提出一种基于图像识别的二次仿真的方法。首先,介绍图像一运行工况一数值仿真结果之间的关系。其次,提出一种二次仿真系统的框架,并详细介绍二次仿真的数学模型、图像识别、聚类等模块。最后,以回转窑火焰图像及回转窑内温度场为对象,实现了系统原型机。通过实际数据进行验证试验,结果表明,该方法能够较准确地给出仿真结果,极大缩短计算时间,从而为回转窑内温度监测提供一种新思路。
A secondary simulation method based on the pattern recognition was proposed. First, the relationship between images, running states and simulation results were introduced. Then, a secondary simulation framework was proposed, and the mathematic model, pattern recognition and running states clustering were explained in detail. Finally, a system prototype machine was achieved by VC++6.0 based on the images and production data sampled from the rotary kiln of an alumina plant. Simulation experiments were carried out between the proposed method and the Fluent software under the corresponding condi- tions. The results indicate that the proposed method can obtain satisfactory simulation result in very short time, which provides a novel view for the temperature monitoring in rotary kilns.