An intelligent supervision system for cylindrical traverse grinding

The paper presents a supervision system that uses techniques of artificial intelligence to monitor, control and optimise the traverse grinding operation. The system consists of two levels which act in parallel to produce parts satisfying the geometrical and surface finish requirements with maximum possible productivity. The objective of the first optimisation level is to maximize the material removal rate, simultaneously satisfying restrictions on surface roughness, out-of-roundness and waviness errors and on grinding temperature. At the same time the second, geometrical control level is responsible for the removal of the initial shape error by stabilising the motion trajectory of the grinding wheel in relation to the part being ground. The performance of the supervision system was evaluated by extensive experiments to prove its effectiveness.