基于主成分分析和最小二乘支持向量机的油田产量预测模型

The Oilfield Production Prediction Model Based on Principal Component Analysis and Least Squares Support Vector Machine

油田产量预测是油田开发生产中的重要工作,也是油田开发决策的基础。为了准确且快速地进行油田产量预测,本文首先采用主成分分析方法对预测模型的输入变量进行降维优化处理;其次,利用自适应粒子群算法对支持向量机的核参数σ和惩罚因子C进行优化;最后,建立基于最小二乘支持向量机预测模型。实验结果表明,该预测模型有效地去除了冗余信息,降低了输入变量的结构复杂度,提高了模型寻优的收敛速度且避免陷入局部最优,有效地提高了油田产量的预测效率和预测精度。

The oilfield production prediction is an important work in oilfield development and production, is the basis of oilfield development decisions. In order to accurately and quickly predict oilfield production, firstly this paper uses principal component analysis method to reduce the dimensionality for the input variables of forecast model; Secondly, it optimizes the nuclear param- eters and the penalty factor of support vector machine by using adaptive particle swarm optimization algorithm; Finally, the fore- casting model based on least squares support vector machine is built. The experimental results show that the prediction model can effectively remove the redundant information and reduce the structure complexity of input variables, and improve the con- vergence speed of model optimization and avoid falling into local optimum, effectively improve the efficiency and the predic- tion accuracy of oilfield production forecast.