摘要
在分析煤层含气量影响因素的基础上,采用基于小样本理论的支持向量机(SVM)回归方法,建立了预测煤层含气量的计算模型。通过对沁水盆地南部目标煤层含气量影响因素分析,建立了煤变质程度、储层压力、温度及煤质特征支持向量机模型并进行了训练和测试。结果表明:SVM模型预测结果和实测结果误差小,为煤层气资源的勘探开发提供参考,是预测煤层含气量的新途径。
Based on the analysis of the influencing factors of coalbed gas content,support vector machine(SVM) regression on the basis of small sample theory,was used to build a calculation model of gas content.Through analyzing the influencing factors of the target coal seams gas content of the Qinshui southern basin,the SVM model of the influencing factors test data,such as coal metamorphic degree,reservoir pressure,reservoir temperature and coal characteristics is established,trained and predicted. The results shows that the error between predicted results and those analysed by sample test is small,the method of building SVM model provides reference for exploitation of the CBM resource,and it's a new approach of exploration coalbed gas content.
出处
《西安科技大学学报》
CAS
北大核心
2010年第3期309-313,共5页
Journal of Xi’an University of Science and Technology
基金
国家"973"项目(2009CB219600)
国家自然科学基金项目(40872104
40730422
2008ZX05-03404)
关键词
煤层气
含气量
支持向量机
coalbed methane
gas content
support vector machine