Research on Closed-Loop Control System Based on Image-Signal Processing of Furnace Flame

摘要

Through the furnace-flame image-signal processing for power plant, effective-temperature field proportions, high-temperature field proportions, centroid offset distances, and circular degrees in high-temperature field can be all obtained. What's more, based on the above data and related signals collected by sensors such as flame detectors as a criterion, the Kohonen's self-organizing neural network is introduced to distinguish the states of furnace flame. Therefore, the opening incremental adjustment is proposed to achieve real-time control of furnace flame.