论文部分内容阅读
以磁为激励源进行轧辊表面及内部裂纹缺陷的检测,基于磁荷理论的裂纹数学模型的建立和神经网络的分段循环逼近的裂纹信号定量分析,实现了裂纹无损定量检测。经在铁磁材料上模拟实验取得了满意效果。
Using magnetic as excitation source, the surface and internal crack defects were detected, the crack mathematical model based on the magnetic charge theory and the quantitative analysis of the crack signal in the piecewise cyclic approximation of the neural network were used to achieve the non-destructive quantitative detection of cracks. After the ferromagnetic material simulation experiments achieved satisfactory results.