Dynamic sensitivity analysis of CNC machine tools in static state

Transcription

1 Adance in Intelligent Sytem Reearch, lume 136 4th Internatinal Cnference n Senr, Mechatrnic and Autmatin (ICSMA 2016) Dynamic enitiity analyi f CNC machine tl in tatic tate uchu Jiang1,a, inyng Ma1,b,*, Caihua Ha1,c, Huanbin He1,d, Cha Qin1,e, Bin Li1,f 1 Schl f Mechanical Science and Engineering, Huazhng Unierity f Science and Technlgy, Wuhan , China a xuchujiang@hut.edu.cn, bmaxyhut@163.cm, c @163.cm, d hehuanbin@126.cm, e @163.cm, flibin999@hut.edu.cn Keywrd CNC machine tl, Operatinal mde, Dynamic enitiity, Tl wear characterizatin Abtract Thi paper mainly dicue the tl cnditin mnitring baed n dynamic enitiity, and explre a methd fr characterizing tl cnditin characterizatin. Firtly, the enitie directin and cmpnent f different rder lw mde were analyzed in the tatic tate. Mdal parameter and the dminant mde are al identified. Secndly, dynamic enitiity analyi f the tl-wrkpiece ytem i analyzed by uing peratinal mde. Fr the dminant mde, the mdal enitiity i high in bth tatic and dynamic tate. Althugh many factr uch a bundary cnditin change a lt in cutting cnditin, the mdal enitie part d nt change. It indicate that the mdal enitiity i the baic attribute. Therefre, it i reliable t analyze the dynamic enitiity f the tl-wrkpiece ytem by the amplitude change f the peratinal mde during the cutting prce. Intrductin Tl cnditin mnitring i ery imprtant fr prce autmatin. The exceie tl wear can caue the machining dimenin ditrtin, increae the crap rate and the prductin ct [1]. Tl cnditin mnitring ha an imprtant effect n lwing tl wear rate, wrkpiece urface quality cntrl and prce ptimizatin. Tl wear mnitring ha tw baic methd currently, ne i direct mnitring methd, the ther i indirect mnitring methd. Mt f the methd are indirect [2-4]. We can ue iual r ptical ignal t analyze the tl tate in direct mnitring methd. Hweer, it i difficult t be ued in the actual machining prce. The mt widely ued mnitring ignal in indirect mnitring methd are acutic emiin, cutting frce and ibratin acceleratin ignal [8]. The bigget limitatin i nt the ening technlgy but the analyi technlgy [9]. The crrelatin between the change f cutting frce and tl wear ha been widely recgnized. There are me adantage f uing ibratin ignal t mnitr the tl tate, uch a rbutne, high reliability and fat repne, which are ery imprtant fr nline real-time mnitring. The reearch f uing ibratin ignal t carry n the tl cnditin mnitring i dicued in thi paper. There are mainly fur kind f characteritic ignal that characterize the tl wear tate: natural frequency amplitude, high frequency band amplitude, tth frequency amplitude and frequency band energy. Cpyright 2016, the Authr. Publihed by Atlanti Pre. Thi i an pen acce article under the CC BY-NC licene ( 65

2 Adance in Intelligent Sytem Reearch, lume 136 Lim fund that there wa a trng crrelatin between tl wear and natural frequency amplitude f the tl rd [5]. Chelladurai analyzed the crrelatin between peak amplitude f high frequency band and tl wear [1]. Barreir analyzed the crrelatin between maximum amplitude f high frequency band and tl wear [7]. Tma analyzed the crrelatin between PSD energy change f tl tth frequency and tl wear [6]. The preent tl wear identificatin methd hae high crrelatin with the cutting parameter, the applicatin f the identificatin methd i limited. S the trend f tl cnditin mnitring i t find a tl wear identificatin methd which i independent f the cutting parameter. Dimla fund that the tl tip wear might better characterize the tl life than the wear f tl ide rack [1]. S we will mainly tudy the crrelatin between tl tip wear and ibratin ignal f turning tl. Change factr f tl-wrkpiece ytem in cutting prce can lead t the change f lw frequency peratinal mde f CNC machine tl tructure. S we will tudy tl wear f turning prce baed n peratinal mdal analyi (OMA) in thi paper. The dynamic change f the tl-wrkpiece ytem in cutting prce i tudied baed n OMA. A frequency dmain methd named p.plymax algrithm can be ued t prce the PSD f ibratin acceleratin repne ignal. Each rder mde f CNC machine tl can be accurately determined in machining prce. We analyze the change factr enitiity f tl-wrkpiece ytem in different directin a well a each rder lw-frequency peratinal mde f different part in cntinuu cutting prce. Thi i defined a the dynamic enitiity methd. Dynamic enitiity change with the cutting excitatin energy. And the crrepnding directin tructure dynamic enitiity parameter will change ignificantly. Thu we can btain the ariatin f wear. The methd i ignificantly different frm the traditinal pectrum analyi. In thi paper, we will ue dynamic enitiity methd t tudy the tl cnditin mnitring. Experimental deign f dynamic enitiity methd Firtly, a cmmn experimental ytem f impact and cutting experiment i hwn in Fig.1. Fig. 1 Experimental ytem The detail f the experimental ytem are hwn in table 1. 66

3 Adance in Intelligent Sytem Reearch, lume 136 Table 1 Detail f the experimental ytem Cmpnent Machine Tl Data acquiitin and analyi ytem Cutting tl Detail CNC lathe K60 LMS SCADAS Mbile SCM05 & LMS Tet. Lab 10B Wrkpiece Vibratin repne enr 45 teel Mitubihi turning tl,type:ptgnr2525m16 PCB-356A15 the meaurement frequency band i 5000Hz Secndly, cutting a large-diameter wrkpiece i taken a an example t detail the cutting prce ytem and arrangement f enr meauring pint a hwn in Fig.2. Fig. 2 Schematic diagram f Prce ytem and enr meauring pint The meauring pint 1 i n the tl rd near the cutting edge. The meauring pint 2 i n the table. The meauring pint 3 i n the Z table. The meauring pint 4 i n the dwn guide and the meauring pint 5 i n the pindle ide. Dynamic enitiity analyi in tatic tate Thi ectin mainly analyze the dynamic enitiity in tatic tate. The enitie directin, enitie cmpnent and dminant mde f each lw frequency mde in 0-120Hz were analyzed. The lw frequency mdal enitiity analyi in tatic tate utilized the impact experiment data. We impacted the Z directin f the tl, Z directin f table,, Y, Z directin f Z table in the impact experiment, then btained the FRF. Fig.3 hw the impact tet in +Z directin f table. 67

4 Adance in Intelligent Sytem Reearch, lume 136 Fig. 3 Schematic diagram f impact tet The frequency repne functin (FRF) f each meauring pint were btained after the impact tet, and then the teady-tate diagram calculated by PlyMA algrithm i hwn in Fig e- 6 g/ N Am pl i t ude 5. 47e Li near Hz Fig. 4 The teady - tate diagram f the mde in FRF In Fig.4, the left rdinate repreent the amplitude and the right rdinate repreent the calculated mdel rder. The red cure in the figure repreent the calculated FRF. The green cure in the figure repreent the mdal indicating functin. When the mdal indicating functin i cle t zer, it mean that the crrepnding frequency i likely t be an rder mde. Dynamic enitiity analyi f the lw frequency mde in tatic tate. The enitiity analyi i cnducted mainly frm the enitiity directin, the enitiity part and the dminant mde. The principle f judgment i: Accrding t the cmprehenie analyi f the FRF f different meauring pint, if the mde peak exit in the certain directin regardle f different meauring pint, and the amplitude i relatiely large, then the mde i enitie t thi directin. With the ame impact, the cmpnent with the larget FRF amplitude i cnidered a the enitie part. Accrding t thi principle, the dminant mde can be determined. 68

5 Adance in Intelligent Sytem Reearch, lume 136 Fig. 5 FRF f the meauring pint when impacting Z directin f table Fig.5 hw that the amplitude f 30Hz and 36Hz are ignificant, they are bth enitie t the Z directin. The amplitude f 30Hz i the larget at the pindle meauring pint, it i enitie t the meauring pint n the pindle. The amplitude f the 36Hz i the larget at the impact pint lcated n table and the amplitude i the ecnd n the pindle away frm the impact pint. S the 36Hz i al enitie t the pindle. Thi i an interactin that take the three-dimeninal excitatin f the tl tip, the tructural ibratin cmplexity a well a the tranfer functin between the ariu meauring pint int accunt. The aim i t determine the enitie mde f the different cmpnent. Mdal Parameter and Senitiity Summarizatin in Lw Frequency Band f CNC Machine Tl By analyzing the frequency teady-tate diagram f each directin and each meauring pint in the fie impact experiment and cmbining with the enitie cmpnent a well a enitie directin f each mde, the reult are ummarized in Table 2. Table 2 9 rder mde and enitiity in 0-120Hz frequency band f the CNC lathe Mde parameter Order Firt rder Secnd rder Third rder Frth rder Fifth rder Sixth rder Seenth rder Natural Frequency Damping (Hz rati % Senitie directin Senitie cmpnent Z Z pindle pindle pindle tl tl pindle Table 2 hw that 35Hz i the dminant mde f the pindle in Z directin, while 85Hz i the dminant mde f the tl in directin and 19Hz i the dminant mde f the pindle in directin. The dminant mde in directin i 85 Hz, and the dminant mde in Z directin i 36 Hz. 69

6 Adance in Intelligent Sytem Reearch, lume 136 Cncluin In tatic tate, 85Hz f directin i the mt enitie t the tl. In Hz frequency band, the mdal enitie directin in the lwer frequency band Hz are mainly Z-directin f the pindle. The mdal enitie directin in the higher frequency band Hz are directin f the tl. And the lwet mde 19Hz and the highet mde 105Hz are directin f the pindle. S the mdal enitiity ha a certain regularity. The dynamic enitiity analyi under perating cnditin wa carried ut. Becaue the rean fr pace i nt detailed. Fr the dminant mde, the mdal enitiity i high in bth tatic and dynamic tate. Althugh many factr uch a bundary cnditin change a lt in cutting cnditin, the mdal enitie part d nt change. It indicate that the mdal enitiity i the baic attribute. Therefre, it i reliable t analyze the dynamic enitiity f the tl-wrkpiece ytem by the amplitude change f the peratinal mde during the cutting prce. Acknwledgement The reearch i upprted by the Natinal Natural Science Fundatin f China under Grant N and , and the Key Prject in the Natinal Science & Technlgy Pillar Prgram f China under Grant n. 2015Z Reference [1] D.E. Dimla, P.M. Liter, On-line metal cutting tl cnditin mnitring. Frce and ibratin analye, Internatinal Jurnal f Machine Tl and Manufacture 40 (2000) [2] S. Puruhthaman, Y.G. Sriniaa, A back-prpagatin algrithm applied t tl wear mnitring, Internatinal Jurnal f Machine Tl and Manufacture 34 (1994) [3] I.N. Tanel, C. Mekdeci, C. Mclaughlin, Detectin f tl failure in end milling with waelet tranfrmatin and neural netwrk (WT-NN), Internatinal Jurnal f Machine Tl and Manufacture 35 (1995) [4] O. Mary, Detectin f tl wear uing multienr reading defued by artificial neural netwrk. Internatinal Sciety fr Optic and Phtnic1991, pp [5] S.S. Rangwala, D. Drnfeld, Integratin f enr ia neural netwrk fr detectin f tl wear tate, Intelligent and Integrated Manufacturing Analyi and Synthei (1987) [6] D.E. Dimla, Multiariate tl cnditin mnitring in a metal cutting peratin uing neural netwrk [Unierity f Wlerhamptn1998. [7] S. Rangwala, D. Drnfeld, Senr integratin uing neural netwrk fr intelligent tl cnditin mnitring, Jurnal f Engineering fr Indutry 112 (1990) [8] K. Zhu, Y. San Wng, G.S. Hng, Waelet analyi f enr ignal fr tl cnditin mnitring: a reiew and me new reult, Internatinal Jurnal f Machine Tl and Manufacture 49 (2009) [9] P.M. Liter, On-line meaurement f tl wear