基于模糊神经网络进行尺寸加工精度智能监控的研究

STUDY ON THE INTELLIGENT CONTROL OF WORKPIECEDIMENSION ACCURACY BASED ON FUZZYNEURAL NETWORKS

提出了基于模糊多层感知器神经网络模型的尺寸加工精度在线智能监控新方法，它对于缺乏精确的数学描述或具有时滞、非线性等复杂实际对象，是一种行之有效的技术，并具有及时性和预报性好的特点。描述了多层感知器神经网络的工作原理，阐明了基于人的模糊经验规则实现工件尺寸精度智能监控的三层感知器控制模型和学习算法，并进行了试验仿真。结果表明，方法有效，具有应用前景。

A new method for on-line intelligent control of workpiece dimension accuracyvia fuzzy multilayer perception networks is proposed. With good forecast in time,it is aneffective means on the occasions of lack of precise mathematical description or of timed-elayed and non-linear system. The principle on multilayer perception networks is given,and the three layer perception model and its learning algorithm derived from human kno-wledge on iatelligent control of workpiece dimension accuracy are described. With expe-rimental simulation, the method holds a promise of application.