基于改进型偏最小二乘法的高炉炼铁工序能耗预测方法

Energy-consumption forecasting method of blast furnace iron-making process based on improved partial least squares

针对小样本环境下,具有自变量之间多重相关性特点的高炉炼铁工序能耗预测问题,从预测角度利用"舍一交叉"验证方法对偏最小二乘回归模型进行了改进,提出了应用改进型偏最小二乘回归建立预测模型的方法。以我国某钢铁厂高炉炼铁工序的能耗预测为例,说明了改进型偏最小二乘回归法与普通偏最小二乘回归法相比,预测误差平方和能够降低86.76%。

Facing to the energy-consumption forecast problem of blast furnace iron-making process with characteristics of fewer samples and independent variables with multiple correlations, this paper improved the partial least-squares regression model from the angle of prediction. A real steel factory＇ s blast furnace iron-making process is taken as an example to show that the sum of prediction error square of the improved partial least-squares regression can be reduced by 86.76% than that of partial least-squares regression.