Journal of Materials Science and Engineering A 5 (5-6) (2015) 221-229 doi: 10.17265/2161-6213/2015.5-6.005 D DAVID PUBLISHING eal Time Chatter Vibration Control System in High Speed Milling Hyeok Kim 1, Mun-Ho Cho 1, Jun-Yeong Koo 1, Jong-Whan Lee 2 and Jeong-Suk Kim 3* 1. School of Mechanical Engineering, Pusan National University, Busan 609-735, epublic of Korea 2. Department of Mechtronics Engineering, Korea Aviation Polytechnic, Busan 616-741, epublic of Korea 3. Department of Mechanical Engineering, Pusan National University, EC/NSDM,Busan 609-735, epublic of Korea Abstract: This paper presents the chatter vibration avoidance method in high speed milling. Chatter vibrations have a bad influence on surface integrity and tool life. So how to get the cutting conditions without chatter vibration is very important. In order to get stable cutting condition, too many parameters are required. So this paper focuses on simplification and real-time control of chatter avoidance program. The developed method uses only microphone signal for chatter vibration sensing. The measuring signal is analyzed by FFT (Fast Fourier transform) method to get whether or not the chatter vibration is generated on cutting condition. If chatter vibration occurs, the developed program suggests stable cutting speed in real time with tool teeth number, damping ratio and chatter frequency. This suggested that the program reliability is confirmed by dynamic cutting forces and surface profiles. Key words: Chatter vibration, milling, impact test, real-time control, frequency response function, lab view. 1. Introduction Because of the acceleration and superior precision of machine tools, the numerous high-quality products are currently produced. However, chatter vibration, which degrades the quality of processed products, remains a problem to be solved. Chatter vibration, which causes problems during the manufacturing process, appears to be a self-excited vibration. When chatter vibration occurs, a chatter mark exhibits a wave pattern appears on the surface of the processed product, and thus, the processing grade is degraded. Additionally, the load on the axial system increases, the tool-life is reduced, and damage occurs owing to the vibration. Accordingly, the processing cost increases, resulting in a drop in productivity [1]. Thus, it is necessary to research a program that can detect chatter vibration and propose a stable spindle speed that can prevent chatter vibration according to the dynamic characteristics of the axial system. A lobe diagram is a graph illustrating the * Corresponding author: Jeong-Suk Kim, tenure professor, research fields: machine tools, dynamics and metal cutting. E-mail: juskim@pusan.ac.kr. relationship between the rotational speed and the depth of cut of the spindle [2]; this is the most fundamental theory. However, because it requires extensive parameters to be applied in the practical industrial field, it is not often used. Thus, in the present research, stable and unstable areas are not determined according to the lobe diagram [3]. Instead, this research aims to easily apply the stable cutting condition through the chatter frequency only when chatter vibrations occur so that chatter vibration can be prevented [4]. The purpose of this study is to develop a virtual dynamic system that proposes a virtual condition to analyze the signal characteristics when chatter vibration occurs during processing and to prevent chatter vibration by using lab view. 2. Virtual Dynamic Machining System A VDMS (Virtual dynamic machining system) was developed to precisely detect chatter vibration that occur during processing, and to effectively reduced it. In existing research on chatter vibration, numerous variables have been used to generate a model that is similar to practical chatter vibration [5].
222 eal Time Chatter Vibration Control System in High Speed Milling However, the present research focused on detecting the occurrence of chatter vibration in real time by using simplified variables. Flowchart of VDMS is shown in Fig. 1. The present program consists of the following six modules: a data filter processing module, a FFT (Fast Fourier transform) analysis module, a cutting force acquisition module, an impact excitation test module, a chatter vibration occurrence notification module, and a stable spindle speed recommendation module. Input values required for running the program are sampling rate, spindle speed, and number of cutting edges. Spindle speed is required to determine whether the detected frequency is a tool passing frequency or a chatter frequency. And the number of cutting edges is a necessary input value for calculating the stable spindle speed. The structure of the front panel of the developed VDMS is shown in Fig. 2 and the position of each module is as follows: The chatter vibration occurrence notification and stable spindle speed recommendation modules are located at Position 1. The input variable and detectedd frequency are marked at Position 2. The module located at Position 3 is for determining the cutting force signal. The FFT analysis module microphone is located at Position 4, and at Position 5 the specifications of the filter for the signal can be selected. The impact excitation test module is located at Position 6. Fig. 1 Flowchart of VDMS process. 2.1 Impact Excitation Test Module This module was operated by loading the excitation response data of the axial system, obtained by using the impact hammer and accelerometer. Among the impact testt data, the magnitude value was used in the impact excitation module, and the magnitude value configured in the two-dimensional array was converted to a one-dimensional array. Then, by using the peak detector, the natural frequency of the primary mode was calculated. The damping ratio δ,, was calculated after establishing two places at which the amplitude of the natural frequency was ± 1/ 2 [6]. The calculated natural frequency and damping ratio were delivered to the FFT analysiss module, which was used to designate the scope of the occurrence of chatter vibration. The prepared module is illustrated in Fig. 3. 2.2Chatter Detecting Method Constructed Module The developedd VDMS conducts FFT analysis while determining the cutting signal. The signal obtained from the microphone is conveyed to the VDMS through the DAQ (Data acquisition) board and transmitted to the FFT analysis module after passing the filter. The frequency content that has the maximumm amplitude of the analyzed microphone signal is detected by using the peak detector. Additionally, to determine the frequency thatt has an amplitude over a certain value, a threshold value was designated so thatt the frequency content could be detected. Generally, it is known that chatter vibration occurs near the natural frequency of the axial system, but nothing is quantitatively established. Because of the preliminary experiment, chatter vibration occurred only within the scope between approximately two and a half times the natural frequency of the axial system. Based on this scope, the preliminary experiment, chatter vibration occurred only within the scope between approximately two and a half times the natural frequency of the axial system. Based on this scope, the scope of chatter frequency f c was established by using
eal Time Cha atter Vibratio on Control Sy ystem in High h Speed Millin ng Fig. 2 2233 Frontt panel of VDM MS. Fig. 3 Block k diagram of im mpact excitatioon test module.. the natural frequency f fn, of the axial system s calcullated in the impacct excitation test module, and a the result was defined by using: u fn/2 < fc < 2fn (1) Next, an algorithm that removees the harm monic content of the t tool passsing frequenccy was addedd by inputting thhe spindle speed s value. The frequeency detected in the FFT anaalysis modulee falls underr the pe of Eq. (1) and has the m maximum am mplitude overr scop the threshold. Thhe frequency ccontent of thee tool passingg freq quency, not the t harmonicc content, waas establishedd as the chatterr frequency.. The prog gram sourcee calcculating the present p processs is shown in n Fig. 3 and iss processed as follows. The siggnals that passsed the filterr are FFT connverted annd inserted into thee onee-dimensionall array after bbeing separateed into f0, df,, and d magnitude. Among the signals, the frequency off the maximum amplitude a is detected. Th his module iss opeerated by loadding the excittation responsse data of thee axiaal system, obttained by usinng the impactt hammer andd accelerometer. Among the impact tesst data, thee mag gnitude value was used in the impaact excitationn mod dule, and thee magnitude value configured in thee two o-dimensionall array w was converrted to a onee-dimensionall array. Theen, by usin ng the peakk deteector, the naatural frequenncy of the prrimary modee wass calculated. The dampinng ratio δ waas calculatedd afteer establishingg two places at which the amplitude off the natural frequuency was ± 1/ 2 [5]. Th he calculatedd natu ural frequenccy and dampinng ratio weree delivered too the FFT analysiss module, whhich was used d to designatee the scope of thee occurrence of chatter viibration. Thee
224 eal Time Chatter Vibration Control System in High Speed Milling prepared module was illustrated in Fig. 2. through the peak detector and compared with the input machining speed to determine whether it was a tool passing frequency. Then, after confirming whether it occur near the natural frequency, the result was presented through the chatter vibration in the comparison operator. The chatter frequency detected in this process was conveyed to the indicator, which indicated whether chatter vibration had occurred. Additionally, the detected chatter frequency was used in the calculation of stable spindle speed, and the damping ratio δ required in this process is the value obtained by the impact excitation test module. The relationship between the chatter frequency f c and stable spindle speed N, which is calculated by Eq. (2) [7]. N = 60f c / (i(1 + δ)n t ) (2) N: Stable cutting speed (revolution per min). f c : Chatter vibration frequency. i: Lobe number. δ: Damping ratio. n t : Number of cutting tool edge. 2.3 Cutting Force Acquisition Module The cutting signal determined by the cutting force acquisition module is displayed in a graph in two ways: moving average method and resultant force method of each component of force. esultant force F is calculated through the component of force in the tri-axis direction, according to Eq. (3). When chatter vibration occurs, the resultant force largely increases compared to the stable machining; thus, the credibility of the chatter vibration detection is measured by using the size of the resultant force. F = (F 2 a + F 2 r + F 2 t ) 1/2 (3) F : esultant force. F a : Axial force. F r : adial force. F t : Tangential force. The module is prepared to add a scale input, which is the software amplifier, to the program, in addition to the hardware amplifier of the cutting signal. 3. Experimental Setup 3.1 Impact Test Setup Because the chatter vibration evasion process of the present research is conducted under the mechanism of detecting chatter vibration through the natural frequency of the axial system, it is prioritized to identify the dynamic characteristics of the axial system. The impact hammer underwent excitation by using Type 8206-002 from Bruel & Kjaer, and an accelerometer obtained a response signal by using Type 4384 from Bruel & Kjaer. In the case of the shock excitation test, the proficiency of the experimenter largely affects the result; thus, the frequency response was obtained through 10 mean value calculations by using only the experimental data, which features the credibility of the coherence function. A schematic diagram of the experiment equipment is shown in Fig. 4. 3.2 Cutting Experimental Setup The present experiment was conducted at the MAKINO V55 tri-axis machining center. egarding the signal during the cutting, data similar to the accelerometer were gathered, and for the sake of convenience in usage and signal measurement, the data were obtained by using the microphone (Bruel & Kjaer, Type 4189). Additionally, the distance between the areas at which the sensing and machining occurred was fixed at approximately 400 mm. The cutting force was obtained by using a Kistler 9257B-type dynamometer, and a Kistler 5019B130-type amplifier was used. The machined material was AISI 1045; the form of the machined product featured a low slenderness ratio, and thus, was machined to a bulk type so that vibration could not occur. Cutting fluid was not used here, and the filter used in determining the signal was a bandpass-type windowed FI filter. The band was within 100 Hz to 5,000 Hz. The UT coating end mill from Unimax was used as the tool.
2255 eal Time Cha atter Vibratio on Control Sy ystem in High h Speed Millin ng Tab ble 1. Cuttingg conditions. Spiindle speed Dep pth of cut (mm)) ad dial of cut (mm m) Feeed rate (mm/minn) Cuttting method Fig. 4 Block k diagram of FFT F module. Fig..6 5,0000 1.0 3.0 1,0000 Slot cutting 5,830 1.0 3.0 1,135 Experimen ntal setup. poin nt of the infleection of maggnitude amon ng the impactt testt data; thus, itt was displayed in a distorrted graph, ass indiicated in Fig. 7. The undistorted magnittude graph off the impact test was w confirmedd through thee Origin Pro** grap ph tool and was shownn in Fig. 8. The naturall freq quency, calcuulated by loadding the experrimental dataa in th he impact exccitation test m module of thee VDMS, wass 1,416 Hz, and thhe damping rratio was calcculated to bee 0.04 41. The scoppe in which the chatter vibration v wass deteected throughh natural freqquency was within w 703 Hzz to 2,832 2 Hz. 4.2 Cutting Forcce Analysis Fig.5 Front panel of VDM MS. There were two t cutting edges e with diaameter of 10 mm. The length of o the tool innstalled in thee tool holder was established at a approximaately 55 m, which w was sligghtly longer, so thhat chatter vibbration couldd occur relatiively easily. The machining m connditions were listed in Table 1, and the scheematic diagram m was shownn in Fig. 6. 4. esults 4.1 esults of o the Impact Excitation Test Te In the casse of VDMS, the peak waas obtained att the As A indicated in Fig. 9, thee size of the cutting forcee wheen chatter vibbration occurss appeared to considerablyy incrrease comparred to that unnder the stablle machiningg con ndition. The cutting c force aaveraged 246 6 N at spindlee speed of 5,830 PM and deepth of cut off 1.0 mm. A cut ting force of o 556 N w was measure d under thee con nditions of 5,,000 PM sppindle speed and 1.0 mm m dep pth of cut, where w chatter vibration occcurred. Thiss wass more than two t times larrger than the cutting forcee of the t stable macchining conddition. The cu utting force att the introductorry and end portions off the cuttingg
226 eal Time Chatter Vibration Control System in High Speed Milling Fig. 7 Impact test result of VDMS. Fig. 8 Impact test result of Origin Pro*. Fig. 9 Cutting forces depending on cutting conditions.
eal Time Cha atter Vibratio on Control Sy ystem in High h Speed Millin ng Fig. 10 2277 nt panel of VD DMS (spindle sp peed 5,000 PM, depth of cu ut 1.0 mm). Fron increased ow wing to the shhock from rottation [9]. 4.3 Front Paanel Verificattion of VDMSS To evaluaate the validiity of the chaatter detectioon of VDMS, whhich was opeerated after the microphhone signal, was determined in real timee during millling machining to verify thhe front paanel under each e condition. Chatter vibration v occcurred at sppindle speedd of 5,000 PM and depth off cut of 1.0 mm m; this machinning condition is described inn Fig. 10. Thee detected chatter frequency was w 2,630 Hz, and a high-ppitched chatteering sound was generated inn the air. Moreover, M it was confirmed thhat a chatter mark appeareed on the surrface of the machhined productt. When the chatter occurrred, the chatter inndicator lightt activated (reed), and the sttable spindle speed was calcculated throuugh the deteected frequency. Spindle speed s of 5,8330 PM is thhe stable spinndle speed, calcculated by thhe chatter frequency f unnder the previoous conditiion. The scope of the recommendeed spindle speeed was betweeen 5,830 PM and 15,164 PM M; thus, 5,8300 PM, whicch was the cloosest ue to 5,000 PM, was selected as th he machiningg valu con ndition. Chatter C vibrattion did not ooccur under th he conditionss of 5,830 5 PM sppindle speed and 1.0 mm depth of cut,, and d the chatter mark m did nott appear on the t machinedd surfface. The peaak frequency ddetected heree was 470 Hz,, whiich is the fiffth content of tool passin ng frequency.. Thee front panel of o this conditiion is illustratted in Fig. 11, and d unlike the front fr panel unnder the cond ditions duringg whiich chatter viibration occuurred, the chaatter indicatorr ligh ht remained green g and thee stable spind dle speed didd not appear. S 4.4 Machining Surface When W chatterr vibration ooccurred, a cross-striped c d chaatter mark apppeared on thee surface; thu us, reliabilityy of VDMS V to disttinguish chattter was confirrmed throughh the surface obsservation. Att spindle speeed of 5,0000 PM M and depth of cut of 1.0 mm, the conditions underr whiich chatter vibbration was ggenerated, thee existence off a ch hatter mark was w confirmed, as shown in the resultss of VDMS. V The surface stattus of this co ondition wass indiicated in Figg. 12, which shows one su urface wheree
228 eal Time Chatter Vibration Control System in High Speed Milling Fig. 11 Front panel of VDMS (spindle speed 5,830 PM, depth of cut 1.0 mm). Fig. 12 Machined surface of workpiece (spindle speed 5,000 PM, depth of cut 1.0 mm). Fig. 13 Machined surface of workpiece (spindle speed 5,830 PM, depth of cut 1.0 mm). the machining was inconsistent and the marks of shaking left and right were generated. The conditions of 5,830 PM spindle speed and 1.0 mm depth of cut are shown in Fig. 13. Here, the chatter mark was not generated, but a subtle trace appeared where the tool passed through. 4. Conclusions In the present study, a dynamic machining system had been developed that can detect chatter vibration in real time and avoid it during the cutting process. The developed program VDMS was designed to determine cutting force and conduct FFT analysis, Impact excitation test, chatter detection, and optimal speed recommendation. By using this program, chatter vibration detection was confirmed throughh a machining experiment. Chatter vibration was detected and stable spindle rotation speed was calculated only with the vibration frequency generated during the machining experiment. Furthermore, it was confirmed that chatter vibration was not generated and stable machining was performed even if the depth of cut was not reducedd during the machining to modify the calculated machining speed. Acknowledgments This research was supported by Doosan Infracore. eferences [1] Jung, N. S. Vibration in 210-17. 2008. Analytical Prediction of Chatter Milling Process. KSME mech. 33 (3):
eal Time Chatter Vibration Control System in High Speed Milling 229 [2] Merritt, H. E. 1965. Analytical Prediction of Stability Lobes in Milling. Annals of CIP 44: 357-62. [3] Yue, J. P. 2006. Creating a Stability Lobe Diagram. IJME Session IT, 301-501. [4] Norizanu, S. and Eiji, S. 2010. Vibration Suppressing Method and Vibration Suppressing Device for Machinine Tool. US 2010/0104388 A1. [5] Lamraoui, M. and Thomas, M. 2014. Indicators for Monitoring Chatter in Milling Based on Instantaneous Angular Speeds. MSSP 44: 72-85. [6] Sergiu, T. C. 1990. Stability in the Dynamics of Metal Cutting. ISBN 20: 0-444-98868-8. [7] Hiromitsu, M. and Toru, Y. 2012. Tracing and Visualizing Variation of Chatter for In - Process Identification of Preferred Spindle Speeds. Annals of CIP 4: 11-6. [8] Norikazu, S., Yusuke, K., Takashi, K., ei, H. and Eiji, S. 2012. Identification of Transfer Function by Inverse Analysis of Self-Excited Chatter Vibration in Milling Operations. Precision Engineering 36: 568-75. [9] Altintas, Y., Eynian, M. and Onozuka, H. 2008. Identification of Dynamic Cutting Force Coefficients and Chatter Stability with Process Damping. Annals of CIP 57: 371-4.