East China University of Sience and Technology
National Key R&D Program of China(2016YFB0303403) and National Natural Science Fund for Distinguished Young Scholars (61725301，61925305)
Rotary Kiln is the core thermal reaction equipment of cement calcination process, whose operation state is closely related to the yield, the energy consumption and the quality of products. Contact temperature measurement can not be installed in the core area inside the kiln due to high temperature and continuous rotation. Rotary Kiln Shell Scanning System(RKSSS) is available to monitor the temperature of kiln shell and reflect the thermal condition inside the kiln indirectly in real time. A new method of abnormal conditions identification which is based on 2D-OTSU-Edge-Detection is proposed in this paper. Using the relative entropy, such method figures out weight coefficient of the model, and thus constructs a fusion model consists of gray gradient information and local gray standard deviation information. At the same time, 1D-OSTU pre-searching strategy is adopted to improve the efficiency of the algorithm. By applying images from the RKSSS, the proposed method is tested and compared with two 1D-OTSU methods. The results demonstrate that the proposed method made a promotion of 7.14% and 34.83% in detection rate, and 62.4% and 70.9% in false alarm probability. Therefore, the proposed method is available to detect the abnormal condition and to extend operation cycle of the rotary kiln.