About this Journal Submit a Manuscript Table of Contents
Advances in Mechanical Engineering
Volume 2013 (2013), Article ID 753631, 9 pages
http://dx.doi.org/10.1155/2013/753631
Research Article

Stiffness Identification of Spindle-Toolholder Joint Based on Finite Difference Technique and Residual Compensation Theory

1College of Mechanical Engineering and Applied Electronics Technology, Beijing University of Technology, Beijing 100124, China
2Design Center, Beijing No.1 Machine Tool Plant, Beijing 101300, China

Received 22 November 2012; Revised 5 January 2013; Accepted 28 January 2013

Academic Editor: A. Seshadri Sekhar

Copyright © 2013 Zhifeng Liu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The chatter vibration in high-speed machining mostly originates from the flexible connection of spindle and toolholder. Accurate identification of spindle-toolholder joint is crucial to predict machining stability of spindle system. This paper presents an enhanced stiffness identification method for the spindle-toolholder joint, in which the rotational degree of freedom (RDOF) is included. RDOF frequency response functions (FRFs) are formulated based on finite difference technique to construct a completed spatial FRF for the joint, where the measured data can be obtained from the piezoelectric acceleration sensors. In order to depress the influence of “modal truncation” and measurement noises, residual compensation theory is introduced to regenerate the RDOF FRF. Experiments are conducted to demonstrate the efficiency of the proposed model in stiffness identification of spindle-toolholder joint, and the accuracy is significantly improved compared to the traditional model.

1. Introduction

High speed machining has been widely used in automotive, aerospace, and other industries. This rising trend has presented new challenges for the machine tool design and application. High speed machining could produce higher material remove rate and higher surface quality. However, the higher machine tool strength and precision are required. The dynamic behavior of the machine tool is the key to determine the precision results. Generally, to mitigate the chatter of machine center, three factors are considered: cutting force conditions, work piece material properties, and the dynamics of the machine tool system.

Process control [15] is an effective way to improve the stability of the system. The system vibration can be monitored and depressed by controller and actuator during machining process. Shiraishi et al. [6] applied state feedback control to suppress machining chatter in turning. Luo et al. [7] presented a semiactive control law to reduce the detrimental motion of the building structure by adjusting the damping and stiffness of some parts. Moreover, the nonadaptive and adaptive output feedback sliding mode controllers were designed to improve the stability of structure [8]. Another method is to optimize the dynamic characterization for improving the stability of machine tool system. The stability lobe diagram has been widely used to predict the stability of the cutting condition. The frequency response function (FRF) of the machine tool is essential to plot a stability lobe diagram [914]. The spindle system is the most important part of a machine tool. The response at the tool tip is mostly affected by the spindle system. The spindle system mainly consists of spindle, supporting bearings, and toolholder. For the spindle system, the spindle-toolholder joint is one of the most flexible sections. A majority of the total stiffness of a spindle system comes from the spindle-toolholder joint [15]. Therefore, it is necessary to study the stiffness identification method of spindle-toolholder joint for marching stability analysis or the spindle system improvement. Much work can be found in this field [1625].

Substructure coupling method is a widely used method to study the spindle-toolholder joint. The spindle-toolholder system is virtually divided into three parts: spindle substructure, toolholder substructure, and the joint, which is usually assumed as a set of springs and dampers. Frequency responses of the substructures and assembly can be obtained by an experiment method or numerical simulation. Then the stiffness and damping of joint are calculated based on substructure coupling theory. Schmitz et al. [12, 1921] detailed this method and called it receptance coupling substructure analysis (RCSA). In their research, the spindle-toolholder system was divided into two substructures and a joint part. The joint consisted of a translational (or radial) spring and a rotational spring, and the stiffnesses were analytically identified by frequency response estimation. Based on this method, Schmitz and Duncan [19] extended the RCSA method to three substructures, in which the cutting tool was considered individually. Namazi et al. [23] modeled the spindle-toolholder joint by uniformly distributing translational and rotational springs. The stiffnesses were identified by minimizing the error between the experimental measured values and the estimated frequency response. Ahmadian and Nourmohammadi [24] also applied the distributed springs to an elastic layer interface.

The rotational frequency response is essential data for rotational stiffness identification. In a previous work, the rotational FRF was estimated by numerical calculation or was obtained by experiments. The numerical calculations are based on FE model. However, it is inevitably questioned for accuracy [26]. The experimental methods usually use angular response transducers directly. These transducers are usually very expensive and have a poor accuracy for measuring the spindle-toolholder assembly [2729]. Laser technique is also applied in some beam experiments [3032]. The noncontact nature of laser vibrometers offers advantages over traditional contacting vibration. By using multiple laser beam configurations, it is possible to measure the rotational vibrations. This method is still under development and is rather difficult to implement.

In this paper, an enhanced stiffness identification method for the spindle-toolholder joint is proposed. The whole assembly is divided into three substructures: spindle substructure, toolholder substructure, and the joint. The dynamical model is established based on substructure synthesis method. A traditional identification model without RDOF is calculated for comparison purpose. Frequency responses of the substructures and assembly are obtained by experiments. The translational FRFs are directly obtained by acceleration sensors. The rotational FRFs are derived from several series of translational FRFs, based on finite difference technique. Some modes cannot be measured due to the limit of experimental frequency range. It is called residual problem, sometimes also referred to as “modal incompleteness” or “modal truncation.” In order to reduce this negative influence and mitigate experimental errors, residual compensation is introduced. The regenerated rotational FRFs are much more ideal, and the identified stiffnesses become more accurate. Stiffness identification for a BT40 spindle-toolholder joint is conducted to demonstrate the efficiency of the proposed method. It is concluded that the identification model is much more accuracy compared to the model which not including RDOF.

2. Stiffness Identification Model of Spindle-Toolholder Joint

Substructure synthesis method is adopted to identify the stiffness of spindle-toolholder joint [12]. The dynamical model is illustrated in Figure 1, which is composed of three sections: toolholder substructure A, spindle substructure B, and joint C. The joint C is assumed to consist of a series of springs and dampers and is connected with substructure A and B rigidly. The connected regions of substructure A and B are both denoted as zone “j”; the unconnected region of substructure A is denoted as zone “a”; the unconnected region of substructure B is denoted as zone “b.”

753631.fig.001
Figure 1: Dynamic model of spindle-toolholder system.

Joint dynamical parameter denotes the inverse of the frequency response , which consists of stiffness matrix and damping matrix shown in the following:

According to [25], can be expressed as where , , , and are translational response of force excitement, translational response of moment excitement, rotational response of force excitement, and rotational response of moment excitement, respectively. The superscripts and refer to the substructures, and the quantities without superscripts refer to the assembly. The subscripts denote the response and excitement situation. For example, the subscript refers to the response of zone “j” under the excitement of zone “a.”

A stiffness matrix with complete spatial degrees can be expressed as where denotes the translation-to-force stiffness, the translation-to-moment stiffness, the rotation-to-force stiffness, and the rotation-to-moment stiffness.

According to results of [21, 23], the elements and have less effect on the dynamic behavior of the spindle-toolholder assembly. Therefore, and can be neglected for the purpose of improving the calculation efficiency. Some researchers [16] proposed the simplified identification model, as shown in (5), which neglected the rotational FRFs. The model is denoted as joint type I in this paper. However, this simplified model is not suitable for the spindle-toolholder system due to significant influence of RDOF. The adopted model including rotational stiffness is denoted as joint type II, as shown in (6) as

3. RDOF via Finite Difference Technique

Finite difference technique [27, 28] can be used to estimate rotational quantities by using translational quantities. Three acceleration sensors are deployed on spindle or toolholder, as shown in Figure 2. , and are the translational displacements of points A, B, and C, respectively. , and are the rotational displacements of the three points. ,  , and can be calculated by where is a transformation matrix, can be expressed as where denotes a constant spacing of the three points, and has an important influence on the estimation. The smaller the is, the more accurate the ,  ,   will be.

753631.fig.002
Figure 2: Finite difference schematic diagram and the sensors arrangement.

Dynamic function of the system under excitements can be expressed as where , and are in the form of column vectors. They denote translational displacements, rotational displacements, forces, and moments, respectively. Frequency response matrix is in the form Here, , and are corresponding to , and quantities in (3).

Equations (7), (9), and (10) can be rearranged to yield where and are irrelevant variables therefore, (11) leads to The completed spatial FRF matrix is a symmetric matrix therefore, So far, if given the translational FRF , the RDOF FRF , and can be derived.

Translational FRF matrices are obtained through experiments. However, due to the limit of sensor measurement range, some low and high modes cannot be picked up. This is called “residual problem,” or “modal truncation.” This negative effect will be amplified when performing multiplications or inverse operations shown in (12) and (14). Therefore, it is necessary to conduct residual compensation.

4. Residual Compensation

A complete FRF for typical mechanical joints with small damping ratio can be expressed as where the subscript j and k refer to the response point and the exciting point, respectively; denotes the rth-order natural frequency; denotes the rth-order modal constant; where and are the jth and the kth items of the rth modal shape expression, respectively. In (15), is obtained through experiments, while the low order quantity and the high order quantity are to be compensated.

A series of points are selected from the experimental obtained FRF curve, which are denoted as , . Substituting these points into (15), we obtain By solving this equation, the modal constants can be derived. After substituting the solved into (15), an ideal FRF regenerated curve can be fitted out. As a result, the low order frequency response and the high order frequency response have been compensated.

5. Stiffness Identification and Verification for the Spindle-Toolholder Joint

5.1. Experiment for Translational FRFs

A BT40 type spindle-toolholder experiment system is built up, as shown in Figure 3. In order to exclude the influence of the toolholder-tool joint, a special toolholder with one cylinder end is designed, as shown in Figure 3. The spindle and toolholder are assembled together with a long and lightweight drawbar. One end of the drawbar is screwed and bolted with the toolholder. While the other end of the drawbar is connected with a tension bolt. The tension force of the bolt can be measured by a stress-strain meter, and this tension force transmits through the drawbar to the toolholder. The assembly is suspended to simulate free degree of freedom (DOF) state. In order to acquire experimental data of zone “a” corresponding to Figure 1, a group of sensors is set at the cylinder end of the toolholder, denoted as sensor group 1 in Figure 3. Sensors for zone “b” and zone “j” are located at the outside wall of the taper joint and are denoted as sensor group 2. Each group consists of three piezoelectric acceleration sensors. The sensor type is PCB model 330B30. An impact hammer is adopted for excitement. The exciting locations are set at the opposite location of the sensors, as shown in Figure 3. The signals of the impact hammer and the sensors are collected and analyzed in an LMS modal analyzer.

753631.fig.003
Figure 3: Spindle-toolholder joint experiment system.

The translational FRFs for zone “a,” “b,” and “j” in (3) are obtained by using the right sensors of each group, while the corresponding rotational FRFs , and can be derived using three translational sensors for each group according to the finite difference theory.

By the hammering test, a series of acceleration FRFs can be obtained by applying 8 kN, 10 kN, 12 kN, and 15 kN drawbar force independently. Figure 4 shows the series curve , which denotes translational-to-force FRF of the spindle-toolholder assembly. The natural frequencies are listed in Table 1. Comparing the natural frequencies under four different drawbar forces, the forth order is the most sensitive one to the drawbar force change with the highest changing ratio 12.7% [12]. Therefore, the forth order’s natural frequency and amplitude value are selected as inputs of proposed model to identify the stiffness of spindle-toolholder joint based on the sensitivity analysis theory. The stiffness of the joint varies with different drawbar forces. However, the identification methods are generally the same. In the following identification and analysis of the joint, the 15 kN drawbar force is adopted to validate the efficiency of the proposed method.

tab1
Table 1: Natural frequencies of the assembly system under four drawbar forces.
753631.fig.004
Figure 4: The series of FRF under four different drawbar forces.

5.2. Finite Element Model

In order to verify the stiffness identification efficiency, a finite element model for the spindle-toolholder system is established in ANSYS. The model is shown in Figure 5, and the geometry dimensions are listed in Figure 6 and Table 2. Material density is defined as , Young’s modulus , and Poisson’s ratio . Timoshenko beam element BEAM189 is adopted for both the toolholder and the spindle. In the joint area, the nodes of the toolholder are one-to-one correspondence with these of the spindle and are connected with 52 Matrix27 elements. Matrix27 is an arbitrary elastic element, which can be specified by stiffness matrix. The stiffness values of the 52 Matrix27 elements are assigned uniformly based on the identified results in the next section.

tab2
Table 2: (a) Geometry dimensions of the toolholder, (b) geometry dimensions of the spindle.
753631.fig.005
Figure 5: Spindle-toolholder system FE model.
fig6
Figure 6: (a) Geometry of the toolholder. (b) Geometry of the spindle.

5.3. Stiffness Identification

Rotational FRFs , , , , , , , , , and in (3) can be derived using (12) and (14). Figure 7 shows the amplitude of the derived , which is the translation-to-moment FRF of the spindle-toolholder assembly. Due to “zero drift” influence of the adopted sensors, the translational FRFs from experiments are distorted near . And the “zero drift” phenomenon will be amplified during deriving rotational FRFs. Therefore, the response range from 375 Hz to 2500 Hz is selected to exclude the influence of “zero drift” phenomenon for the derived , as shown in Figure 7.

753631.fig.007
Figure 7: Amplitudes of the derived .

Once the rotational FRFs are calculated, the joint stiffness can be identified using (2) and (3). The forth-order modal parameters are selected to identify the stiffness of spindle-toolholder joint. The forth-order natural frequency and the corresponding stiffness value are shown in Figure 8. The forth-order natural frequency is 1975 Hz, and the identified translational and rotational stiffnesses are and , respectively.

fig8
Figure 8: (a) Dynamic translational stiffness. (b) Dynamic rotational stiffness.

For the purpose of validating the proposed method, a single translational stiffness is assigned to the Matrix27 elements for the joint type I, while both the translational stiffness and rotational stiffness are assigned to the Matrix27 elements for the joint type II. Natural frequencies can be calculated by FE modal simulation as shown in Table 3. Compared with the experiment values, the natural frequencies adopting the joint type I are distorted. The errors of the 1st–3rd order natural frequency are 68.9%, 17.2%, and 7.7%, respectively. Nevertheless, the natural frequencies adopting the joint type II are relatively accurate. The error of the first-order natural frequency is 10.1%, which is much lower than the error of the type I. Although the errors still exist, they have been reduced significantly, as shown in Figure 9. It can be concluded that the joint type I model is invalid and the results with small error can be obtained by adopting the joint type II model. However, a further enhancement has to be developed to meet accurate FE simulation.

tab3
Table 3: Natural frequencies of the joint type I and the type II models.
753631.fig.009
Figure 9: Error comparison of the joint type I and the type II models.
5.4. Stiffness Identification after FRF Residual Compensation

Due to the range limitation of the adopted acceleration sensor, the “modal truncation” phenomenon will occur during the experiment. And the measurement noises are brought into the measured signals. All of these will affect the identification results. In order to mitigate the effects of “modal truncation” and measurement noises, the residual compensation method is introduced to further enhance identification accuracy of the spindle-toolholder joint. A series of points on the experimental curve are selected, and a new FRF curve can be fitted out based on formula (15). Figure 10 shows the translation-to-force FRF before and after residual compensation. The regenerated data are much better.

753631.fig.0010
Figure 10: Amplitudes of before and after residual compensation.

After the FRF curve is regenerated, the joint stiffness can be calculated adopting the regeneration data for the joint type II. The stiffnesses are denoted as and , respectively, as shown in Figure 11. The identification results of and are and Nm/rad, respectively. The stiffnesses are also assigned to the Matrix27 elements in the FE model as depicted in Section 5.3. The 1st–4th natural frequencies before and after residual compensation are listed in Table 4 and Figure 12. An obviously improvement has been obtained by introducing the residual compensation method. For the first order especially, the error reduces from 10.1% to 0.7%. The maximum error appears at the fourth order with error 6.1%. Therefore, we can conclude that the joint type II model adopting regeneration data is more efficient in the stiffness identification by introducing the residual compensation method. The identification results can meet the requirement of accurately predicting machining stability of spindle system. Simultaneously, structural optimization can be implemented based on the identified stiffness of spindle-toolholder joint in the future.

tab4
Table 4: Natural frequencies of the joint type II model before and after residual compensation.
fig11
Figure 11: (a) Dynamic translational stiffness after residual compensation. (b) Dynamic rotational stiffness after residual compensation.
753631.fig.0012
Figure 12: Error comparison of the joint type II model before and after residual compensation.

6. Conclusions

In this paper, an enhanced stiffness identification method was proposed for the spindle-toolholder joint. The joint type I and II were established, in which the joint type II include the rotational stiffness. Due to the difficulty of directly measuring the rotational FRFs, the finite difference technique was introduced to estimate the rotational FRFs of spindle-toolholder system. The error of first-order natural frequency for the joint type II was 10.1%, which is much lower than that of the type I. However, the identification result was affected by the “modal truncation” and measurement noises. We introduced the residual compensation method. After the FRF curve was regenerated, the joint stiffness can be calculated by adopting the regeneration data for the joint type II. An obviously improvement has been obtained, in which the error of first order is reduced from 10.1% to 0.7%. Experiment results showed the efficiency of the proposed model in stiffness identification of spindle-toolholder joint.

Acknowledgment

This research is supported by the National Science and Technology Major Project of China (no. 2010zx04012-011).

References

  1. S. Kohn-Rich and H. Flashner, “Robust fuzzy logic tracking control of mechanical systems,” Journal of the Franklin Institute, vol. 338, no. 2-3, pp. 353–370, 2001. View at Publisher · View at Google Scholar · View at Scopus
  2. M. Wang and R. Fei, “On-line chatter detection and control in boring based on an electrorheological fluid,” Mechatronics, vol. 11, no. 7, pp. 779–792, 2001. View at Publisher · View at Google Scholar · View at Scopus
  3. R. A. Ibrahim, “Recent advances in nonlinear passive vibration isolators,” Journal of Sound and Vibration, vol. 314, no. 3-5, pp. 371–452, 2008. View at Publisher · View at Google Scholar · View at Scopus
  4. K. J. Kalinski and M. A. Galewski, “Chatter vibration surveillance by the optimal-linear spindle speed control,” Mechanical Systems and Signal Processing, vol. 25, no. 1, pp. 383–399, 2011. View at Publisher · View at Google Scholar · View at Scopus
  5. F. Palacios-Quiñonero, J. Rubió-Massegú, J. M. Rossell, and H. R. Karimi, “Semiactive-passive structural vibration control strategy for adjacent structures under seismic excitation,” Journal of the Franklin Institute, vol. 349, no. 10, pp. 3003–3026, 2012. View at Publisher · View at Google Scholar
  6. M. Shiraishi, E. Kume, and T. Hoshi, “Suppression of machine-tool chatter by state feedback control,” CIRP Annals: Manufacturing Technology, vol. 37, no. 1, pp. 369–372, 1988. View at Publisher · View at Google Scholar · View at Scopus
  7. N. Luo, J. Rodellar, J. Vehí, and M. De la Sen, “Composite semiactive control of a class of seismically excited structures,” Journal of the Franklin Institute, vol. 338, no. 2-3, pp. 225–240, 2001. View at Publisher · View at Google Scholar · View at Scopus
  8. N. Luo, J. Rodellar, M. De La Sen, and J. Vehi, “Output feedback sliding mode control of base isolated structures,” Journal of the Franklin Institute, vol. 337, no. 5, pp. 555–577, 2000. View at Scopus
  9. J. Tlusty, “High-speed machining,” CIRP Annals: Manufacturing Technology, vol. 42, no. 2, pp. 733–738, 1993. View at Publisher · View at Google Scholar · View at Scopus
  10. J. Tlusty, “Design of machine tools: drives and structures,” in Manufacturing Processes and Equipment, pp. 559–578, Prentice Hall, Upper Saddle River, NJ, USA, 2000.
  11. T. L. Schmitz and B. P. Mann, “Closed-form solutions for surface location error in milling,” International Journal of Machine Tools and Manufacture, vol. 46, no. 12-13, pp. 1369–1377, 2006. View at Publisher · View at Google Scholar · View at Scopus
  12. T. L. Schmitz and K. S. Smith, Machining Dynamics: Frequency Response to Improved Productivity, Springer, New York, NY, USA, 2009.
  13. Y. Altintaş and E. Budak, “Analytical prediction of stability lobes in milling,” CIRP Annals: Manufacturing Technology, vol. 44, no. 1, pp. 357–362, 1995. View at Publisher · View at Google Scholar · View at Scopus
  14. S. Engin and Y. Altintas, “Mechanics and dynamics of general milling cutters. Part I: helical end mills,” International Journal of Machine Tools and Manufacture, vol. 41, no. 15, pp. 2195–2212, 2001. View at Publisher · View at Google Scholar · View at Scopus
  15. E. I. Rivin, J. Agapiou, C. Brecher, et al., “Tooling structure: interface between cutting edge and machine tool,” CIRP Annals: Manufacturing Technology, vol. 49, no. 2, pp. 591–634, 2000. View at Publisher · View at Google Scholar
  16. E. B. Kivanc and E. Budak, “Structural modeling of end mills for form error and stability analysis,” International Journal of Machine Tools and Manufacture, vol. 44, no. 11, pp. 1151–1161, 2004. View at Publisher · View at Google Scholar · View at Scopus
  17. D. M. Shamine and Y. C. Shin, “Analysis of No. 50 taper joint stiffness under axial and radial loading,” Journal of Manufacturing Process, vol. 2, no. 3, pp. 167–173, 2000. View at Publisher · View at Google Scholar
  18. M. Wang, D. Wang, and G. Zheng, “Joint dynamic properties identification with partially measured frequency response function,” Mechanical Systems and Signal Processing, vol. 27, pp. 499–512, 2012.
  19. T. L. Schmitz and G. S. Duncan, “Three-component receptance coupling substructure analysis for tool point dynamics prediction,” Journal of Manufacturing Science and Engineering, vol. 127, no. 4, pp. 781–790, 2005. View at Publisher · View at Google Scholar · View at Scopus
  20. G. S. Duncan, M. F. Tummond, and T. L. Schmitz, “An investigation of the dynamic absorber effect in high-speed machining,” International Journal of Machine Tools and Manufacture, vol. 45, no. 4-5, pp. 497–507, 2005. View at Publisher · View at Google Scholar · View at Scopus
  21. T. L. Schmitz and G. S. Duncan, “Receptance coupling for dynamics prediction of assemblies with coincident neutral axes,” Journal of Sound and Vibration, vol. 289, no. 4-5, pp. 1045–1065, 2006. View at Publisher · View at Google Scholar · View at Scopus
  22. S. S. Park, Y. Altintas, and M. Movahhedy, “Receptance coupling for end mills,” International Journal of Machine Tools and Manufacture, vol. 43, no. 9, pp. 889–896, 2003. View at Publisher · View at Google Scholar · View at Scopus
  23. M. Namazi, Y. Altintas, T. Abe, and N. Rajapakse, “Modeling and identification of tool holder-spindle interface dynamics,” International Journal of Machine Tools and Manufacture, vol. 47, no. 9, pp. 1333–1341, 2007. View at Publisher · View at Google Scholar · View at Scopus
  24. H. Ahmadian and M. Nourmohammadi, “Tool point dynamics prediction by a three-component model utilizing distributed joint interfaces,” International Journal of Machine Tools and Manufacture, vol. 50, no. 11, pp. 998–1005, 2010. View at Publisher · View at Google Scholar · View at Scopus
  25. T. Yang, S. H. Fan, and C. S. Lin, “Joint stiffness identification using FRF measurements,” Computers and Structures, vol. 81, no. 28-29, pp. 2549–2556, 2003. View at Publisher · View at Google Scholar · View at Scopus
  26. J. E. Mottershead, A. Kyprianou, and H. Ouyang, “Structural modification. Part 1: rotational receptances,” Journal of Sound and Vibration, vol. 284, no. 1-2, pp. 249–265, 2005. View at Publisher · View at Google Scholar · View at Scopus
  27. M. L. M. Duarte, Experimentally-derived structural models for use in further dynamic analysis [Ph.D. thesis], University of London, London, UK, 1996.
  28. M. L. M. Duarte and D. J. Ewins, “Rotational degrees of freedom for structural coupling analysis via finite-difference technique with residual compensation,” Mechanical Systems and Signal Processing, vol. 14, no. 2, pp. 205–227, 2000. View at Publisher · View at Google Scholar · View at Scopus
  29. N. Mehdi, Mechanics and dynamics of the tool holder-spindle interface [M.S. thesis], University of British Columbia, Vancouver, Canada, 2006.
  30. J. R. Bell and S. J. Rothberg, “Rotational vibration measurements using laser Doppler vibrometry: comprehensive theory and practical application,” Journal of Sound and Vibration, vol. 238, no. 4, pp. 673–690, 2000. View at Publisher · View at Google Scholar · View at Scopus
  31. M. Martarelli, C. Santolini, and P. Castellini, “Characterization of rotating structures in coast-down by means of continuous tracking laser doppler vibrometer,” in Modal Analysis Topics, vol. 3 of Conference Proceedings of the Society for Experimental Mechanics Series, pp. 525–532, Springer, New York, NY, USA, 2011. View at Publisher · View at Google Scholar · View at Scopus
  32. L. Xiang, S. Yang, and C. Gan, “Torsional vibration measurements on rotating shaft system using laser doppler vibrometer,” Optics and Lasers in Engineering, vol. 50, pp. 1596–1601, 2012. View at Publisher · View at Google Scholar