Abstract
Our understanding of the complexities of the flow mechanism in Shale plays has not kept up with our industry’s interest in these prolific and hydrocarbon rich formations. Furthermore, massive multi-cluster, multi-stage hydraulic fractures, that have proven to be essential for economic recovery from Shale plays, have significantly increased the complexity of the flow behavior and consequently have made the modeling efforts more challenging.
In this paper, the application of a recently developed AI (Artificial Intelligence)-based reservoir modeling approach on Marcellus Shale is presented. In this approach, data mining and pattern recognition techniques were used to initiate modeling of the hydrocarbon production (dray gas and condensate) from Marcellus Shale. Instead of imposing our understanding of flow and transport in shale gas media, which is a very complex and non-linear system, we allow the production history, reservoir characteristics, and hydraulic fracturing data and operational constraint to force their will on our model and determine its behavior.
In this work, the full-field history matching process was performed on a Marcellus shale asset including 135 wells with multiple pads and different landing targets. The full field AI-based Marcellus Shale model then used for forecasting the future well/reservoir performance to assist in planning field development strategies. The goodness of match quality is self-evident, thereby validating this modeling approach. Nevertheless, to examine the model validity in the forecasting mode, the field data was partially matched and then attempted forecasting. Taking validation one-step further, the production performance of a recently drilled well, which was completely blind to the model (was not involved during training and initial validation), was predicted and compared with actual field measurement.
Furthermore, sensitivity and economic analysis are performed in order to identify the impact of different reservoir descriptions (e.g. different reservoir characteristics, stimulation and completion factors) and rank the impact of above-mentioned parameters on the Net Present Value (NPV) of investing on gas wells producing from Marcellus Shale.
