基于MABC-SVM的含瓦斯煤体渗透率预测模型  

Prediction model on permeability of gas-bearing coal based on MABC-SVM

作　　者：汤国水[1] 张宏伟[1] 韩军[1] 宋卫华[1] 

机构地区：[1]辽宁工程技术大学矿业学院,辽宁阜新123000

出　　处：《中国安全生产科学技术》2015年第2期11-16,共6页Journal of Safety Science and Technology

基　　金：国家自然科学基金资助项目(51104085)

摘　　要：对含瓦斯煤体的渗透率的有效预测,可为瓦斯抽采和瓦斯灾害的防治提供理论指导,采用改进的人工蜂群算法(MABC)和支持向量机(SVM)相结合对其进行预测。应用改进的人工蜂群算法优化支持向量机的核函数参数C和g,提高了支持向量机的预测准确性。选取有效应力、瓦斯压力、温度和煤的抗压强度作为影响含煤瓦斯渗透率的主要影响指标,结合实验室测试数据,建立MABC-SVM含煤瓦斯渗透率预测模型。研究结果表明:该模型具有较强的泛化能力,可以相对准确有效的对含煤瓦斯渗透率进行预测,为瓦斯渗透率的研究提供了新的研究思路。Effective prediction on permeability of gas-bearing coal will provide theoretical guidance for gas drainage and gas disaster prevention. Modified artificial bee colony algorithm and support vector machine were combined to predict coal gas permeability. The kernel function parameters C and g of SVM were optimized by modified artificial bee colony algorithm,and the prediction accuracy of SVM was improved. The effective stress,gas pressure,gas temperature and coal compressive strength were selected as main impact indicators of coal gas permeability. Combined with laboratory test data,the prediction model of coal gas permeability based on MABC-SVM was established. The results showed that the model has a strong generalization ability,which can be relatively accurate and effective to predict the coal gas permeability,and it provides a new research idea for study on gas permeability.

关 键 词：瓦斯渗透率 支持向量机 人工蜂群 优化算法 

分 类 号：X936[环境科学与工程—安全科学]