油田生产动态预警模型研究

Research on Dynamic Early Warning Model of Oilfield Production

随着大庆油田的开发进入中后期,油田生产中的各个专业都积累了大量宝贵的专业数据,这些数据不仅反映了石油开采的发展轨迹,并且记录了油田开发过程中出现的产量异常下降情况,以及针对异常所采取的增产措施。文中建立了基于支持向量机的动态预警模型,通过寻找历史生产数据中的变化规律,找到生产异常报警形成模式,通过对油田开发动态数据的实时监测,可以达到提前判断出产量的异常下降,以便及早采取针对性措施,确保原油生产的良性运行。通过对大庆油田第八采油厂的历史样本进行了训练和验证,证明模型对于油田生产中发生的异常情况具有较高的的预测准确性。实验证明模型的有效性,并且提出对模型进一步改进的相关方法。

As the development of Daqing oil field in the middle and late stage of oil field production, various profession has accumulated a large amount of valuable data,these data not only reflect the petrolanm exploitation development path,but also record the abnormal production declined in the process of oilfield development,and the take measures to increase production of the abnormal. Based on the support vector machine, dynamic early warning model is established. By finding the historical production data of changes in production, find abnormal alarm formation mode. Based on the real-time monitoring for oil field development data, can judge in advance the abnormal production decline, and take targeted measures to ensure the benign operation of crude oil production. The eighth of Daqing oil field plant history samples are trained and validated,it is proved that the model for oilfield production abnormal condition has a high prediction accuracy. Experimental results show that the model is effective, and the model is further imnroved correlation method.