Intelligent Sliding Mode Control of Cutting Force During Single-Point Turning Operations

Abstract

A novel intelligent control architecture has been developed to regulate cutting forces during single-point turning operations. A self-adapting Sliding Mode Controller (SMC) accounts for parameter variations and unmodeled dynamics in the cutting process. A unique artificial neural network, the 2-Sigma network, statistically bounds modeling uncertainties between a low-order, linear dynamic model and the actual cutting process. These uncertainty bounds provide “localized” gains for the SMC, thus reducing excess control activity while maintaining performance. Initially, the 2-Sigma networks are trained off-line using experimental data from a variety of operating conditions. In the final implementation, the 2-Sigma networks are updated on-line, allowing the SMC to respond to parameter variations and unmodeled dynamics. Experiments conducted on a CNC lathe demonstrate the exceptional performance and robustness of this control system.