神经网络系统在稠油区块优先动用排序中的应用

Application of neural network system to ordering of development sequence for heavy oil blocks

稠油油藏开发是一项投资大、风险大的工程。我国目前有大量的未动用稠油油藏。如果在这些油藏投入开发前,对它们的优先动用顺序进行排列,并划分出目前经济技术条件下没有开发价值的稠油区块,这将对稠油油藏的动用提供指导性意见,避免重大经济损失。为此,选取9个油藏参数,应用神经网络系统对大量已开发稠油油藏样本进行训练,以提取各参数的权重值;并在此基础上对稠油油藏进行综合评价,评价值高的稠油油藏就优先动用。经实例验证,该方法的评价结果与实际开发数据十分相符。该方法避免了人为给定参数权重的主观性,增加了评价的科学性。

Heavy oil reservoirs development means huge-investment and high risk project. A lot of heavy oil reservoirs in our country are not producing. If we can order the prior development sequence of these reservoirs and eliminate some viscous oil blocks without development value under current economic condition and technical levels, the helpful guide can be provided to avoid serious economic loss. Therefore, 9 reservoir parameters are selected, and neural network system is applied to obtain the evaluation parameters＇weighting values. Then, heavy oil reservoirs are comprehensively evaluated and reservoirs with higher values will be developed preferentially. Through some cases testing, it shows that the evaluation results can match with practical development data. This method can avoid man-made subjectivity for parameters weighting, and make the evaluation more scientifically.