State Monitoring And Identification For Electrical Submersible Pump Unit

With the rapid increase of world oil demand and the reducing of easy extraction oil day by day, Electrical Submersible Pump (ESP), as an important oil extraction equipment, plays an important role in secondary or tertiary oil recovery. Normally, ESP works underground of deeper than 1km, the high temperature, high pressure and vibration of formidable working environment effects reliability and high performance of ESP. Research on ESP state monitoring has important significance and practical value to improve the oil extraction technology with ESP and to develop ESP products with self-owned intellectual property.This dissertation is supported by the cooperative project of Daqing Petroleum Administrative Bureau, based on the ESPs that produced in Daqing Oilfield Powerlift Pump Industry Co., Ltd, the multi-sensor down-hole tool (MDT) for ESP and its Engineering Realization are carried out. Considering the project requirement and characters of engineering, comparatively theory and experiment work on multi signals sampling, transmission, modulation and demodulation of pressure, temperature, vibration, etc. are developed. The displacement identification of ESP and reliability analysis of MDT are also finished.Reference to research status of MDT and by analyzing the special working environment of the ESP, the multi-signal are carried with motor power line. Signal extraction and conditioning about status parameters of temperature, pressure, vibration and current leakage is presented, the electric power supply to the MDT is put forward also. Tests and analysis of unbalance voltage and higher harmonic generated by the frequency converter on no and full load, transmission of multi-sensor signals with power line carrier, modulation, demodulation and conditioning technology are carried out.State monitoring and fault diagnosis of the down-hole sensors are carried out. Based on the state monitoring method of RBF networks series forecasting, the time series forecasting model of sensor is built. an initial center choosing method is used to adjust the radius of the initial clustering center automatically with sample series. According to the changing trend of the sensor output, the sensor fault is identified, experiments on temperature sensor verified the identification results. The pressure sensor output is calibrated with temperature signal by software.Displacement of the ESP is the key parameter to ESP working status. Considering the composition of well fluid and the structure of the flow gauge, it's difficult to measure the displacement of the oil well directly. Based on the relation between the ESPU displacement and the pumping lift, the dynamic model of displacement to intake pressure, discharge pressure and the working current of the electrical submersible motor is built with neural networks. The displacement of ESP is identified with recurrent neural networks, the networks learning algorithm is also improved, value ranges of amplification factor and momentum are given. Experiment results verifies the identification of ESP displacement.Reliability analysis of the MDT. On analyzing the reliability of ESP, fuzzy reliability analyzing method is adopt to analyze the reliability of the MDT by backward interference model, so as to master the distribution law of the lifetime if the MDT and to improve the basic component of high fault probability.Experiment tests of MDT in lab and in testing oil field are carried out. Temperature sensors are calibrated and drifts of pressure sensors, vibration sensors and current leakage sensors to temperature are compensated, preliminary study of state monitoring and fault diagnosis of ESP based on MDT is executed. Experiment results show that the developed MDT can monitor the ESP working status in real time.Research on key technology of MDT for ESP and its engineering is the base of producing MDT with self-owned intellectual property.