| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 805179 | Precision Engineering | 2012 | 9 Pages |
In this paper, two techniques are described for experimentally identifying the spindle-machine receptances required for tool point frequency response prediction using Receptance Coupling Substructure Analysis (RCSA). In the RCSA approach, the tool–holder–spindle-machine assembly is separated into three components: the tool, holder, and spindle-machine. The spindle-machine receptances are measured and archived. These receptances are then analytically coupled to beam models that represent the tool–holder. The spindle-machine dynamics are determined using: (1) a synthesis approach where a direct frequency response measurement of a standard artifact inserted in the test spindle is combined with a cross frequency response measurement to calculate the required rotational receptances; and (2) a new Euler–Bernoulli beam approach where the direct frequency response measurement is fit using an assumed (fixed-free) form of each mode within the measurement bandwidth. Experimental results are included for two spindles and four tool–holder combinations. The veracity of the new Euler–Bernoulli beam approach, which requires only a single measurement, reduces noise, and improves tool point dynamics prediction accuracy, is demonstrated.
► Using the Receptance Coupling Substructure Analysis (RCSA) technique, the tool–holder–spindle-machine assembly is separated into three components: the tool, holder, and spindle-machine. ► The spindle-machine receptances are measured and archived and the tool and holder are modeled using beam elements. ► These receptances are then analytically coupled to predict the tool point receptance, which can then be used as input to milling process models. ► Two approaches are described for experimentally identifying the spindle-machine receptances. ► A new Euler–Bernoulli beam approach, where the measured receptance is fit using an assumed (fixed-free) form of each mode within the measurement bandwidth, is presented and validated.
