Date of Graduation
Statler College of Engineering and Mineral Resources
Petroleum and Natural Gas Engineering
A method has been developed that locates and determines well-to-well hydraulic fracture interference (frac-hit) in shale plays using hard data. This method uses Artificial Neural Networks (ANN) with designated parameters and target outputs in conjunction with graphs of gas flowrate, tubing pressure, and cumulative gas prediction. The method was created to address the significant increase in frac-hit occurrences due to the infill wells being completed in shale plays. The production data of the well is first cleaned to eliminate outliers in the initial timeframe of the well and periods of no production so that the ANN model can be accurately trained. The model then predicts daily gas flowrate and is graphed against the wells cumulative gas and tubing pressures. The location of the section of variance from real data versus the predicted results will indicate a phenomenon at a given instant. This can indicate frac-hits through graphing a plot of a parent wells tubing pressure, gas flowrate, and cumulative gas production against a new child well at the location of variance that was observed in the model prediction.
The results of ANN training and test results accurately predicted cases where frac-hits are observed in the given field. This model also was able to predict the onset of the frac-hit which correlated to the same time that a new well was being completed in the area. This method allowed further research into the results since it was able to provide predicted flowrates at the time periods of frac-hits rather than only the time of the hit. Therefore, the ANN model was determined to be an adequate choice in analyzing frac-hits due to the sheer volume of information that can be taken away from the results.
Chamberlain, Dennis Wayne Jr., "Application of Machine Learning on Fracture Interference" (2018). Graduate Theses, Dissertations, and Problem Reports. 3698.