Mechanism and Machine Theory 42 (2007) 168182

Mechanism and Machine Theory

Shift dynamics and control of dual-clutch transmissions

Department of Mechanical Engineering, University of Michigan-Dearborn, Dearborn MI 48128, United States

Abstract Shifts in a dual-clutch transmission (DCT) are realized by torque transfer from one clutch to another without traction interruption due to the controlled slippage of the clutches. The timing of engagement and disengagement of the two clutches is critical for achieving a smooth shift without engine are and clutch tie-up. This paper presents an analytical model for the simulation, analysis and control of shift dynamics for DCT vehicles. A dynamic model and the control logic for the integrated vehicle have been developed using Matlab/Simulink as the simulation platform. The model has been used to study the variation in output torque in response to dierent clutch pressure proles during shifts. Optimized clutch pressure proles have been created for the best possible shift quality based on model simulation. As a numerical example, the model is used for a DCT vehicle to simulate the wide-open throttle performance. Vehicle launch and shift process are both simulated to assess transmission shift quality and validate the eectiveness of the shift control. 2006 Elsevier Ltd. All rights reserved.
Keywords: Dual-clutch transmission; Automatic transmissions

1. Introduction There has been a clear trend in the automotive industry in recent years towards increased ride comfort and fuel eciency. As the power transmission unit, transmissions play an important role in vehicle performance and fuel economy. There are currently several types of transmissions and the associated technologies that oer dierent performance priorities when t into a vehicle [1]. Manual transmissions have an overall eciency of 96.2%, which is the highest eciency value for any type of transmission. Current production automatics have been improved to provide an eciency of not more than 86.3%. Belt type CVTs have an overall eciency of 84.6%, however, the major advantage of CVT is that it allows the engine to operate in the most fuel-ecient manner [2]. Automated manual transmissions have the same eciency of manual transmissions and oer operation convenience similar to conventional automatic transmissions. There exist two technically feasible designs for automated lay-shaft gearing transmissions. One uses a single clutch and is basically a manual transmission with an added-on control unit that automates the clutch and shift operations. In this design, there is an interruption of
*

Corresponding author. Tel.: +1 313 593 5539. E-mail address: anding@[Link] (Y. Zhang).

0094-114X/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/[Link].2006.03.002

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

169

torque during a gear change since the engine is cut-o by the clutch during shift. This torque interruption leads to unanticipated passenger felt jerks due to vehicle acceleration discontinuity and is highly uncharacteristic of conventional automatic transmissions. The other design uses a dual-clutch system between the engine and the transmission and overcomes the shortcomings of the single clutch version [3]. The two clutches are engaged alternatively in dierent speeds and power transmission continues during a shift through the control of clutch slippage. A shift process involves the engagement of the oncoming clutch and the release of the ogoing clutch. This results in shift characteristics that are typical of clutch-to-clutch shifts commonly seen in conventional automatic transmissions. It is a common practice in the automotive industry to use analytical models for the prediction and assessment of new types of powertrain systems. A great deal of research eorts have been focused on the modeling and control of vehicle transmission, such as conventional automatic transmissions [46], continuously variable transmissions [7,8], and hybrid systems [912]. Various formulation methods and programming techniques have been used in these researches to model the dynamics of vehicle powertrain and simulate the performance of transmission control. Typically, the equations of motion are rst derived separately at the component level and then integrated into the overall vehicle system. The integrated system models are either implemented in generically developed codes or in object oriented programming environment. As compared with the maturity of technologies for conventional automatic transmissions, the modeling and control of dual-clutch transmissions is still a new area and technologies associated with the DCT design and control are still at the early stage of development. This paper presents an analytical model for the simulation, analysis and control of the launch and shift processes of DCT vehicles. The research work is concentrated on modeling the vehicle dynamics during shifting and establishing a simulation tool for the analysis and optimization of shift control using clutch pressure proles as the control signals. Matlab/Simulink is used as a simulation platform to develop the dynamic model and control logic for the integrated vehicle powertrain model. The simulation model is used to study the variation in output torque in response to dierent clutch pressure proles. Optimized clutch pressure proles are obtained to achieve the best possible shift quality based on model simulation. As a numerical example, the model is used for a vehicle equipped with a dual-clutch transmission to simulate the wide-open throttle performance. Vehicle launch and shift process are also simulated to assess transmission shift quality and validate the eectiveness of the transmission control strategy. 2. DCT structure and model description The dual-clutch transmission is shown schematically in Fig. 1. The transmission has six forward speeds and a reverse speed. The transmission input shafts are designed as quill-shafts, with one solid shaft positioned inside another hollow shaft. The solid shaft carries on it the second, fourth, sixth and the reverse gear, while the hollow shaft carries the rst, third, fth gear. Clutch 1 (CL1) connects all the odd gears and Clutch 2 (CL2) connects all the even gears to the input. Synchronizers are positioned between two gears similar to conventional manual transmissions. When in a particular gear, the respective clutch and synchronizer are engaged and power ows from the engine through the clutch and synchronizer to the output shaft. The other clutch remains open and the remaining gears freewheel. To change a gear the ogoing clutch is slowly released and the oncoming clutch simultaneously engaged. It is this feature of DCT that allows for uninterrupted torque transfer even during a gearshift. The powertrain is modeled as an integrated multi-degree-of-freedom system in which each element is a lumped mass model, as shown in Fig. 2. The engine is modeled as a mass inertia and accepts throttle angle as the input to produce a mean torque at the crankshaft. The drivetrain consists of component models of the dual clutches, layshaft transmission gearsets, driveshaft, dierential and a vehicle model that incorporates the road load and aerodynamic drag. All gears and synchronizers are modeled as non-compliant elements and are represented as mass inertias. The input and output shafts are modeled as compliances and are represented by torsion spring-damper assemblies. Clutches and synchronizers are modeled as friction elements with hydraulic pressure as control signals. The output of the simulation model is the torque at the axle and vehicle speed. The relative angular velocity across the two clutches is monitored for decision making in the control logic. The following assumptions are made in the development of the model:

170

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

6
SYN6 Intermediate Shaft 2

5
SYN5R

R
Final Drive Pinion 2

CL2

CL1

SYN24

SYN13

Input Shaft

Intermediate Shaft 1

Final Drive Pinion 1

1 Output

Fig. 1. DCT stick diagram.

CL2

R
I1

i / p
Ii/p

K1

e
Ie Engine Cm Im

h
Is Ih

Km C1 Input Shaft

CL1
im
I2

31 1

o
ia I3

K2

im

C2 Output Shaft

Fig. 2. DCT dynamic model.

The engine output torque is interpolated in terms of the throttle angle and RPM from the engine map. Gears have no backlash. All the mechanical losses are modeled as a part of the vehicle drag. Delays due to hydraulic actuation system are not considered. Clutches are modeled as Coulomb friction elements. Temperature eects of the drivetrain are neglected.

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

171

3. Component models The DCT system consists of complicated component level subsystems such as the engine, clutches, and vehicle road load. These component models are described in the following. 3.1. Engine The mechanics of the engine assembly is modeled with a two degree-of-freedom system: one is the rotational inertia of the moving parts and the other is the inertia of the engine and transmission supported on their mounts. As an assumption, the engine is modeled as a mean value torque generator that does not include the engine transients. Engine output torque is interpolated corresponding to engine speed and throttle position from an engine map modeled as a look-up table. For each throttle opening (TA), the engine torque from idle to redline is a function of engine angular velocity (xe) such that, T e f T A ; xe where Te is the engine torque, xe is the engine angular velocity and TA is throttle angle. 3.2. Clutch The clutches in a DCT function as primary gear changing elements in addition to their normal purpose of disconnecting the engine from the driving unit. Based on the clutch geometry and friction characteristics, the clutch torque is calculated as follows:   3  2 Ro R3 i T CL lF n n 2 3 R2 R2 o i where l is coecient of friction that can be formulated as a function of clutch slip, Fn is normal force on clutch face that depends on the apply pressure, Ro is outside radius of friction disc, Ri is inside radius of friction disc, n is the number of friction discs. For simplicity, the clutch torque is modeled as a function of the relative angular velocity and the hydraulic pressure in the clutch piston in terms of a look-up table. The clutch has three operation states and the torque transmitted in each state is described by the following equation: 8 > ClDxP app slipping < 3 T CL T closed > : 0 open where C is the constant reecting the clutch dimension, Dx is the relative angular velocity between the clutch input and output ends and Papp is the pressure in the clutch piston. T is the torque applied on the clutch when the vehicle is running in particular speed with the clutch fully closed. As shown in the equation above, the clutch torque during a shift depends only on the clutch apply pressure and the coecient of friction for a given clutch. The apply pressure is the control signal and the coecient of friction depends on the clutch slip, i.e., the relative velocity of the two clutch sides. 3.3. Synchronizer In the dual clutch design, gears of the next speed are pre-engaged while the vehicle is running in a current speed. Gear engagement and synchronization do not occur simultaneously with the shifting process. The two gears to be synchronized are freewheeling on the shafts and hence, the synchronizer is modeled as a simple friction element. The synchronizer torque is proportional to the friction force as shown below: T syn f lF ; Dx13 4 1

where Tsyn is the synchronizing torque, F is the normal force acting on the friction cone, l is the coecient of friction of the synchronizer ring and Dx13 is the angular velocity dierence across a synchronizer, between the rst and third gears as indicated by the subscripts.

172

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

3.4. Vehicle road load The vehicle road load model accounts for the rolling resistance, aerodynamic resistance and transmission losses as formulated in the following: RA C d V 2 lf W W sin h T RL RA Rd 5 6

where Cd is the air drag coecient that depends on the body style and dimension, W is the vehicle weight, lf is the rolling coecient, h is the grade angle and V is the vehicle velocity. TRL is the road load torque on the wheel and Rd is the wheel radius. 4. System model The dynamic model for a dual-clutch transmission is illustrated in Fig. 2. The engine mounts, input and output shafts are modeled as spring-dampers to account for the compliance of these components. Gear shafts are modeled as coupled lumped masses. There are two sets of dynamic equations for the model. The model follows one set of equations of motion when the vehicle is running in a particular gear and the transmission provides a mechanical link between the engine and the wheels. During a gearshift it follows another set of equations as there is no direct link between the engine and the wheels as the system is in a dynamic shift state. Presented below are the generic dynamic equations for model operation in any given gear and also during an upshift. The equations for a specic gear and a specic gear change can be written by choosing the correct gears and considering Clutch 1 open in second, fourth and sixth gears and Clutch 2 open in rst, third and fth gear modes. 4.1. Operation in a particular gear 4.1.1. Engine and input shaft Im dxm To Tm dt T m K m hm C m xm dxe Ie T e T i=p dt T e T e T A ; xe 7 8 9 10 11

T i=p K 1 he hi=p C 1 xe xi=p

where Tm is the torque at the engine mount, Ti/p is the input shaft torque, Km and K1 are the stiness of the engine mounts and input shaft respectively, Ie and Im are the mass moments of inertia of the engine output shaft and the engine-transmission block, C1 and Cm are the damping coecient of the input shaft and the engine mount, hm, he and hi/p are the angular displacements of the engine mount, engine crankshaft and input shaft respectively, xm, xe and xi/p are the angular velocities of the engine mounts, engine crankshaft and input shaft, respectively. To is the reaction torque applied on the mount of the engine-transmission assembly, whose magnitude is the same as the output torque. 4.1.2. Transmission and output shaft I i=p dxi=p T i=p T CL1 T CL2 dt & T im engaged T CL1 ; T CL2 0 open dxim To T im it I eq dt ia 12 13 14

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

173

T o K 2 ho hw C 2 xo xw T o T RL W dv g dt Rd dv dxw Rd dt dt

15 16 17

where K2 and C2 are the stiness and damping coecients of the output shaft, respectively. ho and hw are the angular displacements of the output shaft and wheel, respectively. Ii/p is the mass moment of inertia of the input shaft. T CL1 and T CL2 are the torques carried by the dual clutches. Tim is the torque on intermediate shaft 1 for the rst, second, third and fourth speeds and on intermediate shaft 2 for the fth, sixth, and reverse speeds as shown in Fig. 1. xim is the angular velocity of the intermediate shaft. xo and xw are the angular velocities of the output shaft and the wheel, respectively. To is the axle output torque. Ieq is the equivalent mass moment of inertia on the intermediate shaft that includes all rotational masses in the power ow path. it is the existing transmission ratio. For example, Ieq for the third speed and the sixth speed is calculated respectively as follows: I eq I 1 I h i2 3 I eq I 2 I s i2 6 for the third speed for the sixth speed 18 19

where i3 and i6 are the gear ratios of the third and sixth speed, respectively. I1 and I2 are the mass moments of inertia of the two intermediate shafts assemblies. Ih and Is are the mass moments of inertia of the hollow and solid shaft assemblies. 4.1.3. Equations of motion during a shift During a shift, the clutch is no longer closed. For an upshift, Clutch 1 is being released and at the same time Clutch 2 is being applied such that both of them are slipping during the shift period. Hence, the engine torque is not directly transmitted to the intermediate shaft by a closed clutch. It is the friction torque in the clutch that is now been transmitted. Since there is no mechanical link between the engine and wheel, the powertrain dynamics are governed by another set of the equations represented as follows. 4.1.4. Engine dxe T e T i=p dt T i=p K 1 he hi=p C 1 xe xi=p Z T e T i=p xe dt x0 e Ie T e f xe Ie 20 21 22 23

where x0 is the engine angular velocity at the beginning of the shift. The engine is a function of its angular e velocity alone since the throttle angle is controlled at a xed value during shift. 4.1.5. Transmission dxi=p T i=p T CL1 T CL2 dt T CL1 f1 P app1 ; DxCL1 I i=p T CL2 f2 P app2 ; DxCL2 DxCL1 xe xh DxCL2 xe xs xe ishift xw ia 24 25 26 27 28 29

174

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

I eq

dxim To T CL1 iodd T CL2 ieven dt ia T o K 2 ho hw C 2 xo xw T o T RL W dv g dt Rd

30 31 32

where xh and xs are the angular velocities of the hollow and solid shafts, respectively. iodd and ieven are the gear ratios of the current and next speeds involved in the shift. ishift is the gear ratio during a shift and is a function of shift time. Papp1 and Papp2 are the clutch pressure signals that serve as the shift control input. The equivalent mass moment of inertia Ieq in Eq. (30) depends on the related shafts and gear ratios involved in the speed change. For instance, Ieq is calculated for a 12 shift as follows: I eq I 2 I h i2 I s i2 1 2 5. Shift control logic The dynamic equations derived above are incorporated in the simulation model in Simulink. Each equation is treated as a matrix of elements and has one quantity that is derived from known values. Sometimes external inputs are used to calculate the unknown and in some cases an algebraic loop is formed, where the output of a future event acts as an input to its driver. Gear-shifting strategy can be dened as the logical combination of a series of interdependent events, which dictate the accurate instance of a gear upshift or downshift as well as the optimum duration of a gear change. Fig. 3 shows the decision block used to build the control logic for the DCT modeling. A shift controller receives input from various sensors like the throttle position, current vehicle speed, current gear position, clutch schedule and engine speed. Based on these inputs the controller makes a
GEAR POSITION

33

DECISION BLOCK FOR CLUTCH ACTIVATION

NEUTRAL?

IN GEAR?

YES CHANGE GEAR?

YES CL1 = 0 CL2 = 0

CL1 ONCOMING (RISING CURVE) ODD TO EVEN? CL2 OFFGOING (FALLING CURVE)

YES

NO

1, 3, 5?

CL1 ACTIVE

2, 4, 6?
EVEN TO ODD? CL2 ACTIVE CL1 OFFGOING (FALLING CURVE) CL2 ONCOMING (RISING CURVE)

Fig. 3. Decision logic block.

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

175

5-4

6-5

Throttle Position (%)

60 1-2 50 40 30 20 10 0 0 10 20 30 40 50 60 70 Vehicle Speed (mph)

80

90

100

110

Fig. 4. Gear shift schedule.

decision of an upshift, downshift or gear position holding. A gear shift schedule used in this paper is illustrated in Fig. 4. At every simulation time step, the threshold shift speeds on the six shift curves, 12, 21, 23, 32, up to 56 and 65 are calculated according to the current throttle angle. To implement shift logic for the dual clutch operation, the following three criteria need to be addressed: 1. To detect the time the shift is initiated. 2. To create a specic rate of engagement and disengagement of the clutches. 3. To determine the completion of shift. The continuous dynamics of the gearshift event is modeled in Stateow/Simulink. A pulse generator is placed in the Stateow subsystem such that it generates a spike every time the gear state changes to a new state. To determine the end of shift, the dierence in angular velocity across the oncoming clutch is monitored. Since this clutch is open before the start of shift, there is a velocity dierential across the clutch. When the clutch locks-up, both sides rotate at the same speed such that the relative angular velocity across the clutch is zero. In this paper each clutch is modeled as a look-up table consisting of a rising pressure curve and a falling pressure curve for the clutch to function as oncoming clutch and ogoing clutch, respectively. A counter is designed such that it generates an input ramp curve for the clutch pressure look-up tables once the gear change pulse is detected. The counter always compares the angular velocity across the oncoming clutch and creates a lock-up signal until the relative velocity becomes zero. The rate of engagement of the clutch thus becomes a function of how quickly the angular velocity drops to zero. For a quick shift, engagement curve will be steeper than that for a slow shift. A downshift is carried out in exactly the same manner, except the clutch pressure curves are reversed and so are the oncoming and ogoing clutches. The ratio change during a shift can be qualitatively explained using a 12 upshift as an example. In the rst gear operation Clutch 1 is in the completely engaged condition and Clutch 2 is disengaged. The second gear is already engaged by the 24 synchronizer. In the rst gear the output torque of the transmission is a function of the engine torque, engine speed, gear ratio and the driveline inertias. If the damping eects of the input shaft are not considered for simplicity, the transmission output torque in the rst gear is To T e i1 I 3 I 2 i2 I 1 i2 xe 2 1 _ ia 34

176

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

The output angular velocity is related to the engine angular velocity: xe xo i1 ia

35

For a xed throttle position, the engine torque in a particular gear is calculated based on the engine speed and a look-up table which has the engine map incorporated in it. This is possible since the clutch is connected and there exists a mechanical link between the engine and the wheels. However, during a gear shift, both the clutches are slipping and are partially engaged. The engine speed becomes a function of the clutch engagement schedule and hence the engine torque also varies as the speed. The engine speed is calculated as follows: R T e T CL1 T CL2 dt xe 36 Ie Two important observations can be made from this equation: (1) Engine angular velocity is not constant and hence the gear ratio follows a waterfall like curve based on the engine speed. xe ii2 a 37 xo i (2) The torque developed by the o-going and on-coming clutch is deducted from the engine torque. Hence, the torque reduction during a gear change depends on the timing and engagement characteristics of the two clutches. The most ecient condition would be when the two clutches can smoothly take up the torque dierential between the initial gear and the target gear output torque.

6. Simulation results and analysis The developed vehicle model is used for the shift control of a midsize passenger car equipped with a dualclutch transmission. The vehicle data involved in the model are shown in Table 1.
Table 1 Vehicle data Parameter Engine size Vehicle mass Tire radius Eective tire rolling radius Dierential gears moment of inertia Final drive gear ratio Final drive gears moment of inertia Transmission gear ratios Odd gears moment of inertia Even gears moment of inertia Engine mount damping constant Engine mount spring constant Engine moment of inertia Input shaft moment of inertia Input shaft spring constant Input shaft damping constant Output shaft spring constant Output shaft spring constant Solid shaft moment of inertia Hollow shaft moment of inertia Intermediate shaft moment of inertia Value 3.0L V6 3000 kg 0.312 m 0.308 m 0.0047 kg m2 3.07 0.0002 kg m2 First 3.14, second 1.98, third 1.37, fourth 1.00, fth 0.76, sixth 0.6, Rev 4.04 0.0023 kg m2 0.0009 kg m2 135 N m s/rad 16269 N m/rad 2.7 kg m2 0.004 kg m2 4920 N m/rad 13.6 N m s/rad 16300 N m/rad 31.1 N m s/rad 0.002 kg m2 0.001 kg m2 0.008 kg m2

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

177

Reduce engine torque

Oscillations

Offgoing clutch disengaging Output Torque

Peak overshoot

Offgoing clutch disengaed

Oncoming clutch engaging

Torque Re-instatement

Torque Phase

Inertia Phase Time

Fig. 5. Output torque prole for typical DCT shift.

The output torque prole of the dual-clutch transmission for a typical upshift is shown in Fig. 5. As seen in the nature of the output torque, clutch-to-clutch shifts consist of two phases, a torque phase and an inertia phase. When a gear shift signal is initiated, the engine torque is transferred through two clutches simultaneously. During this time neither of the two clutches is completely engaged. The transmission output torque drops to the lowest value during the torque phase due to shift transients. The speed ratio is not changed yet during the torque phase. In the inertia phase that follows the torque phase, the engine has to be decelerated in order to achieve speed synchronization with the target gear. During this phase, the kinetic energy of the powertrain components is transferred to the vehicle while engaging the next gear, causing a brief rise in the output torque. The inertia phase is responsible for most of the torque recovery and this is where the torque overshoots occur. The two key aspects considered in the clutch pressure control are: (a) As the oncoming clutch reaches its critical capacity, the ogoing clutch pressure is brought to zero. This ensures that the clutches will not tie-up and prevent causing a sharp torque overshoot, which is felt as a vehicle jerk by the driver. (b) Before the oncoming clutch reached its critical capacity, the ogoing clutch is held above its critical capacity to maintain continuity in torque transmission and prevent the engine from aring that is felt as a power loss by the driver.

6.1. Launch To launch a vehicle from neutral, Clutch 1 is controlled to slip with the rst gear engaged. The launch control is through a look-up table for the clutch pressure and the normalized pressure in the clutch cylinder is modulated. Various output torque proles at launch, shown in Fig. 6, can be obtained based on dierent launch pressure proles of the clutch cylinder as shown in Fig. 7. For an aggressive launch, the slip torque

178

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

Fig. 6. Output torque during launch.

Fig. 7. Clutch pressure proles for launch.

prole has a steep slope and clutch locking torque is achieved in a very short time. Likewise, for a slow start the same torque prole has a gentle slope and the time to lock the clutch is greatly increased. An optimized

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

179

launch is achieved between the above two extremes. Although steady state torque is seen to be achieved in the same launch time, the clutch is hooked up quite early in case of option 2. This results in a pronounced vehicle jerk, however such a situation might be desirable when the vehicle is being driven in Sport mode. 6.2. 12 Upshift The simulation was run for an upshift from rst gear to second gear at wide open throttle. The vehicle speed is approximately 20 mph when the shift is initiated. The output torque during shift is plotted in Fig. 8. At the beginning of the shift, the pressure on the ogoing clutch is sharply reduced to the threshold for disengagement. This causes the output torque to drop, which depicts the torque phase. The oncoming clutch pressure is ramped up to the desired pressure level and the output torque begins to regain its value in the inertia phase. Fig. 9 shows the relative angular velocity across the oncoming and the ogoing clutches. The output torque variations mainly depend on the clutch piston ll pressures. The simulation model was tested for three dierent clutch pressures and the eect on output torque proles was studied. Fig. 10 shows the three clutch pressure proles for the shift control. Each set of twin clutch pressure curves has a dierent point for ogoing clutch disengagement and also a dierent rate of engagement for the oncoming clutch. The pressure prole in red is optimized in terms of the shift smoothness. For a higher than optimum pressure applied at the clutch, the torque overshoot is clearly observed. On the other hand, a large torque dip called torque hole is observed when the clutch ll pressures are lower than optimum. The three torque proles are plotted for comparison in Fig. 11. 6.3. 32 Downshift A 32 downshift is simulated and the results are shown in Fig. 12. A downshift is initiated either when the throttle is suddenly depressed for a passing maneuver or when a brake is applied. In this model, the signal

Fig. 8. Output torque in a 12 upshift.

180

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

Fig. 9. Relative angular velocities of the dual clutches.

Fig. 10. Clutch pressure proles for 12 shift.

builder incorporates a brake model. In the Stateow gear selector, an additional condition is included to check if brake signal is prevailing. When the brake signal is detected the transmission is ready for a downshift.

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

181

Fig. 11. Eects of pressure proles on output torque.

Fig. 12. Output torque during a 32 downshift.

7. Conclusion This paper is aimed at studying the dynamic characteristics of the shift events of dual clutch automated layshaft transmissions. A simulation model integrating the engine, the transmission and the vehicle environ-

182

M. Kulkarni et al. / Mechanism and Machine Theory 42 (2007) 168182

ment has been built using Matlab/Simulink software tool. The main objective was to create an integrated engine-driveline system model and simulate the behavior of the DCT in events like launch, upshift or downshift. The importance of clutch apply timing and clutch pressure control signals are quantitatively analyzed and explained using model simulation in this paper. To understand the eects of clutch apply timing, three dierent clutch pressure proles are modeled and the resulting torque outputs are comparatively studied as a case study. By modifying the clutch pressure proles interactively with the simulation results, optimized clutch control can be achieved for smooth vehicle launch and gear shifts. It can be concluded that an eective and detailed analysis of the DCT shift characteristics can be performed using the model presented in the paper. References
[1] G. Lechner, H. Naunheimer, Automotive Transmissions: Fundamentals, Selection, Design and Application, Springer, Berlin, New York, 1999. [2] M.A. Kluger, Denis M. Long, An Overview of Current Automatic, Manual and Continuously Variable Transmission Eciencies and Their Projected Future Improvements, SAE paper 1999-01-1259, 1999. [3] W. Grobpietsch, T. Sudau, Dual Clutch for Power-Shift Transmissions A Traditional Engaging Element with New Future, VDIBerichte Nr. 1565, 2000, pp. 259273. [4] D.H. Kim, K.J. Yang, K.S. Hong, J.O. Hahn, K.I. Lee, Smooth shift control of automatic transmissions using a Robust adaptive scheme with intelligent supervision, International Journal of Vehicle Design 32 (3/4) (2003) 250272. [5] Y. Zhang, Z. Zou, X. Chen, X. Zhang, W. Tobler, Simulation and analysis of transmission shift dynamics, International Journal of Vehicle Design 32 (3/4) (2003) 273289. [6] T.W. Megli, M. Haghgooie, D.S. Colbin, Shift Characteristics of A 4-Speed Automatic Transmission, SAE paper 1999-01-1060, 1999. [7] Z. Zou, Y. Zhang, X. Zhang, W. Tobler, Modeling and simulation of traction drive dynamics and control, Journal of Mechanical Design, ASME Transaction 123 (4) (2001) 556561. [8] M. Yasuoka, M. Uchida, S. Katakuta, T. Yoshino, An Integrated control Algorithm for an SI Engine and CVT, SAE paper 1999-010752, 1999. [9] M. Goetz, M.C. Levesley, D.A. Corolla, Integrated Powertrain Control of Gearshifts on Twin Clutch Transmissions, SAE paper 2004-01-1637, 2004. [10] L.W. Tsai, G. Schultz, N. Higuchi, A novel parallel hybrid transmission, Journal of Mechanical Design, ASME Transaction 123 (2) (2001) 161168. [11] K.L. Butler, M. Ehsani, P. Kamath, A Matlab-based modeling and simulation package for electric and hybrid electric vehicle design, IEEE Transactions on Vehicular Technology 48 (6) (1999) 17701778. [12] B.K. Powell, K.E. Bailey, S.R. Cikanek, Dynamic modeling and control of hybrid electric vehicle powertrain systems, IEEE Control Systems Magazines 18 (5) (1998) 1733.