Kinematical synthesis of 1-dof mechanisms using finite elements and genetic algorithms

The use of finite elements in both kinematic analysis and synthesis of mechanisms has shown good performance. Its only drawback is the need for an initial reasonable quality solution.

To address this topic, the use of genetic algorithms with a finite-element-based error function is proposed in this paper. This approach has not only shown good behaviour with simple mechanisms, but also with complex kinematic chains.

The energy-based error function, which has demonstrated such good behaviour in the optimization via second-order Newton–Raphson methods, presented some limitations. To solve them, a new distance-based error function has been developed. Simultaneously, new finite elements have been created to represent some joints and kinematic elements. The need to address some specific configuration problems has led to the development of a correction algorithm.