A multi-component mechanical system is a mechanical assembly composed of several parts that are connected through joints. To design such a system, isolating its different components and designing (optimizing) them separately and putting them back together is a common routine practiced most often. However, this simplistic approach may be fundamentally wrong, as to design a system toward its optimal performance, it must be considered as a whole.
Designing a system (not just mechanical ones) can be perceived as a bi-level optimization problem (i.e. design process). In the first level, the system itself must be designed. This means deciding on the type and number of the components which it is made up. The second level, then, finds the properties of each of the parts selected previously such that the desired performance criteria are met.
As an example, suppose the goal is to design a landing gear for a given airplane from the following set of primitive components and joints so that at the touchdown moment it can stand the applied loads while causing a minimum displacement/acceleration for the airplane.
There is an infinite number of ways of combining these components and it is a highly time-consuming and computationally intensive task to explore all these options and select the best amongst them. Therefore, to search this large design space, we need to have a directed way to efficiently iterate through it. And this is the focus area of this project. Our findings and methods developed so far have been published as provided below:
- Multi-speed Gearbox Synthesis using Global Search and Non-convex Optimization, accepted for CPAIOR, 2020
- Configuration Design of Mechanical Assemblies using an Estimation of Distribution Algorithm and Constraint Programming, IEEE Congress on Evolutionary Computation (CEC), 2019
- Design optimization of dynamic flexible multibody systems using the discrete adjoint variable method, Computers & Structures, 2019
- Constraint-oriented programming approach to mechanical assembly design, U.S. Patent, 2019 (pending)
- Techniques for applying generative design to the configuration of mechanical assemblies, U.S. Patent, 2019 (pending)
- Efficient sensitivity analysis for generative parametric design of dynamic mechanical assemblies, U.S. Patent, 2019 (pending)