ClosedLoopControl

Background

As seen in the Open Loop Control example, a simple voltage profile fed to solenoids properly enables to change gears but a certain amount of shift shock is clearly observed. Reducing such shift shock is a significant challenge for OEMs and suppliers because shift quality is one of the most important factors for stress free driving experience and comfort. One solution to achieve smooth transients is to use a closed loop controller that can accurately follow an optimal pressure profile. In contrast to an open loop controller whose only function is input a prescribed profile to the system, a closed loop controller monitors the output from the system as feedback to the controller to modulate the input to the system. Closed loop control can also significantly reduce calibration effort as many different clutch pressure profiles will be needed for different vehicle conditions (different engine load conditions). The closed loop controller in this example consists of four PID controllers which can be switched between depending on clutch states. The results show good shift reduction that will give better shift quality.

Control Concept

The control concept is shown in Figure 1. The input to the controller is the pressure in the Clutch Chamber (measurement) and output to the system is a voltage signal to the Solenoid (actuator). The PID controller evaluates the error between the measured pressure and a pressure command which is defined in tables within the controller, and then produces the solenoid voltage to the system. As described earlier, the PID gains can be changed depending on the clutch state detected by the clutchState model. The clutch states are defined as four different states as shown in Figure 2.

1. Steady (Low Pressure) and Fill

2. Engagement

3. Steady (High Pressure)

4. Disengagement

In this example, Steady (Low Pressure) and Fill states are treated as the same state because they share the same PID parameters according to a reference paper. Therefore, the controller in this example has 4 sets of parameters for 4 clutch states. By changing these parameters, the effect on shift shock behavior can be analyzed.

Figure 1. System concept for closed loop control

Figure 2. Definition of clutch state

Analysis

Closed loop control shows better results than the result from open loop control as shown in Figure 3.

Figure 3. Vehicle inertia acceleration

Figure 4 shows how the controller behaves within this model. The top diagram shows a comparison between the command signal (red) and actual output (blue). The output pressure follows the command well during the initial lower steady, fill, engagement and higher steady states. The error tends to increase in disengagement and in following lower pressure phases because of a lack of flow rate. Nevertheless, the torque transfer is decreased during this period because of a lack of pressure to fill the clutch, as shown in the middle diagram. The bottom diagram illustrates the clutch state detection. The PID gains are changed according to this detection.

Figure 4. Controller and Plant behavior

References

Virinchi Mallela and Zongxuan Sun:

Modeling and Control of a Novel Architecture for Automatic Transmissions

ASME Proceedings | Control Design Methods for Advanced Powertrain Systems and Components