Avalable onlne at www.scencedrect.com Proceda Engneerng 16 (2011 ) 363 368 Internatonal Workshop on Automoble, Power and Energy Engneerng (APEE 2011) Modellng and Co-smulaton Based on AMESm and Smulnk for Lght Passenger Car wth Dual State CVT F. Xe, J. Wang*, Y. Wanga State Key Laboratory of Automotve Smulaton and Control, Jln Unversty, No.5988, Renmn Street,Changchun, 130025, Chna Abstract Co-smulaton platform was constructed usng MATLAB/Smulnk and AMESm for lght passenger car wth dual state CVT. The vehcle dynamc model, hydraulc system and controller model were establshed based on the cosmulaton platform. Through smulaton analyss of typcal workng condtons, t was valdated that the cosmulaton platform was effectve and practcal for R&D of passenger car wth CVT. 2010 Publshed by Elsever Ltd. Selecton and/or peer-revew under responsblty of Socety for Automoble, Power and Energy Engneerng Open access under CC BY-NC-ND lcense. Keywords: Modellng; Co-smulaton; Lght Passenger Car; Dual State CVT 1. Introducton Dual state CVT s a combnaton of optmal matchng of CVT and torque converter. Vehcle wth dual state CVT has good performance to start up, low speed crawlng at the stage of low speed movng, and good drvng and economy performance at the stage of hgh speed movng. Furthermore, engne effcency can be mproved, and vehcle s power and economy performance can be ncreased to great degree. 2. Co-smulaton Platform Structure Modelng and smulaton usng MATLAB/Smulnk tool for dynamc analyss has a very hgh level of accuracy, but also has the conflct between model smplfcaton and precson for complex nonlnear system. LMS Imagne.Lab AMESm software provdes graphcal modelng approach, whch elmnates the tedous mathematcal modelng, code programmng and has ensures hgh modelng effcency, but has the weakness of complex control system modelng. In ths paper, co-smulaton platform based on * J. Wang. Tel.: +86 13504434523 E-mal address: wjh@jlu.edu.cn, wjhsage@163.com 1877-7058 2011 Publshed by Elsever Ltd. do:10.1016/j.proeng.2011.08.1096 Open access under CC BY-NC-ND lcense.
364 F. Xe et al. / Proceda Engneerng 16 ( 2011 ) 363 368 Smulnk and AMESm was establshed combnng the advantages of the both. The block dagram s shown n Fg.1 (a). 3. Dynamc Model of Lght Passenger Car wth Dual State CVT 3.1. Engne Model The dynamc characterstcs of the engne are smplfed as a frst order lag nerta consderng engne output torque s n non-steady-state condtons n most cases [1]. The non-steady-state torque value can be obtaned through correctng the steady-state values by experence factor. Fg.1 (b) shows the numercal model of steady-state. Fg.1 (a) Co-smulaton Platform Based on AMESm and Smulnk (b) Steady-state numercal model of Engne output torque Wth a frst order lag nerta, the dynamc characterstcs of the engne are expressed as: s T = e T, / k s+ 1 e s e ( ) ( ) Where, T e s engne torque of non-steady-state; s openng; e s engne speed; s lag tme; k s dynamc ft coeffcent. 3.2. Torque Converter Model T s output torque of steady-state; s the throttle (1) The nput and output characterstcs of torque converter are expressed as follows: K = / T SR = / TR = T / T t t (2) (3) (4)
F. Xe et al. / Proceda Engneerng 16 ( 2011 ) 363 368 365 Where, K s the capacty factor; TR s torque rato; SR s speed rato; turbne speed; 3.3. CVT Model T s mpeller torque; Tt s turbne torque. s mpeller speed; t s The torque balance equaton of nput and output shaft of CVT [2-5] s: ( Te Ie d e / dt) CVT Td = Id d d / dt (5) Where, Ie s engne nerta; Td s the equvalent nerta of CVT output shaft converted from vehcle drvng resstance; I s the equvalent rotaton nerta of CVT nput shaft converted from flywheel and prmary pulley; d e I s the equvalent nerta of CVT output shaft converted from secondary pulley and fnal drve; s the angular velocty of CVT output shaft; transmsson effcency. 3.4. Tre Model d CVT s CVT transmsson rato; s Consderng tre rotaton nerta, HBPacejka s selected as the tre model, whch can be expressed as follows: Fx = D sn( C arctan( Bs E( Bs arctan( Bs)))) (6) Where, Fx s the tre longtudnal force; s s slp rato; D s the peak factor; B s the stffness factor; E s the curve shape factor; C s the shape characterstc factor. 3.5. Hydraulc Control System Model 3.5.1. Hydraulc System By smplfyng the hydraulc crcut approprately, retanng the man control valves, such as the speed rato control valve, clampng force control valve [6], the hydraulc control unt was establshed whch was shown n Fg.2(a). 3.5.2. Controller Model Torque converter controller only consdered the lock/unlock state control. In order to take nto account the transmsson effcency and lock-up comfort requrements and avod excessve torque mpact, torque converter controller selected the couplng pont as the lock-up control pont. When the torque converter speed rato ncreased to the couplng pont, the controller would send lock-up command. When the vehcle speed was lower than the threshold value whch was preconfgured (such as 15km/h), the controller would send unlock command. Clampng force control can guarantee the safety of torque delvery, whch can mprove transmsson effcency and the key components lfe. The clampng force controller consders both the transfer of engne torque and the current rato smultaneously. The controller receves speed, torque and pressure sensor sgnals n real tme to calculate the value of current engne torque and transmsson rato at ths tme, then determnes the target clampng force control command to clampng force control valve. Rato controller accordng to the selected drvng mode and the drver s ntenton (the throttle openng and the brake pedal poston and other parameters) to adjust the actual rato to track the target rato changng [7-9]. Takng nto account the complex operatng condtons and good adaptaton to nonlnear
366 F. Xe et al. / Proceda Engneerng 16 ( 2011 ) 363 368 systems, a fuzzy rato controller was desgned. Input sgnals are the dfference between the target rato and the actual rato, and devaton of the rate of change, the control sgnal s the duty cycle sgnal to control rato control valve. Fg.2 (b) s the fuzzy controller output surface. The vehcle dynamc model based on co-smulaton platform s llustrated as Fg.3 (a). Fg.2 (a)amesm Model of CVT Hydraulc Control Unt (b) Fuzzy Controller output surface Fg.3 (a)vehcle Model based on co-smulaton platform (b) Smulaton analyss results under cycle operatng condton 4. Smulaton Analyss under Typcal Operatng Condtons To valdate the dynamc characterstcs and control effect of vehcle co-smulaton model, typcal cycle condtons were selected, whch ncluded passenger car s start up, acceleraton, deceleraton, ncreasng n drvng resstance, neutral gear poston and stop. The smulaton results were shown as Fg.3 (b). As shown n Fg.3 (b), when smulaton began, the car started up and ncreased the velocty. As the speed ncreased, the CVT transmsson rato decreased from maxmum to mnmum. After mantanng a constant speed, then reduced the throttle openng, whch made the car slow down, then the prmary pulley cylnder draned and pressure down. The actual rato tracked the target speed to ncrease. Then acceleraton agan, wth the speed ncreasng, pressure of prmary cylnder ncreased at the same tme. When the drvng resstance ncreased, the load torque of transmsson ncreased as well as the pressure of secondary pulley cylnder to ensure effectve torque transferrng. Subsequently, the transmsson lever
F. Xe et al. / Proceda Engneerng 16 ( 2011 ) 363 368 367 was set to neutral gear, the pressure of prmary and secondary pulley cylnders draned at the same tme, and the speed rato of CVT automatcally went back to the maxmum value to ensure start up agan. At last the gear lever was set n the park and the car stopped [10]. From the smulaton results, t was valdated that the vehcle dynamc model based on co-smulaton platform can well adapt to the typcal operatng condtons. The hydraulc control unt can accord wth the drvng condtons and drver's ntenton to control the hydraulc actuators to realze correspondng drvng behavor, such as start up, acceleraton, deceleraton, neutral gear poston and stop. 5. Concluson A modelng and co-smulaton method for lght passenger car wth dual state CVT was ntroduced. The co-smulaton platform was based on Smulnk and AMESm. The vehcle power-tran dynamc model and hydraulc system model were constructed by AMESm, and the controller models were bult by MATLAB/Smulnk. Through smulaton analyss results, t was valdated that the co-smulaton platform was effectve and practcable for R&D of dual state CVT. Acknowledgements The authors would lke to thank Prof. Mngshu Lu, Prof. Youkun Zhang and Dr. Shupe Zhang at Automoble and Tractor Laboratory of Jln Unversty for great help durng the computng smulaton and bench test. References [1]Zou, Z., Zhang, Y., Zhang, X., and Tobler, W. Modelng and Smulaton of Trancton Drve Dynamcs and Control. ASME J. Mech. Des. 2002; 123(4):556-561 [2]Carbone, G., Mangalard, L., and Mantrota, G. Theoretcal Model of Metal V-Belt Drves Durng Rato Changng Speed. ASME J. Mech. Des. 2000;123:111-117 [3]Srvastava, N., and Haque, I. On the Transent Dynamcs of a Metal Pushng V-Belt CVT at Hgh Speeds. Int. J. Veh. Des. 2005;13(2):175-194 [4] Carbone, G., Mangalard, L., and Mantrota, G. Influence of Clearance Between Plates n Metal Pushng V-Belt Dynamcs. ASME J. Mech. Des. 2002;124:543-557 [5] Carbone, G., Mangalard, L., and Mantrota, G. Influence of Pulley deformatons on the Shftng Mechansm of Metal Belt CVT. ASME J. Mech. Des. 2005; 127:103-113 [6]Heon-Sul Jeong, Hyoung-Eu Km. Expermental Based Analyss of the Pressure Control Characterstcs of an Ol Hydraulc Three-Way On/Off Solenod Valve Controlled by PWM Sgnal. Journal of Dynamc Systems, Measuresment, and Control. 2002; Vol. 124, p.196-205 [7]Sun, D. C. Performance Analyss of a Varable Speed-Rato Metal V-Belt Drve. ASME J. Mech., Transm., Autom. Des. 1988; 110:472-481 [8]Lebrecht, W., Pfeffer, F., and Ulbrch H. Analyss of Self-Induced Vbratons n a Pushng V-Belt CVT. 2004 Interantonal Contnuously Varable and Hybrd Transmsson Congress. San Francsco, CA. No. 04CVT-32 [9]Akehurst, S., Parker, D. A., and Schaaf, S. CVT Rollng Tracton Drves A Revew of Research Into Ther Desgn, Functonalty, and Modelng. ASME J. Mech. Des. 2006; 128(5):1165-1176 [10]Brokowsk, M., Km, S., Colgate, J. E., Gllespe, R., and Peshkn, M. Toward Improved CVTs: Theoretcal and Expermental Results. ASME Internatonal Mechancal Engngeerng Congress and Exposton. New Orleans, LA. 2002 Appendx A.
368 F. Xe et al. / Proceda Engneerng 16 ( 2011 ) 363 368 A.1. Basc Parameters of the Vehcle Model Parameter Symbol Unt Value Vehcle Mass M Kg 1560 Engne Max. Power P emax Kw 56 Engne Max. Torque T emax Nm 115 Frontal Area A m 2 1.68 Wnd Resstance Coeffcent C D 0.32 Rollng Resstance Coeffcent f 0.012 Wheel Rollng Radus r w m 0.27 CVT Rato Range CVT 0.442~2.45 Center Dstance Between Prmary and A p m 0.15 Secondary Pulley Fnal Drve Rato 0 5.249 Mn. Radus of Prmary Pulley r p m 0.03057 Mn. radus of Secondary Pulley r s m 0.03231