基于有限元和神经网络的主动加热红外无损测厚技术

提出了通过对平板进行单面主动加热,根据平板的热响应测厚的红外无损测厚技术．使用有限元数值计算,得到了不同厚度、加热热流、加热温度、加热区域形状、加热面积下的碳钢平板热响应．使用BP神经网络(NN)建立了热响应与板厚之间的非线性映射,神经网络的训练和测试表明该方法具有较高的精度和较好的推广性．实际应用中可使用红外热像仪记录平板表面的热响应．提出的红外无损测厚技术具有单向性、不接触等优势,具有良好的应用前景．

Active heating infrared non-destructive thickness measurement technology based on finite-element and neural network

Infrared non-destructive thickness measurement technology based on the active heating and thermal response of the plate is proposed. Finite-element numerical calculation method is used, and thermal response of carbon steel plate is investigated under different thickness, heat flux, heating temperature, heating region shape and heating area. BP neural network (NN) is adopted to build up the non-linear mapping between the thermal response and the plate thickness. Training and testing show that this method could has high accuracy and good extendibility. Infrared camera can be used to record the thermal response in practical application. The infrared non-destructive thickness measurement technology proposed in this paper has the advantage of unilateralism and non-contact as well as good application foreground.