应用基于密度聚类的频繁模式增长算法进行煤层气产能分析

Analysis of CBM Productivity by Using FP-Growth Based on DBSCAN

由于传统的煤层气产能分析算法存在影响因素不够全面,运行效率低和人为设置聚类参数缺乏说服力的问题。因此,该文在煤层气产能分类的基础上,对分类结果进行回溯,挖掘煤层气产能影响因素的规律,将基于密度聚类算法(Density-Based Spatial Clustering of Application with Noise,DBSCAN)与频繁模式增长算法(Frequent-Pattern Growth,FP-Growth)关联度分析算法优化结合,提出新的基于DBSCAN的FP-growth煤层气产能分析模型,找出影响煤层气产能的关键因素及其对应的参数范围。该文是深度学习与煤层气开发交叉学科的应用与研究,致力于煤层气产能分析评价体系的构建,为提高煤层气单井产气量,提升措施选井的决策效率有积极影响。

The traditional analysis algorithm of CBM productivity has the problems such as the incomplete influence factors,low operation efficiency and unconvincing artificial setting of clustering parameters,so based on the classification of CBM productivity,this paper traces classification results,excavates the laws of the influence factors of CBM productivity,optimizes and combines the correlation analysis algorithm based on DBSCAN and FP-Growth,puts forward a new analysis model of FP-Growth CBM productivity based on DBSCAN,and finds out the key factors affecting CBM productivity and their corresponding parameter range.This paper is the application and research of deep learning the interdisciplinary of CBM development,and is committed to constructing the analysis and evaluation system of CBM productivity,which has a positive impact on improving the gas production of the single CBM well and promoting the decision-making efficiency of measure well selection.