【摘要】 针对大庆油田某试验井潜油电机使用情况,给出了在工频和变频条件下应用多层反馈神经网络RMNN(recurrentmultilayer neural network)实现电机转速辨识的方案,以便对潜油电机动态运行进行实时监测。鉴于潜油电机独特的高温工作环境,给出了无速度传感器条件下辨识潜油电机动态转速的RMNN模型。通过在井上对潜油电机定子电流、电压等参数的采集,着重研究潜油电机启动、稳定运行以及电源频率变化、负载变化对辨识效果的影响。研究结果表明,基于RMNN模型的潜油电机动态运行的速度辨识误差精度为0.4%,可满足试验井潜油电机转速辨识的需要。更多还原

【Abstract】 Based on an electric submersible pumping system in Daqing oil field,the project is psesented to identify rotating speed of pumping motor with and without converter by recurrent multilayer neural network(RMNN) and to monitor dynamic status of the electric submersible pumping system at the downhole.Considering the special condition of high temperature,the RMNN models of dynamic rotating speed of sensorless electric submersible pumping system are given.By sampling motor stator current and power supply voltage above the shaft,the influences of starting current and steady state operation,variations of power supply frequency and load,impacts of downhole high temperature environment on identification effect are investigated.The research results show that the speed identification precision of dynamic state of the motor has reached 0.4% by the proposed RMNN model,which can satisfy the requirements of electric submersible pumping system.更多还原