摘要
振动信号能够全面反应电潜柱塞泵的运行工况,在电潜柱塞泵的故障诊断中,针对不同工况下的振动信号进行分析尤为重要。提出了一种基于振动信号分析的电潜柱塞泵故障诊断方法,采用改进的隐层神经元数变动的神经网络系统,根据不同情况选择不同节点数,使其达到提高诊断精度及缩短诊断周期的双重要求。对电潜柱塞泵正常运行以及动子不平衡、动子机械磨损运行时的振动信号进行分析研究,利用小波包提取振动信号的能量特征,利用神经网络识别故障。现场试验结果表明,网络训练后实际输出达到允许误差范围,网络测试结果也能与实际状态相对应。该方法能够对电潜柱塞泵进行有效、准确的故障诊断。
The vibration signal can fully reflect the operating conditions of the electric submersible plunger pump.It is very important to analyze the vibration signal under different working conditions in the fault diagnosis of the electric submersible plunger pump.In this paper,a fault diagnosis method of the electric submersible plunger pump based on vibration signal analysis is proposed.The neural network system with improved neurons in the hidden layer was designed and used to choose different number of nodes according to different situations.The dual requirements of improving diagnostic accuracy and shortening diagnosis cycle.The vibration signals of the electric submersible plunger pump normal operation as well as the movable imbalance and mechanical wear are analyzed and researched.Using wavelet packet to extract energy characteristics of vibration signal,neural networks is used to identify faults.Field test results show,the actual output of network training reaches the allowable error range.After the network test results can also correspond to the actual state.The method can effectively and accurately diagnose the electric submersible plunger pump.
作者
杜伟山
DU Weishan(Production Engineering Research Institute,Daqing Oilfield Company Limi ted,Daqing 163712,China)
出处
《石油矿场机械》
2018年第4期8-13,共6页
Oil Field Equipment
基金
中国石油科学研究与技术开发项目"采油
井下作业和储层改造技术研究与应用"(2016B-4104)
关键词
电潜柱塞泵
振动信号
神经网络
故障诊断
electric submersible plunger pump
vibration signals
neural network
fault diagnose
