Development of a test method for the comparative analysis of bicycle saddle vibration transmissibility

Transcription

1 Avalable onlne at Proceda Engneerng 60 ( 2013 ) th Asa-Pacfc Congress on Sports Technology (APCST) Development of a test method for the comparatve analyss of bcycle saddle vbraton transmssblty Ncola Petrone a* and Federco Gublato a a Department of Industral Engneerng, Unversty of Padova, Va Veneza 1, Padova, 35100, Italy Receved 20 March 2013; revsed 6 May 2013; accepted 9 June 2013 Abstract Vbratonal comfort s one of the most mportant parameters evaluated by cyclsts durng bcycle rdng. A method for measurng the saddle vbraton transmssblty was developed wth the am of comparng the comfort propertes of bcycle saddles wthout any nfluence of the full bcycle frame, usng a quanttatve approach. Three dfferent bcycle saddles were mounted on a stff seatpost clamped to the extremty of a vertcal servohydraulc cylnder: saddles were loaded by a UNI standard wooden dummy bottom carryng deadweghts. After applyng a snusodal sweep of 0.6mm ampltude to the cylnder, from 1 to 100 Hz, wth steps of 2.5 Hz, the applcaton of two pezoelectrc accelerometers allowed evaluatng the magntude of the transfer functon H between the nput acceleraton at the cylnder shaft and the output acceleraton at the dummy, n the range Hz. The three curves of the transfer functon magntude were compared and analyzed n order to ntroduce a Vbratonal Comfort Index and to rank the tested saddles from the engneerng quanttatve pont of vew The Authors. Publshed by Elsever Ltd. Open access under CC BY-NC-ND lcense. Selecton 2013 Publshed and peer-revew by Elsever under responsblty Ltd. Selecton of the and School peer-revew of Aerospace, under Mechancal responsblty and of Manufacturng RMIT Unversty Engneerng, RMIT Unversty Keywords: Bcycle; saddle; comfort; vbratons; transmssblty 1. Introducton Comfort s one of the parameters evaluated by cyclsts durng bcycle rdng. It nvolves multple aspects such as the poston of cyclst on the bcycle, the feelngs at the bcycle-body nterface and the vbratons perceved by the cyclst durng cyclng on rregular roads. Focusng on the body-bcycle nterface, the pressure dstrbuton n the perneal regon can be correlated wth a pressure comfort ndex, whereas the dynamc transmsson of vbraton to the body can be assocated wth vbratonal dscomfort. * Correspondng author. Tel.: ; fax: E-mal address: ncola.petrone@unpd.t The Authors. Publshed by Elsever Ltd. Open access under CC BY-NC-ND lcense. Selecton and peer-revew under responsblty of the School of Aerospace, Mechancal and Manufacturng Engneerng, RMIT Unversty do: /j.proeng

2 Ncola Petrone and Federco Gublato / Proceda Engneerng 60 ( 2013 ) The entty of vbratons perceved by the cyclst durng road cyclng at a gven speed on a specfc road surface depends on the geometry of the bcycle and on the mass, nerta and structural characterstcs of ts man components [1-3]. Beng the last bcycle component to be n contact wth the body bottom tssues, the saddle s a key element n the enhancement of vbratonal comfort, together wth cyclng shorts and ther paddng [7]. The modern constructon of bcycle saddles expresses the research of manufacturers towards saddle concepts, constructon and producton technologes that wll be able to satsfy the customers requrements of optmal pressure dstrbuton, vbraton dampng, frcton propertes and thermal/perspraton qualtes, all combned wth the mnmum cost and mass. The Qualty Functon Deployment (QFD) approach can successfully support the development of better products after the dentfcaton of a matrx correlatng the customers qualty requrements to the products engneerng characterstcs: the latter are the results of quanttatve laboratory tests that shall be avalable to the product developers for comparng the products. Am of ths work was to develop an expermental method enablng to measure and compare the saddle vbraton transmssblty as an engneerng product characterstcs. On the bass of former experences appled to the study of bcycle wheels [4, 5] and of bcycle shorts [7] a new laboratory test method was developed to evaluate the vbraton transmssblty of dfferent bcycle saddles models n terms of trend and magntude of ther transfer functons. 2. Materals 2.1 Tested saddles Three racng bcycle saddles (named A, B, C) were selected for the study. Saddles were dfferent for mass, ral materal, shell constructon and paddng thckness. The three saddles were chosen to come from dfferent manufacturers and to have mass dfferences not greater than 11% as well as dfferent constructve solutons (posteror splts) or materal solutons (gel nserts). Table 1. Geometrc and constructve descrpton of the tested saddles. Saddle Type Length [mm] Wdth [mm] Heght [mm] Mass [g] Saddle A 272,0 155,0 46,1 210 Saddle B 260,0 136,0 39,2 196 Saddle C 280,0 142,0 53,0 220 Rals Materal Ttanum alloy Vanox alloy CroMoly alloy Shell Materal Carbon fber Short posteror splt Carbon fber Long posteror splt 10% Carbon fbre Central Gel Insert Paddng thckness Low Low Medum 2.2 Test bench and nstrumentaton A commercal alumnum seatpost was cut at about 50mm from the saddle clamp and was welded to an adaptor alumnum plate enablng to fx t to the 15 kn load cell of a 200 mm stroke vertcal MTS 242 servohydraulc cylnder. The vertcal axs of the cylnder was therefore ntentonally n lne wth the center of the saddle clamp (see Fgure 1.a). The statc vertcal load actng on the saddle was appled by a wooden dummy bottom shaped accordng to UNI 10814, typcally used for the durablty tests of chars. The dummy was free of swvel about three axs by means of a ball-socket jont connected to a steel swngng arm hnged horzontally to a fxed structure 1.3 m apart. The dummy was restng on the saddle and t was loaded by a set of dead weghts n order to have a total statc load of 50 kg on the saddle, as measured by the cylnder load cell (see Fgure

3 290 Ncola Petrone and Federco Gublato / Proceda Engneerng 60 ( 2013 ) b). The dummy poston of the saddle was ntally adjusted after placng a Novel Plance bke saddle mat over each saddle and measurng the pressure dstrbuton wth a cyclst seated on a racng bcycle properly adjusted. The saddles were consstently clamped n the mddle of the rals, wth null roll & ptch angles of the upper tangent plane wth respect to the horzontal, measured wth a dgtal nclnometer. The dummy longtudnal poston wth respect to the cylnder was adjusted n order to reproduce as much closely as possble the pressure dstrbuton recoded on the saddle wth a real cyclst (Fgure 1.c). Acceleratons at the support plate (nput) and at the dummy bottom upper surface (output) were recorded usng two unaxal pezoelectrc accelerometers (SoMat SAPE HLS 1010, +/- 50 g full scale, Hz bandpass) and a SoMat EDAQ lte at 5 khz per channel Fg. 1. (a) Expermental test setup: (1) Dead weghts of 34.5 kg (2) Swngng arm and ball jont fxaton system (3) Output accelerometer (4) Bottom shaped dummy (5) Saddle under test. (6) Dummy Seatpost (7) Input accelerometer (8) Load cell (9) Servohydraulc cylnder. (b) Inferor vew of the Dummy-Saddle nterface. (c) Comparson between the Dummy Bottom (left) and a cyclst pressure dstrbuton (rght) on the same saddle C. 3. Methods Each saddle was excted by the actuator shaft dsplacement at constant ampltude (0.6 mm) and frequency rangng from 1 to 100 Hz wth dscrete ntervals of 2.5 Hz after 2.5 Hz. The actuator shaft acceleraton a n and the dummy bottom vertcal acceleraton a out were respectvely the nput and the output varables of the dynamc system consttuted by the seatpost, the saddle and the dummy. A frequency analyss of both actuator shaft (nput) and Dummy (output) acceleraton sgnals was performed for each command block. Dscrete Fourer Transform DFT of sgnals was computed by FFT algorthm. In order to obtan leakage-free spectrums wth 0.1 Hz resoluton, a total of samples were consdered on each FFT executon, correspondng to an nteger number of nput sne functon perods. The comparson between the ampltude of each nput acceleraton spectrum A n (f) obtaned from the measured sgnal analyss and the respectve theoretcal value was performed to verfy the correct executon of the related command block.

4 Ncola Petrone and Federco Gublato / Proceda Engneerng 60 ( 2013 ) Transfer functon magntude H k (f) was calculated for each k-th tested saddle, pont by pont at each exctaton frequency f. Each transfer functon value H k (f ) was calculated as the rato between the output spectrum magntude peak value A out,k (f ), calculated at the exctaton frequency, and the correspondng nput spectrum magntude A n,k (f ) as expressed n (1). H k (f ) = A A out,k n,k ( f ( f ) ) (1) Transfer functons obtaned from equaton (1) were weghted by the frequency weghtng curve W K proposed by ISO [8], whch expresses the human body sensblty n seated poston along the vertcal Z axs (Fgure 2.a). The percent relatve dfference r (f ) between maxmum and mnmum values of H k (f ) computed for tested saddles was calculated (3) at each frequency n order to evaluate how dfferences between vbratonal transmssblty of dfferent saddles vared wthn the consdered frequency range. ( H k ( f )) mn H k ( f ) max H ( f ) ( ) ( ) max ε ( f ) = 100 (2) r k A Vbratonal Comfort Index, based on the nverse value of the area below the transfer functon curve n the 1- F Hz nterval, was ntroduced as a possble engneerng characterzaton of saddles (3). Both the orgnal transfer functons and the ISO weghted transfer functons (Fgure 2.b) were used for the ndex calculaton. 1 VCIF = F = 100Hz F H() f df 1, 40Kz From the observaton that only n the nterval between 1 Hz and 40 Hz the weght functon showed values hgher than -10 db, the VCI values were calculated over the full range (F = 100 Hz) and the smaller range of hgher human sensblty (F = 40 Hz). 4. Results and dscusson The shape of the curves orgnally obtaned for the three saddles resulted to be smlar: t was characterzed by a man resonance peak around Hz, wth a transmssblty magntude around 20 db, and mnor peaks between 25 and 35 Hz (Fgure 2.a) where transmssblty magntude resulted around -10 db. From 20 Hz onwards, all saddles tended to attenuate vbratons and the transmssblty magntude were all lower than - 40 db at 100 Hz. The larger dfferences between curves appeared around the man peaks and towards 100 Hz. Saddle C showed a transmssblty curve consstently below the other two saddles, n the range 10 Hz-100 Hz: on the contrary, ts curve was hgher from 2.5 Hz to 10 Hz. Curves of saddle A and B were almost overlappng on all the range, wth saddle B showng a smaller transmssblty from 70 to 100 Hz. The rankng of saddles vbraton based on the transmssblty curves resulted to vary wth the frequency range consdered: after the calculaton of the drect Vbratonal Comfort Index, spannng over all the Hz range, saddle C resulted to have the hghest Vbratonal Comfort Index and saddle A the lowest. Saddle B had a VCI 100 that was +4.6% greater than A and saddle C VCI 100 was +20.8% greater than A. Lower dfferences were found when the ISO weghted curves (Fgure 2.b) were consdered for the VCI W100 computaton: saddle B had a VCI W100 that was +7.8% greater than A and saddle C VCI W100 was (3)

5 292 Ncola Petrone and Federco Gublato / Proceda Engneerng 60 ( 2013 ) % greater than A. An nterestng result was that when usng the ISO weghted curves, saddle B had a behavour much closer to saddle C than to saddle A, whereas saddle B was closer to saddle A f the ndex was calculated on the orgnal unweghted curves. A smaller effect was ntroduced by the change of nterval of ntegraton, from 100 Hz to 40 Hz, as shown n Table 2. Fg. 2. (a)the resultng curves of the three saddles transfer functons and the frequency weghtng curve accordng to ISO ; (b) The resultng ISO weghted curves of the three saddles transfer functons. Table 2. The comparson of Vbratonal Comfort Index calculated wthn 100 Hz or 40 Hz, on the orgnal and ISO weghted transfer functons. Saddle VCI 100 VCI 40 VCI W100 VCI W40 A B (+4.6%) (+6.3%) (+7.8%) (+6.3%) C (+20.8%) (+17.4%) (+10.1%) (+9.0%) These results suggest the need of performng a subjectve evaluaton test sesson n order to defntely correlate the cyclsts saddle comfort evaluaton results wth the engneerng results and to dentfy whch VCI F formulaton expresses n the most proper way the cyclst s feelngs. Despte these nterestng results testfy the ablty of the developed test method n quantfyng the dfferences among dfferent saddles from the vbratonal pont of vew, t s also worthy to hghlght the lmtatons of the study. A frst lmtaton can be recognzed n the fact that the test setup s conventonal and t nvolves the use of a wooden dummy that, despte shaped n accordance wth an nternatonal standard, t has not the elastc, dampng and frcton propertes of a human bottom. Ths can be seen as a dsadvantage wth respect to the am of characterzng the saddles wth an engneerng characterstcs that should ntentonally correlate wth the perceved vbratonal comfort: however, techncally speakng, t was ntentonally decded to tests the saddle nvolvng the lowest number of external factors that can nfluence the tests results and wth the most repeatable and reproducble set of test tools. For ths reason, the adopton of a more anthropomorphc dummy, wth a replca of the pelvs bone geometry and of the human soft tssues by a cast of slcon or other materals was not consdered at ths stage: prevous experences n ths drecton regardng the mechancal characterzaton of seat cushons showed that t can be very dffcult to reproduce the dummy consstency

6 Ncola Petrone and Federco Gublato / Proceda Engneerng 60 ( 2013 ) throughout the world and that the durablty of the dummy surface and bulk propertes can be hardly acheved. The second lmtaton that can be hghlghted regards the small number of tested saddles: as usually occurs, the test method was recently developed and s now beng tested on a wder number of saddles. The man lmtaton of the work can be found n the absence, for the moment, of nformaton regardng the perceved comfort behavour of the saddles, based on customers questonnares: ths wll be the frst further development of the work, n order to check f the VCI ntroduced n the work s clearly correlated wth a vbratonal perceved comfort score of the saddle. A second nterestng development of ths study wll consst n the correlaton of the vbraton transmssblty curves wth constructve and structural saddle characterstcs n order to support the manufacturers process towards more comfortable products fulfllng the customers requrements. 5. Conclusons The works report the expermental method developed for the engneerng characterzaton of bcycle saddles n terms of vbratonal transmssblty. The use of a vertcal servohydraulc cylnder holdng the saddle, the applcaton of a standard bottom shaped dummy loaded on the saddle and the use of two accelerometers allowed to acqure the transmssblty curves of three saddles and to ntroduce an ntegral Vbratonal Comfort Index (VCI) to rank the saddles. Saddle C resulted to perform as the best saddle, showng a VCI % hgher than the worst performng saddle, saddle A. Results of the present work showed that developed method was able to detect dfferent saddle vbraton transmssblty and to support a QFD approach to the mplementaton of nnovatve and performng bcycle saddles. References [1] Stone C., Hull M.L., Effect of rder weght on rder-nduced loads durng common cyclng stuatons, Journal of Bomechancs, 28 (4), [2] Hastngs A.Z., Blar K.B., Cullgan F.K., Pober D.M., Measurng the effect of transmtted road vbraton on cyclng performance, The Engneerng of Sport 5, Davs, [3] N. Petrone, F. Gublato. Methods for evaluatng the radal structural behavour of racng bcycle wheels, The 5th Asa Pacfc Congress on Sports Technology (APCST2011). [4] Petrone N., Gublato F., Comparatve analyss of wheels vbraton transmssblty after full bcycle laboratory tests, 40 Convegno Nazonale AIAS, 7-10 settembre 2011, Palermo. [5] Federco Gublato, Ncola Petrone (2012). A method for evaluatng the vbratonal response of racng bcycles wheels under road roughness exctaton. In: -. Proceda Engneerng. PROCEDIA ENGINEERING, vol. 34, p , ISSN: , Lowell, Massachusetts, 9-13 July 2012, do: /j.proeng [6] Mark Oleman, Raluca Marn-Peranu, Mha Marn-Peranu, (2012). Measurement of dynamc comfort n cyclng usng wreless acceleraton sensors. In: -. Proceda Engneerng. PROCEDIA ENGINEERING, vol. 34, p , ISSN: , Lowell, Massachusetts, 9-13 July 2012, do: /j.proeng [7] G Marcoln, A Paol, FA Panzzolo, G Basco, N Petrone, A method for the analyss of cyclst shorts wth dfferent pads for perneal area protecton: comparson between drum and road tests. Proceda Engneerng, Volume 2, Issue 2, June 2010, Pages [8] ISO , Evaluaton of human exposure to whole-body vbratons, 1997.