基于传递函数模型的油田产油量预测方法

针对油田产油量预测方法无法考虑因素动态关系的局限性,首次应用时间序列传递函数模型建立了考虑因素动态关系的多因素油田产油量预测模型。应用传递函数模型的原理并按照建模步骤,以胜利油区某油田新井产油量预测为例,说明了传递函数模型数据准备、模型识别、参数估计、诊断校验及预测过程。研究结果表明,该预测模型既能在繁杂因素中剔除无关因素的干扰,识别影响油田产油量预测的主控因素,又能考察变化非同步且具有滞后性的因素动态关系;拟合程度好,预测精度高达98.4%。

Oilfield output prediction method based on transfer function model

Dynamic relationship among factors have not been taken into consideration in the application of oilfield output prediction method. Time series transfer function method is firstly applied to build a multiple factor oilfield output prediction model considering dynamic relationship. Taking new wells output of one oilfield in Shengli Region for example, data preparation, model recognition, parameter estimation, diagnosis and checking and prediction process of transfer function model are illustrated according to its modeling work flow. The study result shows that this prediction model can remove irrelevant elements from multifarious factors and identify the main controlling factors on oilfield output prediction, and it can also observe the dynamic relationship of factors with non-synchronization and hysteresis. The precision is high and up to 98.4%.