基于长短时记忆神经网络的潜油电泵故障预警

Fault early warning of electric submersible pump based on long short-term memory neural network

以潜油电泵机组的运行电流为主要判别依据,将长短时记忆神经网络应用于潜油电泵运行状态预测中,对于特征不明显的故障类型,利用潜油电泵井运行电压、运行电流、功率、油压、井口温度和瞬时流量数据预测下一时刻的电流值,并利用单分类支持向量机模型来预判潜油电泵机组的运行状态,从而实现潜油电泵的故障预警。最后,利用实际生产数据对模型进行验证。结果表明,所提方法预测准确度较高,可将报警时间提前1 h,实现故障的预警及诊断。

Using the operating current of the electric submersible pump unit as the main criterion,this paper proposes to apply the long and short--term memory neural network to the prediction of the operating state of the electric submersible pump.For the fault types with unobvious characteristics,the operating voltage,operating current,power,oil pressure,wellhead temperature and instantaneous flow data are combined to predict the current value at the next moment.The one-class support vector machine model is used to predict the operating status of the electric submersible pump unit to achieve the early warning of the fault of the electric submersible pump.Finally,Actual production data are used to verify the model.The results show that the prediction accuracy of the method proposed in this paper is high,and the alarm time can be advanced by one hour,and finally the early warning and diagnosis of fault can be achieved.