| Sensitivity and parametric analyses are powerful tools in assisting engineers to improve design, control and operation of mechanical systems. Sensitivity derivatives help designers to understand the behaviour of the system at hand and modify the design to achieve more satisfactory results. Moreover, these derivatives are beneficial in optimum design, optimal control algorithms, and parameter identification methods. In addition to all these applications, sensitivity derivatives can be used to develop methods for parametric analysis of multibody systems; the formulation of such a method is also described in this thesis.;The thesis includes several sections: first, the sensitivity analysis of multibody systems is categorized into two groups: the inverse- and the forward-dynamics based problems. Next, an overview of the literature on different mathematical methods for calculation of the sensitivity derivatives is presented. Then, the dynamics parametric analysis problem is formulated. A parametric analysis method is presented, which relies on the interpretation of a linear space based on sensitivity derivatives. Further to the analysis mentioned above, a parametric analysis toolbox is created in the Simulink environment which complements the Multibody Toolbox (MuT) developed by the Canadian Space Agency. To illustrate the material, two problems are considered to apply the proposed sensitivity and parametric analyses approach on them. First, parametric analysis of contact/impact dynamics for a five-bar linkage is performed. The performance indicator is the pre-impact constrained motion space kinetic energy for contact/impact problems. The performance indicators of these contact/impact systems are related to characterizing the intensity of the contact transition during a change in topology. Second, a combined physical-virtual system is considered and effect of the variation of the inertial parameters of the virtual model on the joint torque variations of such a system is investigated. Simulation and experiments are performed using a setup that contains a physical device and a virtual dynamic system. Experimental results are compared to the simulation results. |
