are improved allowing operator to perform appropriate recovery with sufficient time margin.