【摘要】 基于气测资料构造适当的综合指标,分别利用模糊模式识别和误差反向传播(BP)神经网络两种方法对储层含油气性进行了分析。结果表明,气测资料与储层含油气性具有较强的相关性。模糊模式识别方法可以用来确定气测综合指标与储层含油气性之间的模糊关系,而且对待识别样品气测资料的随机性具有较强的适应性;神经网络法能够较准确地建立气测综合指标与储层含油气性之间的非线性关系,但在待识别样品气测资料具有随机性的情况下,识别结果具有随机性。利用模糊模式和BP神经网络方法对储层含油气性进行识别具有一定的可行性,能快速为储层含油气性分析提供一定的参考依据,并且前一种方法的识别效果优于后一种方法。更多还原

【Abstract】 The methods of fuzzy model identification and error backpropagation(BP) artificial neural network(ANN) were used to analyze oil-bearing reservoirs based on gas logging information.The results show that there is a noticeable correlativity between gas logging data and oil-bearing reservoirs.The distinction of the principles results in the difference of the identifying results that the former is fuzzy and the latter is random. The fuzzy correlativity between gas logging data and oil-bearing reservoirs can be determined and the random property of gas logging data can be adjusted by fuzzy model.The method of error BP ANN can indicate the non-linear correlation between gas logging data and oil-bearing reservoirs accurately and be hard to adjust the random property of gas logging data.The methods used to identify oil-bearing reservoirs are feasible.The identifying results show that the former method is better than the latter method.更多还原