Automotive suspension system modelling and controlling

In both academic and industrial fields, suspension system modelling and associated control design influence vehicle response. Ideal hydraulic force models have been used in active suspension studies for decades, but few studies have investigated hydraulic effects, which are the core of system force generation. Accurate mathematical subsystem modelling is essential in representing physical subsystems and enhancing design estimation control. This thesis details the mathematical modelling of both passive and active suspension and controller design for a quarter-car test rig. When using a conventional passive model, a significant difference between the experimental and simulation results was found for improved modelling of body movements. This led to an investigation in how to resolve this issue, accordingly, the consideration of a new term (friction force) was researched. Establishing a nonlinear friction force became a vital aspect of this work. In addition, emphasis was placed on hydraulic modelling and unknown model parameters that were experimentally identified. This experimental work is unique and helpful for advancing knowledge of any system. A new approach to implementing the friction force was used to identify the system through the transformation of a ¼ car model to one Degree of Freedom (DOF) and two-DOF models. This reduced the model complexity and allowed the parameters to be identified from a series of transfer functions linking vehicle parts and the hydraulic models. Simulation and experimental results were then compared. The hydraulic component model is crucial to the formulation of accurate active control schemes. Full-state feedback controls were realised by Pole-Assignment (PA) and Linear Quadratic (LQ) optimal method. Simulation results suggest that even though the performance of active suspension designed by the PA method is superior to that of passive suspension, it still possesses a design constraint, similar to a passive system, as the design is a compromise between the effects of natural frequency and transmissibility. With a different design concept, the LQ method provided a better solution as it reduced energy consumption by 65% and effectively shifts the dominant natural frequency to a very low-frequency range. Thus, allowing the damping rate to be increased to its critical value with the smallest effect on transmissibility. iv It was estimated for experimental work that the identified model with the LQ controller might be used to predict the dynamic responses of the actual system within a certain range of the design parameters due to the considerable difference between the initial condition of the test rig and the linearised operating design. The servovalve produced issues that did not allow validation of the controller. Both simulation and experimental results, with several conditions, showed consistent agreement, between experimental and simulation output, consequently confirming the feasibility of the newly approved model for passive and active suspension systems that accounted for the actual configuration of the test rig system. These models, that subsequently implemented the nonlinear friction forces that affect the linear supported body bearings, are entirely accurate and useful. The nonlinear friction model captures most of the friction behaviours that have been observed experimentally. Additionally, the models of the nonlinear hydraulic actuators, covered by the dynamic equation for the servovalve, are moderately precise and practical. The suggested Proportional Integral (PI) control successfully guided the road hydraulic actuator and validated the control strategy. The suggested PA and LQ controllers for active systems successfully guided the system to achieve the targets. Ride comfort and handling response are close to that expected for the passive suspension system with road disturbances, whereas there were clear response enhancements for the active system