基于BP神经网络的加药控制系统研究

摘要

浮选是最重要的选矿方法之一,煤矿浮选工艺自动加药控制问题是影响浮选效果的重要因素；传统加药控制主要是由加药工根据生产经验来控制,为了解决浮选精度低这一问题,文中提出采用BP神经网络作为加药控制系统加药量的预测模型,并采用动量项方法对其进行优化,通过仿真得出动量项法网络和测试样误差曲线,说明带动量项改进的神经网络更容易满足网络精度的要求,并且该方法能够有效地降低误差、减小浪费.%Flotation is one of the most important mineral metheds, automatic dosing control problem of coal flotation process is an important factor of influence flotation effect. In traditional medicine control, workers mainly according to production experience to control, according to the low flotation accuracy, this paper puts forward the BP neural network as dosing control system of the dosage prediction model, and the method was improved. Through the simulation results, the method can reduce the error and reduce waste effectively.