Realizing desired quality through gray-box models integrating first principles and statistical analysis

点击翻译按钮获取中文摘要

A gray-box model integrating a physical model and a statistical model was developed, and also locally weighted PLS that can cope with changes in process characteristics was developed. The gray-box model was applied to a control problem of molten steel temperature in a tundish in the steel industry. The locally weighted PLS was applied to various processes in pharmaceutical, semiconductor, and petrochemical industries. The achievement of this research was significant and awarded a prize from The Iron and Steel Institute of Japan ; further development and industrial applications are expected.