Abstract
Production of coal seam gas (CSG) wells relies on continuous extraction of water to depressurise coal below the desorption pressure. Progressive cavity pumps (PCPs) are common artificial lift systems for dewatering CSG wells in Queensland. These pumps are susceptible to failure primarily due to the interference of gas and fines, which are naturally available in CSG wells. The pump failure leads to a significant downtime which requires an expensive well intervention to replace the pump. This paper aims to develop a procedure to detect incipient pump failure to better plan maintenance and optimise well performance. The proposed procedure is developed using a time-series statistical process control (SPC) approach. This approach was applied to the data provided by gas companies and tested against failure data of PCPs from 21 vertical CSG wells in Queensland over a one-year production period. These data include the measurements of pump torque, speed, speed setpoint and efficiency, water flowrates, and wellhead pressure. The results show promise, allowing detection and prediction of more than 90% of the PCP failure events, within 14 hours to 111 days in advance of the recorded failure date. These findings enable gas operators to arrange the maintenance and replacement of the pumps, reducing the downtime and thereby increasing the production of gas wells.
