论文部分内容阅读
油田产量预测是油田开发动态分析最重要的内容之一,也是油田开发优化决策的基础。在介绍最小二乘支持向量机(LS-SVM)及遗传算法(GA)的原理基础上,建立LS-SVM-GA模型,并用该模型对某气田天然气产量进行预测。通过二个性能指标将其与LS-SVM和BP神经网络模型进行对比,结果表明,在样本有限保证一定精度的情况下,LS-SVM-GA模型的预测精度较高,范化能力较强,能够利用该模型对气田天然气产量进行预测。
Prediction of oilfield production is one of the most important contents in the dynamic analysis of oilfield development and also the basis of oilfield development optimization decision-making. Based on the principle of least square support vector machine (LS-SVM) and genetic algorithm (GA), the LS-SVM-GA model is established and the natural gas production of a gas field is predicted by this model. Compared with LS-SVM and BP neural network models, the results show that LS-SVM-GA model has higher prediction accuracy and stronger normalization ability under the limited accuracy of sample, The model can be used to predict gas field gas production.