摘要
桥梁在其服役过程中容易产生桥体损伤,导致其承载能力下降、使用功能降低。频率参数在实际应用中测试获取容易,是良好的损伤辨识指标。考虑到神经网络技术收敛速度慢等缺点,采用遗传算法对其权值及阈值进行优化获取。采用频率变化平方比参数作为遗传优化神经网络的输入参数,以简支梁桥为数值模拟对象,实现了其损伤位置识别。
Bridge is easily damaged during its service which leads to the dechne of its carrying capacity and reduction of its function. Frequency parameters, which are easily obtained in test of actual application, are good damage recognition indexes. Considering the deficiencies of slow convergence rate of neural network technology, genetic algorithm is adopted to optimize the access to the weight and threshold and square ratio parameters of frequency change are adopted as the input parameters of neural network of genetic optimization. Taking simply supported beam bridge as the numerical simulation object, the damage position recognition is realized.
出处
《北方交通》
2012年第11期82-84,共3页
Northern Communications
关键词
简支梁桥
损伤识别
遗传优化神经网络
模态频率
Simply supported beam bridge
Damage recognition
Neural network of genetic optimization
Modal frequency
