Error Compensation Using Inverse Actuator Dynamics

On this page:

Abstract

Research Method

Results

Abstract

The objective of this project is to model the surface features generated during a diamond turning operation as a result of intentional and unintentional tool motions. The intentional tool motions are from axis command and fast tool motion programs while the unintentional motions are due to tool vibration or error motions in the part. The goal will be to discover the range and quality of features that can be fabricated using the axes available on a machine and then to use that information to select the cutting conditions that optimize the desired surface shape.

To eliminate induced vibrations within the fast tool servo, the Variform integrates an internal second-order filter on the command input signal yielding the better closed-loop performance with a broader bandwidth. This improvement, however, delays the response to input commands producing form errors on the surface. An algorithm using surface decomposition will be developed to generate open loop input commands for the dynamics compensation. Since the command does not calculated from the feedback errors, there is no delay time of the response.

Research Method

The investigation has been developed into four stages:

• Modeling the dynamics of the Variform characterizes as a 2nd order dynamics system. The amplitude gains and phases were illustrated through a frequency spectrum and tabulated for a manual control adjustment. Research is currently investigating the validation of this model.
• Based on the concept of Digital Signal Processing, convolution theorem will be introduced to routinely generate the modified input. The dynamics of the fast tool servo will be measured. The algorithm implementation will experimentally justify the viability of the concept.
• The algorithm will be developed for a general tool path to discover unintentional tool-motion errors and hardware limitations.
• Integrating the tool servo to a diamond turning machine will explore the range and quality of feature in order to properly select cutting conditions to optimize the surface shape.

Results

Using the characteristic equation of Laplace Transfer Function as an approximate model of the system dynamics can create a look-up table containing the values of attenuation and phase respected to the operation frequency. In accordance with spatial frequencies of a desired surface, the attenuations and phases were manually selected from the table to modify the input in the manner that counteracts the dynamics. The validation was analyzed and exemplified through the numerical simulation.

The following faculty, students, and PEC affiliates are involved in this project: