The established correlations of vertical multiphase flow in tubing are usually applicable to systems with similar laboratory and field conditions as were used for developing the correlations. This study concludes that the most suitable correlations for East Kalimantan Fields are Hagedorn and Brown (Hagdorn et al., 1966), Ansari (Ansari et al., 1994), and Duns and Ros (1963). In particular, this study presents an Artificial Neural Network model as a guidancee tool to choose the best vertical multiphase flow correlation. The model was developed from data from 73 fields in East Kalimantan region. Several Artificial Neural Network models for predicting the best correlation of vertical multiphase flow were investigated. The models were developed using 60 sets of training data, 6 sets of cross-validated data, and 13 sets of prediction data. The models were structured in several forms depending on the number of input variable used which were wellhead pressure, API gravity, water cut, GLR, liquid total production, vertical depth of gas lift mandrel, and gas total production. The results show that the best model has a 70% prediction accuracy.
展开▼
机译：管道中垂直多相流的已建立的相关性通常适用于与实验室和野外条件相似的系统，这些条件已用于开发相关性。这项研究得出的结论是，东加里曼丹油田最合适的相关性是Hagedorn和Brown（Hagdorn等，1966），Ansari（Ansari等，1994）以及Duns and Ros（1963）。特别是，本研究提出了一种人工神经网络模型作为选择最佳垂直多相流相关性的指导工具。该模型是根据东加里曼丹地区73个油田的数据开发的。研究了几种预测垂直多相流最佳相关性的人工神经网络模型。使用60组训练数据，6组交叉验证的数据和13组预测数据开发了模型。根据所使用的输入变量的数量，模型以几种形式构建，这些参数包括井口压力，API重力，含水率，GLR，液体总产量，气举芯棒的垂直深度和气体总产量。结果表明，最佳模型的预测精度为70％。
展开▼
