成都理工大学学报（自然科学版）2025，Vol.52Issue(5)：951-965,15.DOI:10.12474/cdlgzrkx.2024092301
基于人工智能的低压致密砂岩储层地层压力预测
Prediction of the formation pressure in low-pressure tight sandstone reservoirs based on artificial intelligence technology:a case study of the Well Jinqian 5H area,Sichuan Basin
摘要
Abstract
The Tianfu gas field in Sichuan Basin,including the Well Jinqian 5H area,exhibits significant potential for natural gas exploration and development in the Shaximiao Formation.It has been previously shown that the effectiveness of gas development is closely linked to formation pressure.The 6,7,and 8 sand groups of the second member of the Jurassic Shaximiao Formation in the Well Jinqian 5H area primarily consist of low-pressure sandstone,while most of the drilled wells are horizontal wells.However,commonly used methods for formation pressure coefficient prediction of wells are difficult to apply to the target strata of the study area.Based on our study of the source rocks,source-connecting faults,and sand group characteristics,two types of source-fault-sand formation pressure systems are constructed,and the genetic formation pressure in the target strata of the study area is clarified.Select five types of well logging parameters as sensitive parameters for standardization.For the different source-fault-sand formation pressure systems,results from two formation pressure coefficient prediction methods are compared for data obtained from different wells.Our results show that linear regression is suitable for formation pressure coefficient prediction for vertical-inclined wells,as well as for horizontal wells within source-fault-sand formation systems that experienced abnormally low formation pressures;conversely we find that neural networks are more suitable for formation pressure coefficient prediction for horizontal wells in strata formed in low to normal pressures.Predictions based on the formation pressure systems and well data result in high accuracy.This study provides ideas and support for formation pressure prediction in low-pressure tight sandstone reservoirs in the Well Jinqian 5H area and the Tianfu gas field.关键词
天府气田/金浅5H井区/沙溪庙组/致密砂岩/地层压力/人工智能Key words
Tianfu gas field/Well Jinqian 5H area/Shaximiao Formation/tight sandstone/formation pressure/artificial intelligence分类
天文与地球科学引用本文复制引用
唐青松,王海霞,邓虎成,吴冬,朱德宇,关旭,王小娟,朱迅,张少敏,李楠,胡丽,赵娟..基于人工智能的低压致密砂岩储层地层压力预测[J].成都理工大学学报（自然科学版）,2025,52(5):951-965,15.基金项目
四川省科技计划项目(2021096). （2021096）
