炼钢过程中Cr的加入是一个逐步调整的过程,由于大量随机因素存在,使得Cr含量的测量值低于钢水中Cr的实际含量,从而导致过多Cr加入,成品钢中Cr含量偏高。为此,提出用RBF(RadialBasisFunction)神经网络模型来估计钢水中Cr含量,寻求一组过程变量,通过主元分析降低输入维数,训练神经网络,对钢水中Cr含量进行有效估计。该方法可以一次性调整Cr的含量到要求范围内。 During the steelmaking process, the addition of Cr is a process of gradual adjustment. Due to a large number of random factors, the measured value of Cr content is lower than the actual content of Cr in the molten steel, resulting in excessive Cr addition and high Cr content in the finished steel. Therefore, it is proposed to use Radial Basis Function (RBF) neural network model to estimate the Cr content in molten steel, seek a set of process variables, reduce the input dimension through principal component analysis, train the neural network and effectively estimate the Cr content in molten steel. The method can adjust the content of Cr to the required range once.