充填钻孔寿命SVM优化预测模型研究

来源期刊:中南大学学报(自然科学版)2014年第2期

论文作者:张钦礼 陈秋松 胡威 高瑞文

文章页码:536 - 542

关键词:充填;钻孔寿命;支持向量机;遗传算法

Key words:backfill; drill-hole life; support vector machine; genetic algorithm

摘    要:基于充填钻孔是充填料浆从地表输送到井下采场的咽喉工程,是矿山正常运转的保障,对矿山充填钻孔使用寿命进行预测十分重要,建立支持向量机(SVM)回归预测模型,用训练集对模型进行训练,以验证集预测值的均方误差作为适应度函数,通过遗传算法(GA)对SVM模型参数进行优化选择,应用优化得到的SVM模型对预测集进行预测。以某矿为例,通过GA得到SVM模型最优参数:适应值(均方误差)为0.011 1,惩罚系数C为 47.076 8,核函数参数σ为2.263 8。采用优化的SVM模型对预测集充填钻孔寿命进行预测,预测结果的最大预测相对误差为8.6%,平均相对误差为5.2%。对比BP神经网络(最大相对误差为13.6%),优化的SVM模型预测结果更加理想,精度更高。

Abstract: Based on the fact that backfill drill-hole is a throat engineering when filling slurry transports to the underground stope from the surface, which is the base to mine normal operation, support vector machine (SVM) regression model was established for predicting service life of mine backfill drill-hole. The mean square error of the value was made as a fitness function. Then, the model parameters were optimized through the genetic algorithm (GA), and the optimized SVM was applied to predict the prediction set. Taking a mine as an example, its drill-hole life was forecast through the GA_SVM model calculation, the optimal parameters were gotten, and the accuracy of the model was given. The results show that the adaptive value (mean square error) is 0.011 1, penalty coefficient C is 47.076 8, and kernel function parameters σ is 2.263 8. The maximum prediction error of forecast results is 8.6%, and the average error is 5.2%. Compared with BP neural network (maximum error of 13.6%), optimized SVM model is better and has higher accuracy, so it has good promotional value in a similar predictive engineering.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号