Abstract
This paper proposes a structural damage identification method using genetic algorithm. This method estimates the location and severity of damage by minimizing the difference between measured and estimated natural frequencies and by using some a priori information on damage such as test results and theoretical guesses. By giving a priori information which contains some correct information on the damage, the importance of a priori information and the influence of information quality on estimation accuracy are examined. As a result, it was shown that better a priori information produces better estimation accuracy and that the influence of errors in a priori information will be reduced in estimation.
