论文部分内容阅读
提出用试验设计与人工神经网络相结合的方法 ,建立了一个激光切割工艺参数的选择优化的智能系统。通过试验设计的方法 ,只需做少数几次切割试验 ,将试验结果输入人工神经网络中进行训练和学习 ,系统便可经过自学得到切割结果与切割参数之间的隐含的定量关系 ,获得切割知识。在实际切割时 ,系统根据学到的切割知识 ,可以对任何给定的切割条件进行推理 ,对切割参数进行结果预测和优化。这种方法既不需要做大量的工艺试验 ,也不是单纯的专家经验 ,而是结合了两者的优点 ,使结果预测建立在既有试验数据又有专家知识的基础上 ,因而更加可靠、准确。该系统在激光方位切割的具体应用表明 ,系统能够准确给出定量的加工参数。
The method of combining experimental design with artificial neural network is proposed and an intelligent system for selecting and optimizing process parameters of laser cutting is established. Through the method of experimental design, only a few cutting tests are needed, and the test results are input into the artificial neural network for training and learning. The system can obtain the cutting quantitatively after learning the implicit quantitative relationship between the cutting result and the cutting parameter know how. In the actual cutting, the system according to the cutting knowledge learned, you can reason about any given cutting conditions, the cutting parameters of the results predicted and optimized. This method does not need to do a large number of process tests, nor is it a pure expert experience, but combines the advantages of the two, so that the results based on established test data and expert knowledge, based on more reliable and accurate . The specific application of this system in laser azimuth cutting shows that the system can accurately give quantitative processing parameters.