【摘要】 提出了一种模具自由曲面超精研磨加工中人机一体化智能控制策略。该系统的初始研磨加工参数由基于神经网络的专家系统决定,能克服传统专家系统无法自学习的缺点。研磨过程中由一个模糊神经网络和人组成的合作智能控制系统实时修正研磨参数,通过人机一体化系统中的“人在回路中”方法,使研磨质量和研磨效率达到最佳。相关试验证明,该系统可以克服常规磁力研磨中工艺参数完全依赖操作者经验的缺点,是实现模具表面磁力研磨自动化的有效途径。更多还原

【Abstract】 Based on the study of the magnetic abrasive finishing (MAF) mechanism, the expert system, the fuzzy-neural network and the human-machine system theory, a human machine intelligent control system in molds free surface finishing process was proposed. The initial MAF parameters were determined by the expert system which was based on a neural network. Therefore, the expert system was able to learn by itself. Furthermore, a human machine cooperation system (composed of human and a fuzzy-neural network) made it possible for the revision of process parameters in real-time. Using this system optimum finishing quality and high efficiency can be attained in a highly dynamic finishing process. The relevant experiment showed that this system could determine the process parameters intelligently instead of completely rely on the operators’ experiences and this kind of control system is an effective way to realize molds surface finishing automatically.更多还原