感謝山衛科技同意轉載「山衛科技電子報」系列專題文章，本篇文章原始連結為http://www.samwells.com/bc/news-tw/news-tech-news-tw/466-polytec-improving-machining-techniques
Improving Machining Techniques
在銑切加工時，工具機-工件系統的結構模態常受到切削力的激發，這種週期性的刺激經由刀具傳遞後在工件表面留波浪形紋路和較大的粗糙度，波紋表面進一步影響下一輪切削，使結構振動幅度逐漸放大的現象稱為顫振。
為了避免顫振的發生，由頻率響應函數可定義出切削穩態圖，提供適當的加工參數組合。但因為航太材料質輕的特性與薄殼的結構需求，當量測頻率響應函數時，加速規自身的重量會嚴重影響量測結果，造成頻率響應函數的誤差，誤判穩態區域。
為了瞭解加速規的質量效應對於薄殼元件的影響，使用 Polytec 雷射測振儀針對有無加速規的量測情況進行實驗，實驗結果顯示加速規的自身重量造成薄殼元件的頻響函數將近 50%的誤差，所以薄殼/質輕元件的頻率響應函數量測應使用非接觸式量測取代傳統加速規。
In milling, structural modes of the machine tool-workpiece system are initially excited by cutting forces. Surface finish profiling due to oscillatory excitation left by a section of the tooth is subsequently removed by the incoming and advancing cutting tooth surface which causes increased structural vibrations. This self-excited cutting phenomenon can become unstable, and chatter vibrations grow until the tool jumps out of the cut, ruins the expected surface tolerances and can even break under excessive cutting forces.
High speed machining is a widely used process in the aeronautical industry to machine low stiffness structures with thin walls and floors where high tolerances are required. Machining of these types of structures may experience regenerative lateral vibrations for some cutting conditions. The inherent nature of variable dynamic conditions of these milling and machining processes are likely to be the culprits by which the finished parts exhibit poor surface finish and lower productivity of those manufacturing processes. Stability lobes diagram methods are common techniques that use dynamic information to define stability regions in which it is possible to find the appropriate and desired combinations of machining parameters. With this technique, the experimental Operational Frequency Response Functions (OFRF) and regular FRF are required to feed the Enhanced, Multistage Homotopy Perturbation Method (EMHPM).
Comparison Between Laser Vibrometry and Existing Measuring Methods
It is well-known that accelerometer mass load on heavy structures has negligible influence on dynamic measurements. However, those effects are not negligible when the workpiece mass is small. Since the accuracy of the OFRF directly affects the stability lobes diagram, it is important to study the accelerometer mass loading effects over the stability diagrams to predict accurately the dynamic behavior when milling thin walled parts.
In order to study accelerometer mass loading effects on thin walled structures, we performed several impact tests at different locations over aluminum 7075 thin-walls (1 x 35 x 50 mm) and collected the corresponding FRFs by using a 0.6 gram accelerometer. We repeated the measurements using a Polytec CLV-2534-2 Compact Laser Vibrometer that allows dynamic measurements without adding mass to the workpiece. The dotted lines shown in fig. 1 represent the frequency response functions acquired using the accelerometer and Laser Doppler Vibrometer (LDV) under the same conditions.
Fig. 1 illustrates significant differences between the FRF functions obtained by using the accelerometer and the laser Doppler device. The laser vibrometer without the accelerometer attached captures two fundamental vibrational modes with peak values at 1105 Hz and 1722 Hz.
However, the measurements performed with the accelerometer exhibit the same dynamic
modes but with peak values at 580 Hz and 1366Hz. These differences in the recorded FRF spectra were noticed during experimental tests because of the sound pressure level produced by the thin wall workpiece. The 525 Hz shift of the FRF value of the first peak mode is attributed to the effect of the accelerometer mass.
modes but with peak values at 580 Hz and 1366Hz. These differences in the recorded FRF spectra were noticed during experimental tests because of the sound pressure level produced by the thin wall workpiece. The 525 Hz shift of the FRF value of the first peak mode is attributed to the effect of the accelerometer mass.
In order to verify our experimental observations, we performed measurements by using the LDV with the accelerometer attached to the workpiece. The results, shown by the
dashed line in fig. 1, are the same as those collected with the accelerometer. This experimental test confirmed the accelerometer mass-loading effects over FRF values, which not only causes a shift of the frequency value of about 48%, but also changes considerably the modal damping of the system. In addition, we expect changes in the calculated stiffness. As expected, the effect of the accelerometer mass increases as 1 mm thickness of wall material is re moved from the workpiece during the machining processes.
dashed line in fig. 1, are the same as those collected with the accelerometer. This experimental test confirmed the accelerometer mass-loading effects over FRF values, which not only causes a shift of the frequency value of about 48%, but also changes considerably the modal damping of the system. In addition, we expect changes in the calculated stiffness. As expected, the effect of the accelerometer mass increases as 1 mm thickness of wall material is re moved from the workpiece during the machining processes.
In order to demonstrate the effects of the accelerometer mass-loading on the dynamics ofthe cutting processes, we compute the stability lobes by using the EMHPM for both accelerometer and vibrometer measurements. The stability lobes plotted in fig. 2 are generated for a ½ inch diameter end mill with 2 teeth, a helix angle of20º, and a radial depth of cut of 0.8 mm.
As we can see from fig. 2, the stable depth of cut values on the computed stability lobes are strongly influenced by the accelerometer mass load. Accelerometer measurements produce a shift in stability lobes not only on spindle speed direction, but also axial depth direction in comparison with vibrometer measurements. For this reason, unstable predicted cutting conditions are experimentally explored by means of the LDV. The beam was aimed on the top center of the thin wall.
For a 26,500 rpm spindle speed, the maximum vibration amplitudes were recorded up to 0.3 m/s (fig. 3), with a confirmed stable behavior through the frequency spectrum which corresponded to the tool passing-frequency and its harmonics. On the other hand, an accelerometer-predicted unstable boundary region was explored by LDV OFRF responses (fig. 4). In this case, the velocity amplitude at 30,000 rpm reached 0.6 m/s. The frequency domain exhibits a chatter frequency.
Conclusion
Fig. 5 shows a comparison between the modal parameter values obtained by using both the FRF of the accelerometer and the laser vibrometer recorded data. It can be clearly seen that an accelerometer of 0.6 grams attached to the thin wall workpiece has a significant effect on FRF measurements.
原文出處: Polytec InFocus, Issue 02 2012
2.加速規
