油田产量灰色-神经网络混合预测模型研究

Research on Mixed Forecast Model of Oilfield Production Output Based on Grey System and Neural Network

提出将灰色预测与神经网络预测结合起来,建立灰色-神经网络混合预测模型,对油田产量进行预测。此方法是指以灰色预测为主,采用神经网络预测对其补偿的预测方法。模型的工作过程是,首先建立既考虑油田产量自身变化规律又考虑其影响因素变化对产量的影响的灰色预测模型,然后利用神经网络识别灰色预测所得的预测值和实际产量值之间的未知关系,去修正灰色预测模型所得的预测值。实例表明,这种方法不仅能够在很大程度上提高灰色模型预测的精度,而且扩展了灰色模型预测的应用范围。

A grey predicting model was combined with neural network to establish a mixed predicting model for oilfield production prediction.Grey prediction was dominated in the method,the neural network was used for compensating the prediction.Its operation procedure included that the grey predicting model was established,in which the rules of oilfield production variation and effect of influencial factor variation in production output were considered,the neural network was deployed to recognize the unknown the correlation between the grey predicted value and actual production output,calibrate the predicted value obtained with grey predicting model.Application shows that the method can be deployed for greatly improving the precision of grey predicting model and expanding its application range.