The focus of this research is on the development of a novel multi-disciplinary design optimization scheme for micro-compliant devices. Topology optimization is a powerful tool that can address the need for a systematic method to design MEMS. It is expected that systematic design methods will make the design of micro devices transparent to the user and thus spur their use. Although topology optimization of MEMS devices with embedded actuation has received a great deal of attention among researchers recently, there is not a significant amount of literature available on the subject. The limited literature available addresses multi-physics topology optimization, which employs the homogenization method. However, the products of this method inherit the drawbacks of homogenized material discretization, including checkerboard pattern, gray-scale material and narrow flexural hinges in the optimum solution. In this thesis, a new topology optimization scheme is introduced that addresses the specific needs of MEMS domain. A new discretization approach with frame-ground structure is introduced. This approach offers significant conceptual and practical advantages to the compliant MEMS optimization problem, including compatibility with MEMS fabrication processes. The design spaces of compliant mechanisms are non-convex and it is critical to employ an algorithm capable of converging to the global optimum without the need to evaluate gradients of objective function. In this thesis, an efficient real-coded genetic algorithm is implemented, which shows a better repeatability and converges to very similar solutions in different runs. This new method of optimization facilitates the use of a coarse subdivision of the design domain rather than the homogenized material method, for the same resolution of shape definition. Therefore, the topology optimization scheme developed in this thesis significantly reduces the computational burden without compromising the sharpness of the shape definition. As the problem of compliant mechanism design is posed as a set of conflicting objectives, a well-posed multi-criteria objective function is introduced which avoids one objective dominating the solution. Moreover, the formulation is modified to incorporate electro-thermal boundaries and enables the optimization of the compliant mechanisms to transfer maximum motion or maximum force at the output. A number of design examples are used to demonstrate the ability of the procedure to generate non-intuitive topologies. Their performance is verified using ANSYS and compared with results from the homogenization method and designs reported in the available literature.
| Identifer | oai:union.ndltd.org:WATERLOO/oai:uwspace.uwaterloo.ca:10012/4042 |
| Date | 08 September 2008 |
| Creators | Mehrnaz, Motiee |
| Source Sets | University of Waterloo Electronic Theses Repository |
| Language | English |
| Detected Language | English |
| Type | Thesis or Dissertation |
Page generated in 0.0023 seconds