基于神经网络的游梁式抽油机示功图软测量研究

发布时间：2018-11-02 16:52

【摘要】：游梁式抽油机示功图是悬点载荷和悬点位移的关系图,示功图可以诊断油井的工作状况。至今没有一种有效方法实时测量示功图,主要原因是载荷传感器容易老化及使用寿命短。针对游梁式抽油机的示功图实时测量难问题,总结了现有的示功图间接测量方法后,提出了基于神经网络的示功图软测量方法。该方法的核心是实时测量游梁式抽油机的相关物理量,如电机功率、游梁摆动角度,利用神经网络的非线性逼近功能,预测抽油机悬点载荷。通过数据通信手段,把相关的现场数据传到数据中心,数据中心的示功图测量单元由神经网络预测抽油机悬点载荷,加上悬点位移,得到抽油机示功图。实验仿真表明基于神经网络的示功图软测量方法是可行的,有推广应用价值。
[Abstract]:The indicator diagram of beam pumping unit is the relation diagram between suspension load and suspension displacement, which can diagnose the working condition of oil well. Up to now, there is no effective method to measure indicator diagram in real time, mainly due to the easy aging of load sensor and short service life. In order to solve the problem of real-time measurement of indicator diagram of beam pumping unit, this paper summarizes the existing indirect measurement methods of indicator diagram, and puts forward a soft measurement method of indicator diagram based on neural network. The core of this method is to measure the relative physical parameters of beam pumping unit in real time, such as motor power, swing angle of walking beam, and forecast the suspended point load of pumping unit by using the nonlinear approximation function of neural network. Through the means of data communication, the relevant field data are transmitted to the data center. The measuring unit of the indicator diagram of the data center predicts the suspension load and the displacement of the suspension point of the pumping unit by the neural network, and obtains the indicator power chart of the pumping unit. The simulation results show that the soft sensor method based on neural network is feasible and worth popularizing.
【作者单位】：沈阳理工大学信息科学与工程学院;
【基金】：辽宁省自然科学基金(20072049)
【分类号】：TE933.1

【参考文献】