Reconstruction of cutting forces through fusion of accelerometer and spindle current signals

Abstract

Analysis of cutting forces is critical for understanding, control and optimization of machining operations. Table-dynamometers are the most common tools to measure the cutting forces directly. Workpiece geometry restrictions, high cost, altered dynamics, and limited frequency bandwidth are challenges associated with the direct measurement of the cutting forces using table-dynamometers. In this study, utilization of cost-effective spindle current sensors along with accelerometers is proposed to indirectly predict the cutting forces at the tool tip during milling operations. First, receptance coupling (RC) method is used to obtain dynamics between the tool tip and the accelerometers which are located on the spindle housing. The application of the RC method requires a series of impact hammer tests along with finite element simulations to identify the dynamics of machine tool structure. The measured accelerations are then integrated in frequency domain to obtain displacement of the spindle housing where the accelerometers are attached. A Kalman filter is then designed and applied to the displacement signals to reconstruct the AC component of the cutting forces. Next, the spindle current signals measured through the hall-effect sensors are processed to acquire the tangential and radial cutting forces. Finally, due to different frequency bandwidth of the measured signals through the accelerometers and hall-effect sensors, the reconstructed cutting forces are fused to improve the cutting force measurement accuracy. The proposed method is also verified by conducting a series of half and full-immersion milling tests, and the reconstructed results are compared to the measurements of a reference piezoelectric-based table-dynamometer. Results show that the proposed sensing scheme can predict the cutting forces effectively only at a fraction of the cost of traditional table-dynamometers without affecting overall workpiece geometries.