【摘要】 应用人工神经网络技术,采用分步诊断的方法,提出了反映结构损伤位置和程度的健康诊断方法。文中通过模态分析得到基于固有频率和位移模态的网络输入特征参数,分别利用概率神经网络和BP网络对结构损伤的位置和程度进行识别,结果显示该分步诊断的方法能够对结构的损伤正确识别。更多还原

【Abstract】 An artificial neural network(ANN) based damage detection method for frame structure is proposed.With the characteristic parameters based on inherent frequency and modal displacement,the artificial neural network of probability (PNN) and BP network were used to identify the depth and position of crack fault in structure.A numerical result shows its suitability and effective for structural damage detection.更多还原