Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

Simulation of Grinding by Means of the Finite Element Method and Artificial Neural Networks turing process it is able to produce high workpiece surface quality. Improvements in its performance have allowed for the use of grinding in bulk removal of metal, maintaining at the same time its characteristic to be able to perform precision processing, thus opening new areas of applica- tion in today's industrial practice. The ability of the process to be applied on metals and other difficult to machine materials such as ceramics and composites is certainly an advantage of this manufacturing method. However, the energy per unit volume of material being removed from the workpiece during grinding is very large. This energy is almost entirely converted into heat, causing a significant rise of the workpiece temperature and, therefore, thermal damage. The areas of the workpiece that are affected are described as heat affected zones. Thermal load is connected to the maximum workpiece temperature reached As a novel contribution, FEM results on surface temperatures are used as input data to ANN models which in turn are able to predict with accuracy surface temperatures for various combinations of grinding conditions. These "hybrid" models are evaluated and compared to plain FEM or ANN models. The proposed analysis exhibits the simulation of grinding both with FEM and ANN, as well as a combination of the two methods, and draws useful conclusions. BACKGROUND Grinding is a manufacturing process character- ized by complex relationships between process parameters, workpiece and cutting tools charac- teristics as well as quality features of the finished products. Researchers have utilized modeling and simulation for several decades, now, in order to