论文部分内容阅读
研究了离焦量、脉冲能量、扫描间距、扫描速度和重复频率等激光加工参数对金属表面着色及微纳结构制备的影响机理,诱导制备了氧化膜、类光栅、凹坑和柱状突起4种结构,这些结构会使不锈钢表面产生薄膜干涉、光栅衍射和陷光等现象。通过Matlab软件在工艺参数与颜色HSB值之间建立了一个单隐含层的反向传播(BP)神经网络,该神经网络的训练均方根误差为0.0078,色相H、饱和度S和亮度B的测试相对误差分别为23%,10.4%和5.6%。该神经网络在一定程度上揭示了工艺参数与颜色之间的映射关系,使用该神经网络模型可以对激光着色效果作出有效的预测。
The effects of laser processing parameters such as defocus amount, pulse energy, scanning pitch, scanning speed and repetition frequency on the preparation of metal surface and the preparation of micro / nano structures were studied. Four types of oxide films, gratings, pits and columnar protrusions Structure, these structures will make the stainless steel surface film interference, grating diffraction and trapping and so on. A BP neural network with single implied layer was established by Matlab software between process parameters and color HSB values. The training RMSE of the neural network was 0.0078, the hue H, the saturation S and the brightness B The relative error of the test was 23%, 10.4% and 5.6% respectively. The neural network reveals the mapping relationship between the process parameters and the color to a certain extent. The neural network model can effectively predict the laser coloring effect.