Abstract
The study of machine-tool dynamic is realised here as “monitoring”, meaning checking and improving the functioning of the machine. The state of processing is followed with certain sensors which signs are processed inside the computer, then it takes the decision of monitoring, meaning the identification of a class from the set of classes (process conditions). For monitoring in turning, it is shown the classes (tool conditions). The vibrograms that represent: - the components variations of the cutting force; - the relative displacement between tool and piece, on the repelling direction; - the power furnished by the electric engine, are realised with the functions RANDN and RAND (from MATLAB). Based on them it is calculating 11 monitoring indices. The class resulted at the experiment (simulate) i, which corresponds to the monitoring indices, we establish with the function REM. The ANN with 11 inputs (the number of monitoring indices) and 8 outputs (the number of classes) is realised with 3 layers. The network is made with the function newff, trained with the function train and simulated with the function sim.
Keywords:
Simulation; monitoring; lathe; artificial neural network; learning, classification; MATLAB