Force Control in Two-Dimensional End Milling Operations Using Recurrent Neural Networks
A neural network control scheme is presented to control the average resultant cutting force in two-dimensional end milling operations. Recurrent neural networks are used as they have the potential to provide better approximation capability due to their dynamical nature. The neural controller, which is trained by using a neural identifier which models the process, specifies the feed rate to maintain the cutting force at the desired level. The feed rate is then resolved along the feed axes using a parametric interpolation algorithm so that the desired part shape is obtained. The experimental results obtained show that the neural control scheme can control the average resultant cutting force well. Also, a scheme for implementing on-line training of the neural controller is presented. Initial experimental results obtained using this scheme show that the neural controller can adapt to a changing environment.
T. Luo et al., "Force Control in Two-Dimensional End Milling Operations Using Recurrent Neural Networks," American Society of Mechanical Engineers, Dynamic Systems and Control Division (Publication) DSC, vol. 57-2, pp. 773-780, American Society of Mechanical Engineers (ASME), Nov 1995.
ASME International Mechanical Engineering Congress and Exposition (1995: Nov. 12-17; San Francisco, CA)
Mechanical and Aerospace Engineering
ASME PVP; ASME NE
Keywords and Phrases
Algorithms; Control equipment; Force control; Interpolation; Metal cutting; Neural networks; Two dimensional end milling operations; Milling (machining)
Article - Conference proceedings
© 1995 American Society of Mechanical Engineers (ASME), All rights reserved.