混凝土结构裂缝检测中网络算法的研究

Research on neural network algorithm in crack inspection of concrete structures

检测混凝土结构缺陷一般采用超声无损检测方法,传统的概率法在后期处理时有其自身的缺点.为了克服这些缺点,提出了采用蚁群优化算法与BP神经网络融合的方法进行后期处理,建立蚁群神经网络的数学模型,实现了蚁群神经网络的训练,并通过实例验证了该方法的有效性.

Ultrasonic nondestructive testing method is used commonly in concrete structure defect inspection, the post-processing of traditional probabilistic method has its own disadvantages. In order to overcome these short- comings, the ant colony optimization algorithm and BP neural network fusion method is presented in this paper, the mathematical model of neural network based on ant colony optimization is established and is trained, and ver- ify the validity of the method.