摘要
多层油藏的传统分层测试方法是基于瞬态流量压力耦合模型建立的,但是由于实际生产过程中进行井下瞬态流量测量具有很大难度,且精度较差,因此传统分层测试方法的后期解释结果受瞬态流量影响很大,经常出现较大误差。利用目前被大规模应用的永久式井下温度和压力传感器,提出了基于瞬态温度压力耦合模型的分层测试方法,在前期正模型和可行性研究的基础上,着重研究了该方法中最为关键的反模型部分。通过建立合理的目标函数,应用Levenberg-Marquardt算法,建立了用于回归多层油藏分层属性的反模型,并编制程序进行求解,实现了通过瞬态实测温度和压力对多层油藏分层属性参数(渗透率、伤害渗透率和伤害半径)的求解。通过双层合采和4层合采油藏2个算例,对该测试方法的测试流程、数据选取要求、解释步骤和解释精确度进行了分析。结果表明,所建反模型可以与前期工作建立的正模型配合使用,解释效果较好。基于瞬态温度压力耦合模型的分层测试方法大大缩短了测试时间,对生产影响较小,解释精度高。
Conventional multilayer transient testing method for multilayered reservoirs is developed based on the transient flow ratepressure coupled model.Due to the complexity and accuracy problem of transient flow rate test,it is very hard to get accurate inter-pretation results for conventional multilayer tests.With the widely applied permanent downhole temperature and pressure sensors,the authors proposed a new multilayer testing method which is based on transient temperature-pressure coupled model.The forward model and the feasibility of the proposed multilayer testing method have been discussed in previous paper,and this paper will mainly introduce the inverse model development.By constructing a proper objective function and using Levenberg-Marquardt algo-rithm,the authors developed and programmed the mathematical regression model for inverting individual layer properties in multi-layered oil reservoirs.Meanwhile,this paper discussed the testing procedure,data selection requirements,interpretation steps and interpretation accuracy by using two synthetic examples.The proposed multilayer testing method is based on transient temperature-pressure coupled model,which shortens testing time and has less impact on normal production.The developed model has a great significance for reservoir management,reservoir monitoring,stimulation planning and evaluation.
出处
《油气地质与采收率》
CAS
CSCD
北大核心
2012年第1期99-103,118,共5页
Petroleum Geology and Recovery Efficiency
基金
国家科技支撑计划项目"低(超低)渗油气田高效增产改造和提高采收率技术与产业化示范"(2007BAB17B03)
关键词
多层油藏
分层测试
井筒流动模型
反模型
井下传感器
焦耳-汤普森效应
multilayered oil reservoir
multilayer transient test
wellbore flow model
inverse model
downhole sensor
Joule-Thom- son effect
