用ＧＰ算法以及修正ＧＰ算法对高炉风量、料柱压差、鼓风压力等高炉过程变量时间序列的相关维数进行了移动估算和预报能力分析。结果表明：在研究范围内不能发现高炉炉况吸引子存在；而预报能力的变化与高炉炉况变化前兆之间存在明显对应关系，可用于炉况辨识。
The correlation dimension for the time series of blast furnace process varisble,such as blast volume, burden pressure drop and blast pressure etc.,have been estimated by using GP and modified GP algorithms.The predictublity analysis for the process variables have also been investigated.The results show that it is difficult to determine the chaotic attractor of blast furnace condition.It has been found that the predictablity for process variables can be used to niterpret and identity the precursor of biast furnace condition.
