基于神经网络的铜转炉内壁蚀损红外检测方法

An Infrared Detection Method of Copper Converter Inner-lining Wear Based on the BP Neural Network

针对铜转炉内壁蚀损状态难于直接检测的问题,提出了一种基于神经网络和红外无损检测技术的铜转炉内壁蚀损在线检测方法,分析了炉表面温度场与炉衬内壁蚀损的对应关系以及由红外热像仪测取炉表面温度场、图像处理与神经网络智能计算炉内壁蚀损量的检测原理,建立了相应的检测系统。试验检测的误差在5%之内,结果表明:建立的铜转炉内壁蚀损红外检测方法是可行的,具有较好的适应性,可满足现场的测量要求。

In order to overcome the difficulty of directly detecting the copper converter inner-lining wear, a new on-line detecting method based on the infrared nondestructive test and BP neural network is presented. The relationship between inner-lining wear and temperature field of converter surface is analyzed and the determination principle of inner-lining wear amount is investigated with the combination of temperature field measurement by infrared thermograph, image processing and edge detection and intelligent calculation such as neural network; and its detection and analysis system is given, by which the measurement error is within 5% in the smelting plant. The experimental detection results show that the infrared detection method of copper converter inner-lining wear based on the neural network is feasible and may satisfy the requirements of detection to the smelting plant.