A Method of Intelligent Scheme Optimization for Non-Stop System Intermittent Production Units

摘要

Long-time and over-frequency system power-off of normal intermittent pumping (NIP) technology may exacerbate wax deposit, hydrodynamic level instability and even system restart failure. As a consequence, non-stop-system intermittent pumping (NSSIP) have been tentatively applied into some wells of low production in the oilfield of Daqing in China. Long time pump stop will reduce well production while frequent pumping will lead to void pumping and high energy consumption. Now, NSSIP production scheme is mostly derived from experience decision, the feed and fetch of pump is unbalanced, and it is doubtful whether the low energy consumption characteristic of NSSIP really works. Thus, it is urgent to carry on the research of intelligent scheme optimization for NSSIP. First, based on the A5 database platform of CNPC, the deliberate non-stop pumping scheme and data acquisition scheme for typical experimental wells will be formulated. Second, data of load, displacement and production rate of pumping wells by NSSIP will be measured in real time, then the pump efficiency vs dynamometer working condition will be calibrated accordingly. The CNN+SNN+HDG model is used for training, and the quantitative diagnosis model of insufficient liquid supply is established to realize the intelligent prediction of pump efficiency through dynamometer recognition. Third, the physical model of bottom hole flowing pressure buildup is established to realize the quantitative evaluation of reservoir liquid supply capacity. Data models for correlation analysis of the pumping time, dynamic liquid level, output rate, system efficiency and other parameters are established. At last, the key parameters that affect the pump balance of feed and fetch and economic benefit of a single well are screened out, and the all above analysis results and models are integrated into the ensemble decision tree model to optimize the most reasonable pumping time and frequency for NSSIP schemes. It is suggested that fluid supply capacity affect the intermittent scheme most. For low permeability and low production well, NSSIP could sustain pump fullness, increase system efficiency and save energy at the same time. Based on the ensemble method for intelligently production scheme optimization, NSSIP could increase system efficiency by 7.8% over normal pumping, and by 4.4% over normal intermittent pumping in average. What's more, NSSIP could reduce energy consumption by 22.14 kWh per day compared with intermittent in average. Non-stop-system intermittent pumping (NSSIP) may effectively prevent pumping system from failures caused by void pumping or long-time system stop. It could also greatly increase pumping and system efficiency and save more energy.