Title: CAD Based Design Optimization of Four-bar Mechanisms: a coronaventilator case study Show Chinese title

Title: 基于CAD的四杆机构设计优化：一种冠状通风机案例研究

Abstract: 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%. Show Chinese abstract

Abstract: 机制设计优化是一个有前景的研究领域，因为它能够在不牺牲性能的情况下实现更节能的机器。 然而，机器制造商实际上并不使用文献中描述的设计方法，因为这些算法需要过多的理论分析。 此外，文献中的设计综合方法主要利用启发式优化器，导致次优局部最小值。 本研究引入了一种方便的优化流程，使工业广泛采用成为可能，同时保证能够揭示全局最优解。 为了确保找到全局最优解，对描述可能设计可行区域的约束的数学表达式非常重要。 因此，讨论了点到点（PTP）平面四杆机构的运动学分析，以获得静态和动态约束。 在可行区域内，通过CAD多体软件生成目标值样本。 这些运动仿真确定了在特定设计参数组合下完成运动所需的扭矩。 稀疏插值技术可以最小化所需的样本数量，从而减少CAD仿真次数。 此外，这种仿真结果的插值使得机器设计师无需深入的分析设计即可用数学描述表示目标。 随后，目标的数学表达式允许全局优化器在可行的设计空间内找到全局最优设计。 在一个具有三个设计参数（DP's）的冠状呼吸机机构案例研究中，从1870次CAD运动仿真中仅使用618次来构建模型，使该机构的RMS扭矩减少了67%。