磁流变液
执行机构
扭矩
粘度
机械工程
材料科学
生物医学工程
计算机科学
工程类
结构工程
物理
阻尼器
复合材料
热力学
人工智能
摘要
The motivation for this study is a magnetorheological (MR) prosthetic knee joint. The MR prosthetic knee is adaptive in real-time via an MR fluid. Problems have been experienced when the prosthetic device is used in demanding situations, like hill-climbing for example. The goal of this project is twofold; to increase the on-state torque output of the prosthetic device and, at the same time, to decrease the off-state torque. To achieve the two goals, a combined MR fluid design approach and an MR actuator design approach is adopted. An MR fluid is designed that is tailored for this specific application. Parallel to the fluid design, an MR rotary brake actuator is designed for the prosthetic knee. The MR fluid design is approached by experimentally evaluating twenty-two potential MR fluid compositions for the proposed device. These fluids are mixed and rated against the design objectives of the prosthetic device. Potential MR fluid compositions include; unimodal MR fluids, with a varying particle size and a variable solid loading, bimodal MR fluids with two grades of micron-sized particles, and bimodal MR fluids with nanoparticles. A fluid figure of merit is defined that is the ratio between the on-state shear yield stress and the off-state viscosity. The fluid with the highest ratio will result in the widest torque range for prosthetic device. The MR actuator design is approached by building torque models of the device and optimizing the models. An off-state and an on-state model are developed where the onstate model is based on a magnetic finite element analysis. Design objectives for the device are: high on-state braking torque, low the off-state rotational stiffness and low weight. Trade-offs between the design objectives are explored with a multi-objective optimization technique. The analysis suggests design improvements that have been realized in a newly built and an enhanced version of the MR prosthetic knee.
科研通智能强力驱动
Strongly Powered by AbleSci AI