Manufacturability in automatic synthesis of planar rigid body spring mechanisms

Abstract

For more than 130 years there have been efforts to automate human services. One such automation is the use of computer-based solving methods to aid engineers in their search for their ideal goal, such as automated planar mechanism design. Contrary to creating mechanisms manually, automatic methods are able to create and evaluate mechanisms in quick succession. However, computers can often reach intricate designs which are not directly feasible in the physical domain. These programs do not focus on being able to manufacture the mechanisms, thus, the designs often have features that make it difficult or even impossible to create physically. Examples are inadequate spacing between joints or a mechanism consisting of many elements for which a good layer assignment is hard to find. To ensure these designs are physically feasible, a manufacturability check can be implemented. Sen et al. created a method to obtain manufacturable planar mechanisms for kinematic designs consisting of rigid bodies and revolute joints. This method identifies structural, kinematic or geometric infeasibilities for all layer assignments. When a mechanism is fully feasible, it can be manufactured. However, this method is still limited. It is only applicable for mechanisms consisting of rigid bodies and revolute joints. This restricts the application space of possible designs. Also, this method is not tested with physical examples. In this research, an extension of automatically created manufacturable mechanisms is developed. Firstly, it extends the method such that spring elements in a design can be checked for manufacturability and surrounding links can adapt their shape to accommodate for the springs, resulting in a wider field of applications. Secondly, a method is introduced to simplify the steps to manufacture the theoretical design. Four mechanism prototypes are built using the available methods and their manufacturability is evaluated. Lastly, this research can be used to find the layer assignment with the least amount of layers, resulting in thinner mechanisms for space-constrained applications.