摘要
An adaptive learning control scheme intended to the on-lineoptimization of sculptured. The scheme uses a back-propagation neuralnetwork to learn the relationships between process inputs and processstates. The cutting parameters of the process model are optimizedthrough a genetic algorithms(GA). The capacity of the proposed schemefor determining optimum process inputs under a variety of processconditions and optimization strategies is evaluated on the basis ofmilling of a sculptured surface using a ball-end mill. Theexperimental results show that the neural network could model thecutting process efficiently, and the cutting conditions such asspindle speed could be regulated for achieving high efficiency andhigh quality. Therefore the proposed approach can be well applied tothe manufacturing of dies and molds.
An adaptive learning control scheme intended to the on-lineoptimization of sculptured. The scheme uses a back-propagation neuralnetwork to learn the relationships between process inputs and processstates. The cutting parameters of the process model are optimizedthrough a genetic algorithms(GA). The capacity of the proposed schemefor determining optimum process inputs under a variety of processconditions and optimization strategies is evaluated on the basis ofmilling of a sculptured surface using a ball-end mill. Theexperimental results show that the neural network could model thecutting process efficiently, and the cutting conditions such asspindle speed could be regulated for achieving high efficiency andhigh quality. Therefore the proposed approach can be well applied tothe manufacturing of dies and molds.
