关键词: 油层污染, 测井数据, 神经网络, 数学模型, 评价

Abstract: The damaged degree of formation can be assessed with logging information. Based on those normal process methods, some practical methods and technique to evaluate reservoir contamination by means of artificial neural network and log parameters are introduced in the paper. According to some wells data from Huabei-Erlian oilfield, we developed its neural network mathematical model to evaluate the damaged formation, and conducted modeling prediction. All the results show that this method is feasible and simple, and can accurately predict the contamination parameters and evaluate the damaged degree exactly and reasonably. It can be taken as a new and successful approach to evaluate formation contamination quantitatively.　

Key words: Reservoir contamination, Log data, Nerv network, Math’s model, Evaluation