用BP神经网络在线检测松节油含量

Application of BP Neural Network in the On-line Turpentine Measurement

在用蒸汽蒸馏连续法生产松香和松节油的过程中,在线检测松节油的含油量是一难点。由于目前尚无此类仪表,因此提出用BP神经网络搭建软仪表模型在线检测松节油含量,以生产过程中较易检测到的两个量(温度和黏度)作为软仪表的输入量,间接检测出松节油的含油量。用MATLAB软件对模型进行了仿真,检验了方法的可行性,说明神经网络在辨识非线性系统模型中具有很好的应用前景。

In the process of producing turpentine,it's hard to measure the turpen tine directly. But based on theory of BP Neural Network, the model of indirect t urpentine measurement could be established. The inputs values are the temperatur e and the mucous,these two values are processed by the BP Neural Network(the sof t instrument).Then the feasibility was validated through simulate with MATLAB5.3 .