Tool Wear Prediction Approach for Turning Operations Based on General Regression Neural Network (GRNN) Technique

Abstract

Detection of tool failure is very important in automated manufacturing. Recent trends, being towards mostly unmanned automated machining systems and consistent system operations, need reliable on-line monitoring processes. A proper on-line cutting tool condition monitoring system is essential for deciding when to change the tool. Many methods have been attempted in this connection. Recently, artificial neural networks have been tried for this purpose because of their inherent simplicity and reasonably quick data-processing capability. The present work investigates the feasibility of using general regression neural networks (GRNN) for estimating the level of the nose wear on the cutting edge. Experimental data of different force components, as well as corresponding nose wear values and the three controlling cutting parameters (speed, feed, and depth of cut) are used to train the neural networks. The technique shows close matching of estimation of nose wear and directly measured wear value. Results indicated the need for inclusion of six input parameters (speed, feed, depth, and three force components) in order to get better prediction capability. However, some parameters, such as feeding and radial force components, have shown higher impact than others (power force component). Therefore, it is possible that the trained neural networks can accurately assess tool wear on-line using an appropriate system.