Active suspension simulation through software interfacing

Active suspensions provide favourable characteristics over traditional passive vehicle suspensions since they are able to find a better compromise between ride and handling, a conflict that plagues all conventional suspensions. However, active suspension has yet to break into production vehicles because of the technical issues that remain to be resolved to improve its implementation.; Effective virtual simulation of such a system requires a method of properly modeling a multi-domain system. Software interfacing is a method that may be used to solve such problems. It allows each sub-system to be modeled in its natural domain software and then links the sub-systems together, allowing the input and output values for each program to be exchanged with the rest of the model. With this technique, modelling simplifications of each domain is avoided by allowing a complete and accurate picture of the system to surface before prototyping begins.; This research focuses on simulating active suspension by combining the multibody dynamic software program of ADAMS (Automatic Dynamic Analysis of Mechanical Systems) with Matlab/Simulink. The purpose is to capture the dynamics of the system which would allow the user to tune and optimize the suspension before prototyping. Since it is geared towards passenger vehicles, this study focuses on the ride behaviour of the vehicle rather than its handling abilities.; A quarter car and full car model are implemented for both the traditional lumped mass model and the Bombardier Iltis utility truck. When interfacing with the Simulink controller, nonlinear and linear versions of the ADAMS vehicle model were used; also a fully-active and semi-active suspension was evaluated for comparison with the passive suspension. In addition, a Kalman filter for state estimation was used with the fully-active controller, while bushings are added to the full car Iltis vehicle.; The nonlinear ADAMS vehicle was able to successfully communicate with Simulink to simulate the above systems. Results also indicate that the linear vehicle models are reasonable in their performance and so are useful for quick preliminary studies. Additionally, the simple lumped mass vehicle demonstrated similar response patterns and features as the more complicated Iltis models, further proving the worth of these models.; Results demonstrate that a fully-active suspension is able to significantly increase ride performance over a passive suspension but at the cost of the suspension displacement. As expected the semi-active suspension performance was intermediate to that of the passive and fully-active system but with the advantage of only dissipating energy and not consuming it. Both semi-active and fully-active controllers performed reasonably well for this investigation, however, shortcomings in each were noticed. The Kalman filter was generally able to estimate the system states which make the fully-active controller use all the more viable for real world application.