焦炭机械强度预测模型的研究

摘要

The coke mechanical strength is an important index of coke quality, and an accurate predictive model of coke mechanical strength can ensure the coking production going well.Predicting the coke mechanical strength by support vector regression and detailedly analyzing the different forecasting effects with the different indexes of the blend coal and the coking process,the effective characteristic parameters are founded to establish the coke mechanical strength model. With the characteristic parameters selected, when 85≤M25≤90 or 6≤M10≤12, the predictive accuracy of the support vector machines model is high, the predictive error is small, the correlation is strong, and the generalization ability of the prediction model is great.%焦炭机械强度是焦炭质量的重要指标,建立准确的焦炭机械强度预测模型可保证炼焦生产顺利进行.把支持向量回归机用于焦炭机械强度的预测,详细地分析了不同配煤煤质特性指标及炼焦工艺指标情况下的预测效果,找到了建立预测模型的有效特征参数.基于所选特征参数,当85≤M25≤90,6≤M10≤12时,支持向量机模型的预测准确率高,误差小,相关性强,同时,预测模型的泛化能力强.