煤体瓦斯渗透率的PSO-LSSVM预测模型  

Model of coal gas permeability prediction based on PSO-LSSVM

作　　者：邵良杉[1] 马寒[1] 

机构地区：[1]辽宁工程技术大学系统工程研究所,辽宁葫芦岛125000

出　　处：《煤田地质与勘探》2015年第4期23-26,共4页Coal Geology & Exploration

基　　金：国家自然科学基金项目(70971059);辽宁省科学研究计划资助项目(2010230004)

摘　　要：结合有关煤体渗透率的众多研究成果,总结出影响煤体渗透率的3个主要因素为有效应力、温度和瓦斯压力。采用粒子群优化算法(PSO)对最小二乘支持向量机(LSSVM)的参数进行优化选择,并以上述3个因素和抗压强度为输入值,以渗透率为目标输出值,建立煤体瓦斯渗透率的PSO-LSSVM预测模型。利用25组数据进行PSO-LSSVM模型与BP神经网络及支持向量机的比较实验,PSO-LSSVM预测结果与实际值拟合程度优于其他两个模型,且具有更小的误差。实验结果表明,采用PSO-LSSVM模型可由有效应力、温度和瓦斯压力对渗透率进行较高精度的预测。Three main influential factors of coal permeability were summarized, which were effective stress, tem-perature and gas pressure. The least square support vector machine was applied to predict permeability. The three factors and compressive strength were used as the input, the permeability as target output. Particle swarm optimi-zation algorithm was used to optimize the parameters of least square support vector machine to improve prediction precision. PSO-LSSVM was compared with BP neural network and SVM by a test. The comparative experiment results show that PSO-LSSVM can be used to predict permeability with high accuracy.

关 键 词：渗透率 瓦斯 粒子群优化算法 最小二乘支持向量机 

分 类 号：TD712[矿业工程—矿井通风与安全] P618.11[天文地球—矿床学]