6 Vbf. 5 No. August 2003 INFLUENCE OF THE HEAD RESTRAINT POSITION ON DYNAMIC RESPONSE OF THE HEAD/NECK SYSTEM UNDER WHIPLASH LOADING SHYH-CHOUR HUANG, RONALD L. HUSTON 2 department of Mechancal Engneerng, Natonal Kaohsung Unversty of Appled Scences, Kaohsung, Tawan department of Mechancal, Industral and Nuclear Engneerng, Unversty of Cncnnat, Cncnnat, OH, USA ABSTRACT The objectve of ths paper s to present modelng and smulaton of the effect of head restrant poston on head/neck dynamcs n rear-end motor vehcle collsons. Although ndvdual njury tolerance levels vary, t t s s beleved that properly postoned head restrants can be benefcal n reducng njury. The Tle paper dscusses the effects of of restrant postonng by hy smulatng a seres of of rearend collsons usng a fnte-segment (lumped-mass) model of of the human frame. It It s s found that proxmty of the restrant to the head s s the prncpal factor n n preventng harmful whplash moton. The fndngs suggest that "smart" head restrants could therefore sgnfcantly reduce whplash nduced njures. Bomed Eng Appl Bass Comm, 2003 (August); (August): 5: 6-69. 6-569. Keywords: whplash, head restrant, head/neck model, smart systems. INTRODUCTION "Whplash" s defned both as an njury and as a moton, or knematc, phenomena. As an njury, whplash s defned as []: a neck njury caused by a sudden jerkng backward, forward, or backward and forward of the head as durng an automoble accdent (Random House College Dctonary). As a moton, whplash s a rearward (extenson-tenson) movement of the head/neck system followed mmedately by a forward (flexon-tenson) movement. Receved: Aug, 2003; Accepted: Aug 8, 2003 Correspondence: Shyh-Chour Huang, Professor Department of Mechancal Engneerng Natonal Kaohsung Unversty of Appled Scences 5 Chen Kung Road, Kaohsung, Tawan 807 E-Mal: shuang@cc.kuas.edu.tw Whplash njures are prevalent. They account for more than half of the automoble nsurance clams n Japan and Canada [7-8]. In the Unted States whplash accounts for more than 25 percent of automoble accdent njures [], Although the symptoms assocated wth whplash can be descrbed (neck pan, headache, muscle soreness) the understandng of the mechansm of njury s less clear. But understandng the mechansm of njury s clncally mportant [2-6]. Of even greater nterest are the means of avodng and preventng whplash njures. Head restrants provde an obvous nstrumentalty n reducng and preventng njures. But for head restrants to be effectve the vehcle occupant needs to have hs or her head near the restrant. Thus, n a rear end collson accdent, t s mportant to mnmze the dstance between the occupant's head and the head restrant, The nterest n whplash njures has stmulated a -32
BIOMEDICAL ENGINEERING- APPLICATIONS, BASIS & COMMUNICATIONS 65 number of studes some even nvolvng volunteers [9-ll]. However, tests wth volunteers need to be lmted to low velocty changes (low "delta v") so that serous njures of the volunteers are avoded. Consequently, data from these tests have a somewhat lmted range of applcablty. An alternatve to volunteer testng s the use of cadavers and dummes. Here, however, the data s also of lmted value snce cadaver and dummy head/neck propertes are dfferent than those of actual vehcle occupants. An attractve alternatve to both volunteer and surrogate tests s the use of bodynamc computer models. Such models, whch consst of rgd and flexble bodes representng the human frame, have been shown to be very effectve n smulatng vehcle occupant knematcs n crashes and n hgh acceleraton envronments. [2-6]. In ths paper we present results of research studyng the numercally smulated response of vehcle occupants n rear end collsons, wth a varety of head restrant postonngs. The balance of the paper tself s dvded nto four sectons wth the followng secton presentng a descrpton of the vehcle-occupant model. The next secton provdes a descrpton of the analyss procedures. The subsequent secton contans the smulaton results and the fnal secton provdes a dscusson and concludng remarks. 2. THE VEHICLE-OCCUPANT MODEL The model has two parts: () the human model of a seated vehcle occupant, and (2) the vehcle nteror model representng the effect of the vehcle nteror on the vehcle occupant durng collsons. The human body model conssts of 5 rgd bodes representng the torso, the head, the neck, the feet and the upper and lower arms and legs. The body segments are connected by sphercal (ball-and-socket) jonts except for the elbows and knees whch are modeled as revolute (hnge) jonts. The ball-and-socket jont s descrbed by three relatve orentaton varables, and the hnge jont by a sngle rotaton varable. These jonts are provded wth nonlnear sprngs and dampers to smulate the behavor of the soft tssue (muscle and tendon), and to lmt the range of moton. The jonts are also provded wth dampng restrants to lmt jont angle rotatons. These restrants are modeled by moments of the form where c s a constant and where 9 max and 6 mn are the maxmum and mnmum values of the rotaton angle. Fgure provdes a representaton of the model. The human model's nteracton wth constrant planes, such as those representng the head restrant and back rest, s sensed by proxmty of the human body segments and the planes of the head restrant and the backrest. The forces appled to the segment are then functons of the depth of penetraton of the colldng segment wth the constrant plane. These forces are represented on the ndvdual segments by a force passng through the segment mass center together wth a couple. The model has two ways of generatng contact forces: () By measurng the speed of mpact (a dampng force) and (2) by measurng the depth of penetraton (a sprng force). These are descrbed n the followng paragraphs: () Dampng force development The relatve velocty between the contactng human body segment and the constrant plane can be resolved nto a normal component Av n anda tangental component Av,. Then the dampng force F d s defned as: F d = cav n (2) where c s a dampng coeffcent. Durng a confguraton wth ncreasng penetraton, the dampng force s exerted opposte to Av n. Alternatvely, when the human segment s reboundng from the constrant (decreasng penetraton) no dampng force s appled. (2) Sprng force development The sprng force p s lnearly proportonal to the penetraton dsplacement and the stffness of the constrant surface. As wth the dampng force the sprng force s exerted through the mass center G of the contactng body segment wth the couple torque T beng the vector product of the poston vector?' locatng G relatve to the contact pont and the sprng force vector. The sprng force may be expressed n components normal and tangental to the constrant surface as: F s = F n a n +^lf n u, (3) where f and t are normal and tangental unt vectors. F n s the normal sprng force component, and where V s the coeffcent of Coulomb frcton. Then f s M = 0 99<0 cee ee > o -cee ee>o or e mm <e<e max and 6 < 6 n and 9 > 9 () T =rxf () 33-
66 Vol. 5 No. August 2003 3. DYNAMICAL ANALYSIS The model of Fgure s a multbody system. Effcent analyss procedures have been developed for generatng the governng equatons of moton of such systems. The volumes of Huston (990) and of Josephs and Huston (2002) document these procedures. Let X r (r=l,...,n) be varables (geometrc parameters) descrbng the confguraton and degrees of freedom of the human body model, wth n beng the number of degrees of freedom. Then the angular veloctes cb k (k=l,...,n) of the segments of the model and the mass center veloctes v k (k=l,...,n) of the segments n an nertal reference frame may be expressed n the forms: x and v., = v v ^ * m k * krm x rn (5) where N s the number of body segments, the m (m=l,2,3) are mutually perpendcular unt vectors fxed n the nertal frame, and where the coeffcents GJ^ and Vkrm are components of the partal angular velocty and the partal velocty vectors of the segments and ther mass centers [9]. These components and ther dervatves play a central role n the analyss. By dfferentatng n Equaton (5) the angular acceleraton of the model segments and ther mass center acceleratons may be expressed as: «k =(»k m,x r +a> krm x r )n m and a k =(v km x r +v krm x r )n m (6) Consder the model of Fgure to be subjected to forces from gravty, from the seat, and from collson surfaces of the occupant compartment (steerng wheel, door, wndsheld, roof, etc.). Such forces, as they are exerted on segments of the model, may be represented by an equvalent force system consstng of a force F passng through the segment mass center, together wth a couple wth torque f. Then the generalzed appled ("actve") force F r for x r may be expressed as [7] *\ V krm^km + krm ^km where F km and T km are n m components of F k and T k (7) Smlarly, let the nerta forces on the model be represented by equvalent force systems on the segments consstng of forces F k * passng through _the mass center together wth a couple wth torque T k. Then F k * and T* may be expressed as [9]. and Fk = -m k a k f k *=-Ik-a k -co k x(k-ro k ) (8) (9) where m k s the segment mass and I k s the central nerta dyadc of the segment. The generalzed nerta ("passve") force F r * for x r may then be expressed as: ^r V krm^km "*" km ^km (0) where F km and T^, are the n m components of F k and T k. Usng Kane's equatons [see agan, Kane and Levnson (985)] the governng dynamcal equatons of moton are F r +F r *=0 (r = l,...,n) () By substtutng from Equatons (5) through (0), the dynamcal equatons may be wrtten n the form: A^+B^+C^x^F, r=l,.,n (2) where the coeffcents Arp, Brp, and Gpr are and A rp = m k v k m V k p n +I kmn CO krm <Q kpn B rp = m k V krmv kpm + I kmn <0 km <O kpn ^rpq e rs m lksn U^kqm kpn Head Restrant Seatback Vv V^ VW)( w ( ] (InertaFrame) X B 0 M C B <f E3 ) E2 "^T (Bl B, C B T Fg. Human Body/Seat/Head Restrant Model. (3) () (5) -3-
BIOMEDICAL ENGINEERING- APPLICATIONS, BASIS & COMMUNICATIONS 67 where the l ksn are n s,n n components of Ik and e s the permutaton symbol defned as: s c r.s, = l/2(r- s)(s - m)(m - r) (6) The generalzed actve forces F r also nclude nternal forces representng the contrbuton of the jont forces and moments due to soft tssue (muscles, lgaments, and dsks). Equatons (2) form a system of n second-order, ordnary dfferental equatons for the n x r defnng the poston and movement of the system. Once the physcal data, ntal values of the x r, and the vehcle moton are specfed, Equatons (2) may be ntegrated numercally. The physcal data conssts of body segment dmensons, mass and nerta values for the segments, and characterstcs of the seat and head restrant. The ntal values of the x r defne the ntal posture of the model. Fnally, the vehcle moton s convenently represented by acceleraton/tme data.. SIMULATION OF REAR END COLLISIONS To smulate occupant dynamcs durng rear-end collsons we subjected the model to a frontal deceleraton pulse. The constrant surfaces were the seatback and the head restrant. To determne and document the protectve effect of the head restrant we vared ts horzontal and vertcal postons, as llustrated n Fgures 2 and 3. Fgures and 5 show the results of the smulaton for the varous horzontal head restrant postons. Specfcally Fgure shows the head angular "\.! -.\.\J: wvj r, /' I- '.I ' I : IV :Tv Fg 3. Vertcal Head Restrant Poston. - * 20 5 0 5 0-5 -0 so j /\ V (V V \ \\> s* ~~~ ~- 00 * ^< )J*<!.^ _ 20 Imum.lll^W. Fg. Head Angular Acceleraton Comparson (Horzontal Head Restrant Poston Varaton). - «> r -.» j,0 I o -0-20 50 / t A \- ^ \\ / ^A- K \^ A. 2 ' 2 Bl >lu:< Fn:r.- -30 L lnvmwanj Fg 2. Horzontal Head Restrant Poston. Fg 5. Head Angular Rotaton Comparson (Horzontal Head Restrant Poston Varaton). -35
68 Vol. 5 No. August 2003 acceleraton whle Fgure 5 shows the head rotaton. It s found (as mght be expected) that the head restrant poston closest to the occupant model head yelds the least head angular acceleraton and also the least head rotaton. Smlarly, Fgures 6 and 7 show the smulaton results for varous vertcal head restrant postons. Specfcally, Fgure 6 shows the head angular acceleraton whle Fgure 7 shows the head rotaton. It s seen that the hghest head restrant poston yelds the lowest head angular acceleraton and rotaton. 5. CONCLUSION These results demonstrate not only the mportance of havng a head restrant but also the mportance of the poston of the head restrant n preventng! I 20 r 5 0 5 0 5-0 ^y 5) A. \K 00 * *$%~^*\ 2UI hrm-lm.ll^-vohll Fg 6. Head Angular Acceleraton Comparson (Vertcal Head Restrant Poston Varaton). 0 30! 20 ", % o -0-20 -30.!) KM) rltktlm.lllm.-l A- Fg 7. Head Angular Rotaton Comparson (Vertcal Head Restrant Poston Varaton). 2 excessve head acceleraton and movement. That s, the hgher the restrant and the closer t s to the head, the better the protecton provded by the restrant. Ths n turn mples that vehcle occupants wth manually adjustable head restrants should adjust them as hgh and as forward as possble. The results also mply that seats havng ntegral head restrants (bult-n restrants), assumng they are suffcently hgh and forward, are to be preferred over manually adjustable restrants, snce many users tend not to adjust manual restrants. Fnally, the results demonstrate the potental utlty of "smart" head restrants whch automatcally adjust close to an occupant head at the begnnng of a rear-end collson. ACKNOWLEDGEMENT The authors gratefully acknowledge support for ths work from the Natonal Scence Councl of the Republc of Chna under grant NSC 90-223-E-5-005. REFERENCES l.harrton, M. B., 989. The Whplash Handbook. Charles C. Thomas, Sprngfeld, IL, p. 6. 2. Bogduk, N., 986. The anatomy and pathophysology of whplash. Clncal Bomechancs (), 92-0. 3. Deng, V. C, Goldsmth, W., 987. Response of a human head/neck/upper-torso replca to dynamc loadng -- I Physcal model and II Analytcal/numercal model. Journal of Bomechancs (20), 7-97.. Walker, L. B., Harrs, E. H., et al., 973. Mass, volume, center of mass, and mass moment of nerta of head and neck of the human body. In: Proceedngs of the 7th Stapp Car Crash Conference, SAE, Warrendale, PA, pp. 525-537. 5. Luo, Z. and Goldsmth, W., 99. Reacton of a human head/neck/torso system to shock. Journal of Bomechancs (2), 99-50. 6. Stemper, B. D., Kumaresan, S., et al., 2000. Headneck fnte element model for motor vehcle nerta] mpact: materal senstvty analyss. Bomedcal Scence Instrumentaton (36), 33-335. 7. One, K., and Kanno, M., 993. Influences of physcal parameters on the rsk to neck njures n low mpact speed rear-end collsons. In: Proceedngs of the Internatonal Conference on the Bomechancs of Impacts, pp. 20-22. -36
BIOMEDICAL ENGINEERING- APPLICATIONS, BASIS & COMMUNICATIONS 69 8. Lubn, S., and Schmer, J., 993. Are automoble head restrants used effectvely? Canada Famly Physcan (39), pp. 58-588. 9. Ewng, C. L. and Thomas, D. J., 972. Human head and neck response to mpact acceleraton. Monograph 2, Naval Aerospace Medcal Research Laboratory Detachment, New Orleans, LA. 0. Szabo, T. J., Welcher, J. B., et al., 99. Human occupant knematcal response to low speed rearend mpacts. Socety of Automotve Engneers (SAE) Paper 90532.. Matsushta, T., Sato, T. B., et al., 99. X-ray study of the human head neck moton due to head nerta loadng. Socety of Automotve Engneers (SAE) Paper 92208. 2. Huang, S. C, 995. Modelng human body moton wth applcaton n crash vctm smulaton. Journal of Appled Bomechancs (), pp. 322-336. 3. Huang, S. C, 995. Bomechancal modelng and smulaton of automoble crash vctms. Computers and Structures (57), pp. 5-59.. Huston, J. C, Harlow, M. W., and Huston, R. L., 978. A Comprehensve three-dmensonal head neck model for mpact and hgh acceleraton studes. Avaton, Space and Envronmental Medcne (9), pp. 205-20. 5. Ten, C. S. and Huston, R. L., 987. Numercal advances n gross-moton smulaton of head/neck dynamcs. Journal of Bomechancal Engneerng (09), pp. 62-68. 6. Huston, R. L., 987. Crash vctm smulaton: Use of computer models. Internatonal Journal of Industral Ergonomcs (), pp. 285-29. 7. Huston, R. L., 990. Multbody Dynamcs. Butterworth-Henemann, Stoneham, MA. 8. Josephs, H., and Huston, R. L., 2002. Dynamcs of Mechancal Systems, CRC Press, Boca Raton, FL. 9. Kane, T. R., and Levnson, D. A., 985. Dynamcs: Theory and Applcaton. McGraw Hll, New York, NY. -37