자동차 공기현가 시스템의 차고 및 자세 제어에 관한 연구

Abstract

Electronically controlled air suspension systems are frequently used in road vehicles (mainly sport utility vehicles and luxury vehicles) to improve ride comfort and driving safety by adjusting vehicle height. The main focus of the air suspension is to adjust the static height of the vehicle sprung mass (height control) and to regulate the roll and pitch angles of the vehicle body (leveling control) regardless of payload variation, in order to facilitate passenger entry and exit, to reduce the air drag on the vehicle, and to optimize the ride comfort and driving safety. There are several challenges in control of the air suspension system, such as the complexity of the control system owing to controlling four air springs simultaneously with six valves and a compressor, nonlinearities due to air compressibility and the airflow through the solenoid valves, time-delay due to airflow transition between air charging and discharging, and uncertainties mainly caused by payload variations. Especially, the air compressibility and time-delay may cause significant vertical overshoots and oscillations of the vehicle sprung mass and make the height and leveling control difficult. Because the dynamic characteristics and payloads are different at four corners of the vehicle body, undesired roll and pitch angles may be generated during the height control. This may not only make the passengers uncomfortable, but also degrade the road handling capability. Therefore, the air suspension system should also level the vehicle body to be maintained at appropriate attitude by controlling the relative lengths of the air springs at four corners. The rule-based control methodologies are conventionally employed in production vehicles equipped with the variable ride height air suspension system. The rule based control methodologies are appropriate for ‘system control’ determining desired heights of four corners of the vehicle body from driving conditions and driver's commands, but not appropriate for ‘actuator control’ achieving the desired heights by controlling actuators (six valves and a compressor) due to limited performances. Conventionally, a stepwise height control as the actuator control has been used. The stepwise height control adjusts the front and rear corners alternately until the four corners reach their corresponding desired heights. However, their control accuracy is unsatisfactory in terms of tracking the desired height and regulating pitch and roll motions: if the air suspension system is controlled for the vehicle body to lift 20 mm, then the steady state error more than 5 mm (25%) is produced and significant pitch and roll motions are also generated. In spite of the unsatisfactory performances, the robust height and leveling control of the air suspension system to solve the above mentioned challenges and improve control accuracy has not been studied in detail yet. Especially, almost no researches on the height and leveling control of an air suspension system in an analytical way have been developed. In this dissertation, in order to improve control accuracy and robustness against complexity, nonlinearities, time-delay, and uncertainties in the automotive air suspension system, a very systematic methodology for designing height and leveling control is proposed. At first, a rule-based height and leveling control (conventional control) for the automotive air suspension system is analyzed and developed to establish the need for additional research in the area of height and leveling control of the air suspension system. After that, a non-model based control, a PD-type fuzzy logic control is proposed and applied to the height and leveling control to improve height control accuracy and regulate the roll and pitch angles of the vehicle body. As the main part of this study, a more advanced model based control scheme called sliding mode control is studied for robust controller design in an analytical way. Lastly, a new fault tolerant control (FTC) algorithm is presented to acquire fault tolerance ability during fault situations. The effectiveness and performance of the proposed control algorithms are verified by simulations and actual vehicle tests.