关键词: 层流冷却, 卷取温度, 案例推理, 混合智能, 前馈补偿, 反馈补偿

Abstract: It is hard to control the coiling temperature in a certain range since the existing laminar cooling process lacks the direct feedback during strip hot rolling. A hybrid intelligent control method was therefore proposed combining the mechanism models with case-based reasoning technology, i.e., composed of four modeling modules: to preset the number of valves of spray header to be opened in cooling zone, to forecast coiling temperature, to compensate for feedforward, to compensate for feedback. The method proposed was simulated with the actual operating data provided by a steel plant and the results showed that the setting value of the number of spray header valves to be opened can be readjusted automatically and timely according to the changing operation conditions, thus controlling the coiling temperature in the range the technical schedule requires so as to improve the microstructure and mechanical properties of the rolled strip.