Adaptive optimal control of machining process using neural networks

The results of adaptive optimal control of a turning process based on the multilayered perceptron neural network are reported. To implement an adaptive control scheme, the characteristics of the process have to be identified in real time. In this study, the multilayered perceptron neural networks were used to identify the relationship between the input and the output of the turning process. The results of the experimental evaluation show that the relationship between the input and the output of the machining process can be effectively simulated with neural networks and that an adaptive optimal control system based on the neural network works reasonably well. The perceptron neural network technique is more favorable than conventional techniques in the sense that it assumes minimal a priori knowledge about the structure due to the self-organizing property of the neural network, and the nonlinearities of the system can be effectively modeled