Artificial Neural Network in Setting Topology Optimization Parameters

Structural topology optimization procedures aim at specifying the best distribution of materials in space so that the functional properties of structures to be optimum. However, successful application of these approaches is highly dependent on the experience of the user in setting initial optimization parameters. The process of finding optimization parameters is a time taking process and computationally costly. In this paper the ANN has been used for setting topology optimization parameters for material design. In particular the interest is to predict the level of isotropy of materials using artificial neural network (ANN). The ANN training with back propagation has been used. Fitted model profile and ANN interaction prediction model has been presented.

Keywords:

Artificial neural network , Topology optimization , back propagation , BESO , bulk modulus.