Autonomous cutting parameter regulation using adaptive modeling and genetic algorithms

In this research, a turning process is modeled adaptively by a backpropagation, multilayered neural network with an iterative learning method, and cutting parameters of the process model are optimized through genetic algorithms (GAs). Some constraints were given on the input conditions and the process outputs to provide for the desired surface integrity and to protect the machine tool. Introducing penalty values, which are included in the fitness evaluation of the GAs, we can solve such a constrained problem. Experimental results show that the neural network has the ability to model the turning process on-line, and such cutting conditions as spindle speed and feed rate can be adaptively regulated for maximizing the material removal rate using the GAs.