基于BP神经网络的波美度测量在结晶器中的应用

Application of BP neural network in measurement Baume mould based on

为了精确控制青海盐湖结晶器的在线工艺过程,研发基于BP神经网络的波美度测量模型。寻求可全局收敛的快速学习算法,以满足系统实时控制和良好的性能需要,通过分析结晶过程的物理化学反应,确定影响波美度预测的主要因素,根据现场提取的KCL、温度、离心机电流等历史数据对神经网络模型进行训练。通过仿真以及对实际值和神经网络预测值的均方差分析,表明该模型可以准确地预测和确定波美度,最终提高结晶器的控制精度。

In order to accurately control the process of the crystallizer, research and development of the Baume degree measurement model that based on BP neural network. For- meeting the real-time control system and good performance. We learning global algorithm, Through the analysis of physical and chemical reaction crystallization process, we determined the main factors about affecting the Baume degrees forecast, According to the training of the neural network model for field extracted KCL, temperature, current and historical data of the centrifuge. Through the simulation and analysis of the mean square deviation of the actual value and neural network prediction, show that the model can accurately predict and determine the Baume degree.Finally to improve the precision of the crystallizer.