论文部分内容阅读
油藏模拟器可有效地预测油气藏的生产动态,具有很好的准确性。但是,在油气藏开发的初期阶段,对储集层物性和生产参数的认识有很大程度的不确定性。而专家系统就像一个筛选工具一样能够用较低的成本和较少的时间模拟油气藏的动态。本文所描述的专家系统是一种预测煤层气藏生产动态的工具,其作用与数值模拟器相类似。专家系统借助于大量的数据而得到训练,对于给定的煤层气藏能够预测大约10年的气和水的生产剖面。同时,模型的其他输出还包括:累积产气量、累积产水量、期望的峰值气产量、达到峰值产量的时间以及煤层气藏废弃时间。本研究还涉及到逆向专家系统的研发,它能够从所期望的煤层气藏方案中筛选出最佳的生产方案。而常规油藏模拟器只有对大量的方案进行研究之后才能提出最佳的设计方案。逆向专家系统能够有效地解决最优化问题。
Reservoir simulator can effectively predict the production of oil and gas reservoirs with good accuracy. However, there is a great deal of uncertainty regarding the understanding of reservoir physical properties and production parameters in the early stages of reservoir development. The expert system, like a screening tool, can simulate reservoir dynamics with lower costs and less time. The expert system described in this paper is a tool to predict the production of coalbed methane reservoirs, and its function is similar to that of numerical simulator. The expert system is trained with a large amount of data and can predict a gas and water production profile of about 10 years for a given CBM reservoir. At the same time, other outputs of the model include cumulative gas production, cumulative water production, expected peak gas production, peak production time, and CBM abandonment time. The research also involves the development of a reverse expert system that will screen the best production options from the desired CBM scenario. The conventional reservoir simulator can only propose the best design plan after studying a large number of schemes. Reverse expert system can effectively solve the optimization problem.