学术文献库

Design optimization of mechanisms is a promising research area as it results in more energy-efficient machines without compromising performance. However, machine builders do not actually use the design methods described in the literature as these algorithms require too much theoretical analysis. Moreover, the design synthesis approaches in the literature predominantly utilize heuristic optimizers leading to suboptimal local minima. This research introduces a convenient optimization workflow allowing wide industrial adoption, while guaranteeing to reveal the global optimum. To guarantee that we find the global optimum, a mathematical expression of the constraints describing the feasible region of possible designs is of great importance. Therefore, kinematic analysis of the point-to-point (PTP) planar four-bar mechanism is discussed to obtain the static and dynamic constraints. Within the feasible region, objective value samples are generated through CAD multi-body software. These motion simulations determine the required torque to fulfill the movement for a certain combination of design parameters. Sparse interpolation techniques allow minimizing the required amount of samples and thus CAD simulations. Moreover, this interpolation of simulation results enables the representation of the objective in a mathematical description without in-depth analytical design analysis by the machine designer. Subsequently, the mathematical expression of the objective allows global optimizers to find a global optimal design within the feasible design space. In a case study of a coronaventilator mechanism with three design parameters (DP's), 1870 CAD motion simulations from which only 618 are used to build a model allowed to reduce the RMS torque of the mechanism by 67%.