Damage detection of a cracked column via a neural network approach

摘要: Publisher Summary The chapter discusses the damage detection of a cracked column via neural network approach. theoretical model using an equivalent rotational spring to describe local flexibility edge crack on simple under compression is employed compute natural frequencies for compression. By considering different compressions, and change sizes locations along column, are computed from characteristic equation derived analytical solution. Then, as inputs, patterns size location outputs, database training tool back-propagation networks (BPN) constructed in chapter. According results testing by trained BPN, numerical example shows that BPN useful predicting applied compressive force size-location column.