Data-driven analysis prompts completion modification delivering an additional 270,000 MCF peak month production
Marcellus Shale, Northeast Pennsylvania
A non-operating interest holder in a horizontal Marcellus project was concerned that their completions were under-designed and not producing at potential. The questions they had were as follows:
- How effective are these completions?
- How would these wells produce if they were completed and fracture stimulated more effectively?
- What are the primary drivers that control well production and recovery?
- How significant is geology and reservoir characteristics on well production?
The objective was to determine the increased production from the use of more effective completion design for minimal cost and without experimenting on actual wells. Experimenting on actual wells is problematic not only because of cost but also due to the lack of information required to normalize out the effects of well-to-well differences in reservoir characteristics.
Using the STRATAGEN well performance analysis service, WELLWORX, that uses multi-well, data-driven linear and neural network analysis techniques, develop a predictive model to forecast productivity. This data driven modeling process requires construction of an accurate forecasting model based on readily available well geology and completion information.
Geology and reservoir quality dominate Marcellus production however, fracture spacing and proppant volume ranked the highest in significance for controllable factors related to contact and conductivity affecting the level of gas production.
|Non Controllable Factors
||Number of clusters
Analysis supported increasing the number of frac stages from 10 to 25 and doubling the treatment volumes. The result was a production increase from nearly 200,000 MCF to 460,000 MCF in the peak month from the new new well completion design. This is a production increase of 2.5 over the previous completion and frac designs.
Other design change involved reducing the number of clusters to 4 and increasing the injection rate to 102 BPM, to increase the injection rate per cluster.
The WELLWORX model estimated production for new test well completion that was within 8% of the actual production rates.