摘要
裂缝系统是个复杂的地质体,其储层物性(主要是渗透性)的改善作用是非线形的,各种评价参数与裂缝发育程度之间的关系也是非线形的,导致对裂缝进行定量评价十分困难,单纯依靠常规测井资料进行裂缝识别,存在主观不确定性及多解性;成像测井虽然直观准确,但成本较高。本文基于人工神经网络理论,开展了常规测井资料识别评价裂缝的研究。结果表明,基于BP神经网络的裂缝性储集层常规测井识别,与成像测井对比具有较好的应用效果。
The system of fractures is a complicated geologic body,the contribution of improvement to reservoir nature is nonlinear,the relationship between various evaluated parameters and extent of fractures growing is nonlinear too,which makes it difficult to evaluate fractures with definite quantity.There is subjective uncertainty and ambiguity in using routine logging data to recognize fractures.Although the imaging logging are intuitive and accurate,the cost is very high.Based on artificial neural network theory,using routine logging data in fracture identification is studied.The result shows that fracture identification has better application effect based on BP neural network compared with imaging logging.
出处
《新疆石油天然气》
CAS
2006年第4期39-42,52,共5页
Xinjiang Oil & Gas
关键词
裂缝识别
常规测井
成像测井
人工神经网络
fracture identification
routine logging
imaging logging
artificial neural network
