Enhancing well deliverability and reducing work-over cost using big data predictive analytics
This project aims to leverage existing data in the CSG industry for enhanced production efficiency of coal seam gas wells and reduce the significant operating cost associated with dewatering of these wells through the application of advanced, predictive analytics techniques, including machine learning, in combination with analytical modelling.
The aim is to reduce expenditure on down-hole pumps (with probable application to other rotating equipment). Extraction of coal seam gas (CSG) is associated with producing a significant amount of water. This water needs to be removed to reduce the flowing bottom-hole pressure (FBHP) to allow gas to flow into the wellbore. Progressive cavity pumps (PCPs), commonly used for dewatering, are prone to failure due to gas interference and fines production resulting in high and unpredictable downtime and maintenance costs. Currently the typical pump life for PCPs is 18-24 months and failures require the well to be shut-in, and maintenance crews deployed to recover and replace the pump, with attendant economic, safety and environmental costs. The cost of workover required for replacement of a downhole pump is reported to be $150,000-$200,000. With over 8,000 producing CSG wells in Queensland, the expenditure committed to replacement of the pumps runs to several hundred million dollars annually.
Extending mean time between failures from 18 to 24 months, based on the above estimates, would be a saving per well of around $25,000 pa. For an active, CSG well stock of say 4,000 (it is currently higher than this), this represents an OPEX saving potential of $100 million pa. There are additional benefits from reduction of deferred production and operational efficiencies as work-overs are moved from un-planned to planned maintenance.
Importantly, flowing bottom hole pressure (FBHP) is a key metric for optimising well performance and enhancement of production. Downhole pressure gauges are expensive, unreliable and complex to install. Furthermore, calibration and maintenance of these down-hole gauges is associated with an expensive and complex well intervention to replace the gauge. Therefore, developing a low-cost method of predicting the FBHP for wells without gauges remains a key research challenge.
This project aims primarily to reduce maintenance costs, improve down-time predictability and optimise pump running to enhance production using advanced, predictive data analytics of historical temporal and spatial data provided by the industry. A secondary (but significant) benefit is expected to be improved prediction of FBHP from pump data.
Besides economic benefits of this project, reducing onsite maintenance lessens access requirements for rigs maintenance activities. This will decrease the disruption to rural activities where the wells are located, ease local traffic, and reduce the environmental impacts associated with well workovers.
This project has two primary objectives:
- Detect incipient faults and predict failure to better plan maintenance and optimise settings and thereby to improve pumps or mean time between failures
- Predict the FBHP from the production and pump data using machine learning
|