关于高炉炼铁中热风炉送风温度预测的研究

Study on Prediction of Air Temperature in Hot Blast Stove of Blast Furnace Ironmaking

高炉炼铁作业时需要多个热风炉连续交替输送高热空气,以强化高炉冶炼、减少用煤量、增加产量,而烧炉期炉温的控制是最为重要的一环,空燃比的设定需要蓄热室平均温度的准确预测,提出了一种采用核模糊c-means的马尔可夫预测法,将热风炉送风温度记录看作是无后效性的事物,利用马尔可夫法实施预测,考虑到工业控制的复杂性,引入了改进的c-means算法,计算下一次送风终了时蓄热室的平均温度,以计算燃烧中所需煤气量。实验结果证明,改进后的聚类算法准确度提高明显;在与传统预测值比较中,改进的预测方法在温度区间划定以及温度值的拟合度都具备优势,为热风炉温度优化预测提供了依据。

The operation of blast furnace needs to heat the air in a row alternately with a number of hot air stoves, which can strengthen blast furnace smelting, reduce the amount of coal, and increase the output, and furnace temperature control period is the most important part of it. The setting of air - fuel ratio requires accurate prediction of the average temperature of the regenerator, based on this, the paper proposed a Markov prediction method based on fuzzy kernel c - means, it takes the record of hot blast furnace feeding＇s temperature as the thing of non - aftereffect property. Using Markov to implement prediction and considering the complexity of industrial control, we introduced the improved c - means algorithm, calculated the average temperature of the regenerator at the end of the air supply, in order to calculate the amount of gas required for combustion. The results show that the improved algorithm improved the accuracy obviously. In comparison with the traditional predictive value, the improved prediction method has the temperature range delineation and the fitting of temperature values advantage.