It is well-known that the hydrodynamic torque converter plays a major role in the transient study of power train systems since it has a great influence on the transient characteristics of a vehicle during gear shifting as well as vehicle launching. To predict accurately the vehicle characteristics, detailed analysis of the hydrodynamic torque converter is required. However, even with the development of a nonlinear dynamic model for the torque converter based on Hrovat and Tobler's paper (1985) is available, it is imperative to calculate both torques from impeller and turbine in order to utilize the dynamic model since it takes torque as an input [3]. In order to obtain the information about necessary but unmeasurable variables, nonlinear model-based estimator is developed using already available and measurable speeds data of impeller and turbine. The hydrodynamic torque converter model includes all necessary dynamics, namely, hydraulic as well as mechanical dynamics. This model was developed and implemented using Matlab/Simulink®, and was validated against the vehicle experimental data [10]. With the validated model, nonlinear model-based estimator is developed to calculate the impeller and turbine torques.
The principle contribution of this work is the design of nonlinear model-based estimator using the sliding mode observer for a hydrodynamic torque converter, which later can be used for studies of power train transient effects as well as transmission related control. Even though the dynamic torque converter has been available, there has not been any torque estimator available despite the benefit of using such a dynamic model based torque estimator. The benefit of the dynamic model-based estimator for torque can be seen from figures 7 and 8 with the help of comparison against the torque estimation using steady-state characteristics of torque converter. In addition, with the use of the developed dynamic model-based estimator, torque which is one of the essential variables in terms of control can be accurately estimated compared to the existing static look-up based torque estimator. The resulting nonlinear model-based estimator with the hydrodynamic torque converter will help study the transient study of power train systems by providing physically justified model for engine and vehicle behavior when driven by large signal variations in either engine net torque output or vehicle load. In addition to the power train study, this model will be very useful for shift control and control of converter clutch applications.