A NEW METHOD FOR DYNAMIC ANALYSIS OF STRUCTURES USING CELLULAR AUTOMATA

The emerging idea of Parallel Computing makes the use of new algorithms inevitable to deal with the problem of structural optimization. Simultaneous analysis and design has been a well known optimization method whose efficiency can be increased by the use of parallel computation methods. Cellular Automata are among those algorithms which their models are composed of regular lattice of cell or automata. Each cell may change its state at discrete, fixed times (iterations) according to a local rule. In this way overall global behavior of a system can be modeled by simple interaction of cells which only know local conditions. Local rules which are dependent on the present state of the cells and their neighbors are based on the behavior of system which should be modeled. This study contributes to the dynamic behavior of multi degrees of freedom (MDOF) structures. The proposed method establishes local update rules which act on the stories of the structures as system cells. Any of the stories interacts with its neighbors, while applying a ground motion record to the base of the structure which leads to its global dynamic response after some iterations. These analysis update rules tend to solve the global differential equation of motion by solving some local differential equations at the cell levels. In this method structural analysis is performed in time iterations with regards to the response of neighboring stories. What makes Cellular Automata an efficient computational tool for modeling large systems is its inherent massive parallel computation capability in which by the use of new parallel computer processors its efficiency can be increased rapidly.

Keywords:

Cellular Automata , dynamic analysis , parallel computing