BP神经网络在铝电解槽温度测量中的研究

针对铝电解槽温度高、腐蚀性强、温度难以直接连续测量的问题,引用BP神经网络技术,选择铝电解槽的输入电流、电压和下料速度作为辅助变量,建立BP神经网络模型,并利用matlab对数据样本进行了标准化处理,实现了BP神经网络的建立、训练和仿真。在预焙阳极铝电解槽进行实验,通过预算温度与实际温度进行对比,结果证明了该方法的有效性。

Research of BP neural network in aluminum cell for temperature measurement

The temperature of the aluminum cell is difficult to be directly measured, because of strong corrosion, high temperature. We use BP neural network and aluminum cell is input current, voltage and speed as the subsidiary variable to establish model. Then we normalize data sample in matlab, finishing the BP neural network training and simulation. In pre-baked anode aluminum cell for experiments, we compared with the actual temperature and the simulation temperature, the result proves the validity of this metho...