화학공학소재연구정보센터
Chemical Engineering Communications, Vol.130, 251-264, 1994
Knowledge Extraction in Chemical Process-Control
Direct access to the control knowledge of a skilled process operator is difficult and time-consuming. We present a framework for extracting process knowledge from human process operations records, using automated data reduction techniques for low-level data analysis and presenting reduced data to a human knowledge engineer for interpretation. We illustrate on a simple neutralization in a continuous stirred-tank reactor and employ a neural network for data reduction.