用函数链神经网络建立油气水层解释模型

Buildind an Oil,Gas and Water Interpretation Model Using Function Link Nervous Nerne

为了克服用BP网络建模进行油气水层评价存在的误差,采用函数链神经网络建立识别模型。该文介绍了用此方法建立的三个模型:气测解释模型、地质测井综合解释模型、气测地质测井综合解释模型,并介绍了现场及完井后的解释评价情况。网络采用增强输入模式的方法,用单层网络训练样本,达到了理想效果,三个模型的识别率都是100%。

To avoid errors of evaluating the oil,gas and water layer with the model built through BP network,the recognition model has been make up through the function link nevous network.The paper presents three models built up with the network,that is,gas logging interpretation model,geo-welllogging comprehensive interpretation model and gaslogging-welllogging comprehensive interpretation model,and it introduccs the interpretation and evaluation.