摘要
钢铁腐蚀是威胁工业安全生产的重要原因之一,脉冲涡流热成像缺陷检测技术可以对钢板表面腐蚀进行检测和评估。结合数值模拟和实际腐蚀检测试验,对利用脉冲涡流热成像检测腐蚀缺陷的特点进行了分析,并提取了温度上升较大的高温区宽度和温度最大值为特征量描述腐蚀的宽度和深度。将试件表面沿线圈方向的温度轮廓作为LSSVM的输入,腐蚀的二维轮廓作为输出,由试验测量数据和有限元仿真计算得到的仿真数据组建样本库。建立了由缺陷的温度场信号到缺陷轮廓图像的映射关系,实现了腐蚀缺陷的二维轮廓重构。
Steel corrosion is one of the most serious problems for industrial safety. Steel surface corrosion can be detected and evaluated by pulsed eddy current(PEC) thermography. Combining numerical simulation and actual corrosion detection experiments, the characteristics were analyzed using PEC thermography to detect corrosions. The width of the high temperature zone and the maximum temperature were extracted to describe the width and depth of the corrosion. The surface temperature profile along the coil was chosen as the input data of the LSSVM (least squares support vector machine). The 2D profile of corrosion was output data. The sample dataset was established with the simulation data and experimental data. The mapping relationship of the thermal image to the defect profile was established. It carried out the 2D profile reconstruction of corrosion defects.
出处
《腐蚀与防护》
CAS
北大核心
2015年第7期652-658,共7页
Corrosion & Protection
基金
国家自然科学基金(51307183)
关键词
脉冲涡流热成像
腐蚀
支持向量机
重构
pulsed eddy current thermography
corrosion
support vector machine
reconstruction