【摘要】 针对传统混凝土内部钢筋检测算法仅能完成混凝土保护层厚度检测而无法同时对钢筋直径进行有效估算的弊端,结合神经网络、支持向量机等软测量算法,在已研制检测硬件结构的基础上通过算法建立线圈响应信号与钢筋直径、保护层厚度间的数学模型。对目前采集的数据进行相应处理后使用MATLAB建立BP神经网络模型,支持向量机模型及算法改进后的BP神经网络模型。通过对比3种数学模型检测误差,最终选择算法改进后的BP神经网络模型作为最终数学模型,其检测效果满足实际要求。更多还原

【Abstract】 In view of the drawbacks that the traditional internal steel rebar algorithm could only deduce the thickness of the concrete cover by a single method and had no ability to estimate the reinforcement diameter at the same time. Combined with soft measurement algorithms such as neural network and support vector machine,the mathematical model between the coil response signals,the diameter of reinforcement and the thickness of protective layer are established by MATLAB after dealing with the data collected at present. By comparing the error of the BP neural network model,the SVR model and the modified BP neural network model,the improved BP neural network model is finally selected as the final mathematical model,of which the detection results meet the actual requirements.更多还原