基于BP神经网络技术的储层流动单元研究

Research on Reservoir Flow Unit Based on BP Neural Network Technology

黄珏油田方4阜一段储层属低孔隙度、低渗透率储层,储层特性较为复杂,在进行储层参数的求取时存在较大误差。结合取心物性资料、测井资料,选用流动带指数IFZ划分方法将取心井储层流动单元划分成Ⅰ、Ⅱ、Ⅲ类,并建立流动单元的识别和划分标准。在此基础上,利用BP神经网络技术对取心井储层流动单元进行学习训练,与测井曲线建立其相关的学习和预测模型,对非取心段储层流动单元进行预测,明显提高了测井解释精度,为储层精细评价提供一种较有效的研究方法。

Fang 4 block reservoir with low porosity and low permeability is complex in Huangjue oilfield, so, reservoir parameter calculated has bigger error. Combining with coring formation property material and log data, the reservoir is divided into three types of flow units by flow zone index（ Ivz ）. Established are the recognition and division standards of the flow units. Based on this, BP neural network technology is used to learn and train the reservoir flow units of coring wells. With such a technology, directly built is the mapping relation between log responses and flow unit types so as to learn and predict the flow units in the coring wells or non-coring wells. Log interpretation accuracy is obviously improved, which provides an effective way for fine reser- voir interpretation.