神经网络技术在桥梁构件声波检测中的应用研究

Application of the neural net technique in the sound wave detection of bridge members

随着桥梁工程建设的发展,其混凝土构件质量检测显得越来越重要。采用非线性的BP神经网络方法技术,利用声波的运动学和动力学特征参数,在已知样本参数的约束下,对桥梁混凝土构件缺陷进行多参数综合检测,就方法技术而言是可行的。通过对某大桥混凝土腹板裂缝采用非线性BP神经网络方法技术进行综合检测,其结果验证正确,证明了该方法技术的有效性,且较常规的线性方法技术有着更高的精度和可靠性。为BP神经网络方法技术的应用探索了一个新的领域。

With the development of engineering construction of the bridge, the concrete component quality testing seems more and more important. Adopting nonlinear BP neural network method technology and utilizing kinematics and dynamics characteristrc parameters of sonic wave, the authors make multi-parameters synthetic detection to the bridge concrete member defects under the restraint of the parameter of the known sample. This is feasible as regards method technology. Through adopting nonlinear BP neural network method technology measure synthetically some bridge concrete web cracks, the result is verified correctly. This proves the validity of this method technology which has higher precision and dependability than the routine linear superposition method, And it has explored a new field for the application of BP neural network method technology.