论文部分内容阅读
针对复杂生产过程中的工况识别问题，利用了模块化的设计方法，将其分为离线建模模块和在线决策模块．建模模块采用自适应延时神经网络构造各类工况相应的预测模型，决策模块通过计算未知工况序列相对于各类预测模型的匹配程度判断其所属类别．
In view of the condition identification in complex production process, a modular design method is used, which is divided into offline modeling module and online decision module. The modeling module adopts the adaptive time-delay neural network to construct the corresponding prediction model for all kinds of working conditions. The decision module judges the classification of the unknown working sequence relative to the matching degree of various prediction models.