07-05-2014, 12:33 PM
Investigation of synchroniser engagement in dual clutch transmission equipped powertrains
Investigation of synchroniser.pdf (Size: 1.05 MB / Downloads: 43)
ABSTRACT
Transient response of a dual clutch transmission (DCT) powertrain to synchroniser
mechanism engagements is investigated using a lumped inertia model of the power-
train. Original research integrates lumped inertia powertrain models for the DCT with a
detailed synchroniser mechanism model and two separate engine models, comprising
of a mean torque model and a harmonic torque model, using torque derived from piston
firing. Simulations are used to investigate the synchroniser mechanism engagement
process in a previously unscrutinised operating environment. Simulations are per-
formed using both engine torque models, with the mean torque model demonstrates
the highly nonlinear nature of synchroniser mechanism engagement, and the power-
train response to the engagement process. Through the introduction of harmonic engine
torques, additional excitation is present in the mechanism during engagement, and
increased vibration of the synchroniser sleeve results. The impact of vibrations is
particularly important to the increased wear of indexing chamfer contact surfaces.
Introduction and background
The integral design and operation of dual clutch transmissions exposes synchroniser mechanisms to new sources of
vibration during actuation. Traditional manual transmissions release the clutch before synchronisation begins, isolating
the synchroniser from forced engine vibration. Whereas, in DCTs, synchronisers are engaged with the engine still driving
the vehicle through the alternate clutch, thus torsional vibrations generated in the engine can now be transmitted to the
synchroniser during actuation, exposing it to increased wear and potentially causing failure of the mechanism through
premature release of the ring as a result of vibration. Current synchroniser mechanism studies [1–4] focus on the impact of
mechanism engagement on the driver or investigation of design and operational variables on the engagement process,
analysing ‘‘double bump’’ phenomena and variation in chamfer alignments on engagement for manual transmissions.
Reliance of synchronisers on torque balancing and its highly nonlinear engagement process make the application of the
mechanism in DCTs potentially problematic [5,6]. The unique design and control of DCTs actuates the synchroniser with
the engine still driving the powertrain, engagement is no longer masked by the torque hole as is performed in manual
transmissions. Indeed it is now exposed to aggressive loading during engagement, including the introduction of forced
vibration from the engine and vehicle acceleration.
Synchroniser mechanism dynamics and control
Synchroniser mechanism dynamics are governed by the torques generated in the cone clutch and chamfered splines,
and to fully reveal these relationships this chapter is divided into three sub-sections. Firstly, the synchroniser mechanism
is described and the process of engagement is defined. Secondly, the control apparatus is presented schematically and
mathematical model discussed. Finally, the torques acting on the synchroniser mechanism and application as the control
torque for modelling are defined, including procedure for modelling the sleeve motion during actuation.
Synchroniser mechanism and engagement process
The synchroniser comprises of three key elements that control the mechanism engagement, shown in Fig. 1. The sleeve,
connected to the lay shaft via a splined saddle to enable axial translation only, receives actuation force from the actuator
arm and hydraulic system to engage matching chamfers on both ring and hub. The ring comprises of external chamfered
splines to mate with the sleeve and an internal cone clutch to engage the hub. The hub is mounted on the target gear and
includes both the external cone clutch friction surface and indexing chamfers to engage the sleeve.
Synchronisation process begins with the sleeve seated in the neutral position, Fig. 2(a), load is applied and the detents
are breached. The sleeve pushes the ring forward and viscous torque in the cone clutch rotates the ring to the blocking
position. Any oil film in the cone is squeezed out as the ring moves forward until dry friction results in the cone. The ring is
then rotated to block sleeve motion with chamfered splines on ring and sleeve in contact, speed synchronisation begins,
Fig. 2(b). This condition is maintained while cone clutch torque exceeds torque in blocking chamfers, typically to the
completion of speed synchronisation. When blocking torque exceed the cone torque the ring is rotated by chamfer torque
to the neutral position and sleeve passes over the ring to engage hub chamfers, initiating indexing, Fig. 2©. The final stage
of the process is the re-alignment of sleeve and hub chamfers, where sleeve moves over the hub chamfer tips to interlock
the target gear with the saddle completing engagement, Fig. 2(d).
Powertrain lumped model formulation
The development of a model of powertrain equipped with a dual clutch transmission is presented in this section. It
follows the basic principles of torsional lumped inertia models that can be found in texts such as Rao [17]. To capture the
basic the responses of a vehicle to torsional transients a minimum of four degrees of freedom is required, capturing
the lowest vibration modes. For broader frequency ranges multibody formulation in this paper uses 15 dof with the
synchroniser open, and 14 dof when closed, based on the structure in Fig. 4 it provides sufficient detail of the main
characteristics of the powertrain.
The model includes elements for engine, flywheel, and clutch drum, the simplified transmission model includes
coupled wet clutches and simplified gearing with idling components lumped at gears or synchronisers, and allows output
via a final drive gear attached to the propeller shaft, transmitting torque to the differential, thence to the hub, and vehicle
inertias.
Damped free vibration analysis
To demonstrate the adequacy of this model the first results presented are for the natural frequencies and corresponding
damping ratios of the powertrain model before and after synchronisation. Applying the eigenvalue problem to the system
matrix complex damped natural frequency pairs are returned as eigenvalues; see Rao [17] for detailed description of
method. From these results the natural frequencies and damping ratios are obtained for both system states.
Referring to Table 2, the lowest natural frequency, 0 Hz, is the rigid body mode (RBM) of the system, with the 4th gear
freewheeling there are two rigid body modes present in the open synchroniser model, one for the powertrain and the other
for the freewheeling gears targeted for synchronisation. While with the synchroniser closed there is only one rigid body
mode for the entire system. Natural frequencies of 6, 34, and 98 Hz are consistent with shuffle, hub, and clutch drum
vibration models. The eigenvalue solutions produced 13 negative paired solutions for both open and closed synchroniser
powertrains, with the other solutions being for the rigid body modes as previously discussed, consisting of very small
values (10 À 13) signifying numerical error. These results demonstrate that this is a semi definite system, in that it is not
physically constrained, including the freewheeling synchroniser in the powertrain for the open synchroniser model, and,
more importantly, indicate that powertrain is stable for both open and closed states.
The direct result of locking the synchroniser is to eliminate one degree of freedom in the powertrain model, and alter
the local inertia in the transmission. Observed from table one is the modification of local natural frequencies 8–11 as a
direct result of synchroniser lockup and increase in local inertia. As the structure of system damping is not modified
significantly it can be concluded that on11 in the open model is reduced below that of on10 rather than both natural
frequencies being reduced in the closed synchroniser free vibration results.
Conclusion
A 15 dof lumped spring–mass powertrain model for the investigation of synchroniser mechanism engagement in a dual
clutch transmission has been developed in this paper for the analysis of synchroniser engagement. This includes the
integration of hydraulic control system, a piecewise nonlinear synchroniser model that incorporates torques from both
cone clutch and control chamfers, and engine models using mean torque and harmonic torques from piston firing. These
were applied to investigate the engagement dynamics of the synchroniser mechanism in a powertrain equipped with a
DCT. Speed synchronisation, ring unblocking, and indexing of the mechanism were all demonstrated to be effective using
the mean torque engine model. It was shown that the impulsive speed change between sleeve and ring upon completion of
unblocking generates sufficient torque to release the locked cone clutch, resulting in a relative speed between sleeve and
hub chamfers before indexing, resulting in increased vibration during indexing.
The introduction of engine torque harmonics to the system was performed to demonstrate the transmissibility of
vibrations to the synchroniser and its influence on engagement. The development of stick–slip in the cone prior to
unblocking was identified in the later stages of speed synchronisation, resulting from the higher torsional vibration in the
system. Increased vibration was also present during indexing, with more substantial response than demonstrated in the
mean torque simulations.