A Development of Shift Control Algorithm for Automated Manual Transmission in the Hybrid Drivetrain

Seoul 2000 FISITA World Automotive Congress F2000A036 June 12-15, 2000, Seoul, Korea

Sungtae Cho1) *, Soonil Jeon1) , Hansang Jo2), Yeongil Park3), Jangmoo Lee1)

1) Seoul National University, Seoul, Korea 2) Hyundai Motors 3) Seoul National University of Technology, Seoul, Korea

In this study, a modified shift control algorithm for improving the shift quality of a parallel hybrid drivetrain with an automated manual transmission (AMT) is proposed. Improving shift quality of general AMT demands the sophisticated control for the engagement of clutch, which is generally known to be difficult. But in the parallel hybrid vehicle, both engine and motor can control the speeds of each clutch plate. Therefore it provides the easier clutch control in shift than general AMT. Consequently, it also permits the much reduced shift shock during the shift period. Furthermore various dynamometer- based experiments are carried out to prove the validity of the proposed shift control algorithm.

Keywords: Parallel Type Hybrid Drivetrain, Automated Manual Transmission, Shift Control Algorithm

The test system for hybrid drivetrain is constructed by at- 1. INTRODUCTION taching motor, inverter, power-connection device, sensor, and actuator and so on to the conventional drivetrain for transit bus. For maximizing fuel economy and minimizing emissions, it is recommended to perform the integrated control of engine, motor and transmission by using an automated transmission. In present days, several automated transmis- sions are developed such as an Automatic Transmission (AT), a Continuously Variable Transmission (CVT), and Automated Manual Transmission (AMT) etc. However there are some limitations to apply an AT and a CVT for heavy-duty vehicles such as transit buses or trucks; AT has the disadvantage of a large amount of power loss as its capacity is increased, and the CVT can be currently applied only for small-sized passenger car. In the case of AMT however, it is easily realized by mounting pneumatic actuators and sensors on the clutch and shift levers of the conventional manual transmission, and it reveals a high 1: induction machine 2: diesel engine 3: transmission efficiency compared with AT, suitable for heavy-duty ve- 4: clutch 5: power connection device 6: inverter hicle. But in the case of heavy-duty vehicle, the shift qual- 7: main controller 8: throttle actuator 9: electric fuel injec- ity of a heavy-duty vehicle is less important than that of a tor 10: battery 11: differential gear 12: wheels passenger car; therefore AMT can be a proper solution for 13-15: x-axis, y-axis, clutch pneumatic actuator a heavy-duty vehicle. ω ω ω 16-18: saw-tooth wheel for mot , eng , veh In this study, the AMT for the hybrid drivetrain for transit Fig.1. Schematic diagram for parallel type hybrid bus is constructed, and an advanced shift control algorithm drivetrain. for improving the shift quality is presented. Adequate controls of motor and engine during shift enable the shift shock to be reduced. To verify the presented shift control algorithm, various dynamo-based experiments are carried out.

2. THE HYBRID DRIVETRAIN TEST SYSTEM INCLUDING AUTOMATED MANUAL TRANSMISSION

2.1 THE PARALLEL HYBRID DRIVETRAIN Fig.2. Photo of Hybrid Drivetrain Test Bed. TEST SYSTEM

1 2.2 CONTROLLER 3. SHIFT CONTROL ALGORITHM AND EXPERIMENTAL RESULT The controller, which is designed to manage the overall operation of hybrid drivetrain, is constructed using TMS320C31 DSP Chips. It contains shift logic and power- 3.1 SHIFT CONTROL ALGORITHM split algorithm, so that it controls the velocity and torque of engine/motor.1) (See Fig. 3, 4)

Fig.3. Schematic diagram of main control unit.

Fig. 6. Flowchart of the AMT control program

Fig. 6 shows the basic algorithm applied on the shifting Fig. 4. Photo of main controller and inverter. control of AMT in this study. The shift sequence can be summarized as follows. 2.2 ENGINE THROTTLE CONTROL SYSTEM 1. If the shift starts, the controller separates the clutch plates and simultaneously sets the refer- The integrated control of engine and motor is required to ence torque output of engine and motor. operate them in their high efficiency region. In this study, the vector control method that was applied in the previous 2) 2. After separating the clutch, the gear is shifted to study is used for the motor control, and the engine con- neutral position by the movement of Y-axis shift trol system controlling the throttle of engine was devel- cylinder, and then the speed control starts to syn- oped. In the case of DC motor system for engine control chronize velocities of transmission input/output that in the previous study, there are some problems : the shafts by dc motor. requirement of feedback control, the weakness to the dis- turbance. To overcome those problems, we developed and 3. After the speed control of transmission in- applied the engine throttle control system using step mo- put/output shafts, the gear is shifted to the next tor. Fig. 5 shows the schematic of the engine throttle con- speed by the movement of X-axis cylinder. trol system. 4. During the step of 1-3, the engagement of synchronizer is completed, and then the engine speed is controlled to engage the clutch plates. 5. For the speed of input clutch plate to follow the speed of output clutch plate, the speed control of engine is executed. The shift is completed by engaging clutch plates after the speed difference between input and output clutch plates is below the reference value.

Fig. 5. Schematic diagram of the engine torque control Fig. 7 shows shift sequence in diagram. system using step motor.

2 1. Start shift 3.2 EXPERIMENT OF SHIFT USING BASIC SHIFT ALGORITHM

There are 3 modes such as a hybrid, an engine and a motor mode in this hybrid drivetrain. Among them, the hybrid and engine modes show exactly same shift characteristics because they controls power-sources, clutch, and shift mechanism in the same way during the shift. In the case of motor mode, there is negligible shift shock because there is no clutch engagement/disengagement3). So we particularly 2. Separating clutch plates, set the output torque of en- studied the shift characteristics in hybrid mode by experiment. gine/motor as zero. And, experiments on 2 ↔ 3 upshift/downshift are carried out because heavy-duty vehicles frequently use upper gear than 2-speed. Fig. 8, 9 show the experimental results of the upshift/downshift that was obtained by applying the basic shift algorithm.4)

The large shift shock arises at the moment of clutch engagement and right after the shift. In the moment of clutch engagement the minus torque is generated, owing that the 3. Moving the Y-axis cylinder, controlling the speed of speed of the clutch plate connected to engine is lower than motor to engage the synchronizer. that of the clutch plate connected to AMT (Fig.8 (a)-(1), (b)).

Fig. 8 (a)-(2) shows the large transient torque generated after shift, the reason is that the controller controls the torque of engine and motor during shift as the reference value, but it suddenly controls them as the driver’s pedal stroke after shift.

And the speed difference of clutch plates in engagement is 4. Moving the X-axis cylinder set to the reference value (50, 200rpm) but it wasn’t per- formed perfectly. The reason is that the controller doesn’t compare the reference speed difference with the present speed difference of clutch plates, but it compares the reference with present engine speed.

Fig.8 (c) shows that large transient torque is generated right after the engagement of clutch, the reason is that the controller raises the speed of engine rapidly to reduce quickly the speed difference of clutch plates for the engagement of clutch. 5. Controlling the engine speed to engage clutch plates. In engaging clutch plates, the speed difference of clutch plates must be reduced within the reference value for the smooth shift by controlling engine speed, but the dynamic characteristics of engine are so tardy, so the overall shift time is prolonged (Fig.8 (b)).

6. Completing shift after clutch is engaged.

(a) Torque (upshift).

Fig. 7. Schematic diagrams of shift sequence.

3 V) Due to the tardy response of engine, the overall shift time is prolonged. 3.3 IMPROVING SHIFT ALGORHIM THROUGH EXPERIMENT

(b) Velocity (upshift).

Fig.10 Torque (Upshift) - Experimental result 3 Fig.10 shows the result of a modified shift algorithm in which the controller restores the output torque of engine and motor linearly from the reference value to the value of

(c) Torque (downshift). driver’s pedal stroke during 1-second after shift. It shows Fig.8 Experimental result 1 (ref 50rpm engagement) the transient torques after the shift is moderately reduced than Fig.8, 9.

Torque (downshift). Fig.9. Experimental result 2 (ref 200rpm engagement). (a) Torque (upshift). The problems of basic shift control algorithm are summarized as follows. I) If the speed of the clutch plate connected to engine is lower than that of the clutch plate connected to AMT during the clutch engagement, the minus torque is generated. II) In downshift, large transient torque is generated right after the engagement of clutch; the

reason is that the controller raises the speed (b) Torque (downshift). of engine rapidly to reduce quickly the speed difference of clutch plates for the engagement of clutch. III) Right after the shift, the transient torque is generated, owing that the controller controls torque of engine and motor during shift as the reference value. But it suddenly controls them as the driver’s pedal stroke after shift. IV) The amount of the jerk during shift rises as

the speed difference between clutch plates (c) Velocity (downshift). increases (see Fig.8 (a), 9). Fig. 11. Experimental result 4 (without load)

4 Contrary to other conventional AMT using the clutch slip control for clutch engagement, we use the engine speed control to reduce the speed difference of clutch plate for clutch engagement. So it has a problem that shift time is prolonged due to the tardy engine dynamic response.

During shift, all power output is zero; the driver experi- ences a bad feeling if the shift time is prolonged. To overcome, the control algorithm is revised for the speed of output shift to follow the speed before the shift by using the motor power, while the controller controls the engine (a) Torque (upshift) speed during the clutch engagement.

Fig. 11-15 shows the result of shift experiments using the modified algorithm, which is explained above. Among them, 35kgf-load is applied to the results of Fig. 12-15 using eddy-current type dynamometer.

Because all the power-output is cut off during the disengagement of synchronizer, the speed reduction is indispen- sable in this step, but after the synchronizer engagement, (b) Torque (downshift) the output speed is maintained uniformly by using the motor torque.

(a) Torque (upshift) (b) Torque (downshift)

(b) Torque (downshift) (c) Velocity (downshift) Fig.12. Experimental result 5 (ref. 50rpm)

5 the clutch engagement by controlling engine speed for the speed difference of clutch plates to be within the 100 rpm than that of controlling engine speed for the speed difference of clutch plates to be within the 50 rpm

From the result of above shift experiments, we can find it is desirable to engage the clutch when the speed of clutch plate connected to engine is higher than that of the clutch plate connected to AMT, and the speed difference of clutch plates to be within the 50 rpm.

In this study, the advanced shift algorithm for the AMT in the parallel type hybrid drivetrain is developed, and we verified from dynamometer-based experiments. The results of this study are summarized as follows; 1. The minus transient torque is generated if the clutch is engaged when the speed of clutch plate connected to engine is lower than that of the clutch plate connected to AMT, and we can eliminate it by an adequate speed control of engine. (a) Torque (upshift) 2. The large transient shock is generated right after shift because of raising the speed of engine rapidly for quickly reducing the speed difference of clutch plates for the engagement of clutch. Re- storing output torque of engine and motor linearly from the reference value to the value of driver’s pedal stroke during 1-second can prevent it. 3. The magnitude of the jerk are lager as the speed difference of clutch plate is larger, so we can reduce it by controlling speed of clutch plate using engine and motor (b) Torque (downshift) 4. The overall shift time is prolonged owing to the tardy response of engine, so we used an algorithm to maintain the output speed during shift, by controlling the motor torque.

REFERENCES

1. Han-Sang Jo, Yeong-Il Park, Jang-Moo Lee, Hyeoun-Dong Lee and Seung-Ki Sul , “A Development of Advanced Shift Control Algorithm for Hybrid Vehicle with Automated Manual Transmission”, International Journal of Vehicle Design - Heavy (c) Velocity (downshift) Vehicle Systems , 1999 Fig.15 Experimental result 8 (ref. 100rpm) 2. S. K. Sun, et. al., "Advanced Control Strategy of Parallel Hy- brid Low Emission Electric Vehicle", IEEE Workshop on Power Fig. 12, 14 show the experimental results of engaging the Electronics in Transportation, Michigan 1996. clutch plates when the speed of clutch plate connected to 3. H.S. Jo, S.T. Cho, Y.I. Park and J.M. Lee “Analysis on the engine is lower than that of the clutch plate connected to Shift Characteristics of Automated Manual Transmission in the AMT, and Fig. 13, 15 shows a contrary case. We can see Parallel Type Hybrid Drivetrain System for Transit Buses”, In- the minus transient torque generated in Fig. 12, 14, but ternational Journal of Heavy Vehicle Systems, Acceped for Pub- Fig. 13, 15 shows no minus transient torque after the lication, 2000 clutch engagement. 4. R. Mizon and R. M. Tuck, "Automatic and Powershift Trans- mission for Trucks", SAE Paper No. 912486, 1991. Fig. 12 and 13 shows the experimental results of control- 5. P. W. Masding, J. R. Bumby, "A Microprocessor Controlled ling the engine speed for the speed difference of clutch Gearbox for Use in Electric and Hybrid-electric Vehicles", plates to be within the 100 rpm, and Fig. 14, 15 show the Trans. Inst. M.C., Vol. 10, No. 4, pp.177-186, 1988. experimental results of controlling engine speed for the speed difference of clutch plates to be within the 50 rpm. From Fig.12-15, larger transient torque is generated after