基于神经网络的铜锍吹炼终点预报技术

The Technology of Copper-matte Converting Endpoint Prediction Based on Neural Network

准确预报铜锍吹炼终点 ,对于铜锍吹炼操作至关重要。采用改进的BP神经网络算法和自适应残差补偿算法建立铜锍吹炼终点组合预报模型。利用某厂实际生产数据进行仿真运行的结果表明 ,本文建立的模型具有较高的预报精度和较强的实时性 ,可用于指导生产实践。

The accurate prediction of copper matte endpoint is of great importance to copper matte converting practice. It is the first time that a prediction model based on an improved BP neural network and error compensation of linear regression has been developed. By simulating test, it is proved that the model possesses upper precision and practicability.