PREDICTIVE MODELING OF PROPPED FRACTURE CONDUCTIVITY IN SHALE GAS RESERVOIRS

Loading...
Thumbnail Image

Date

2024-04-17

Authors

Yegeubayeva, Alin

Journal Title

Journal ISSN

Volume Title

Publisher

Nazarbayev University School of Mining and Geosciences

Abstract

Hydraulic fracturing is a well completions technique that induces a network of flow channels in a reservoir. These channels are characterized by fracture conductivity, a measure of how easily a liquid or gas flows through the fracture. Fracture conductivity is influenced by several variables including proppant size, proppant concentration, and hydraulic fracture characteristics. The purpose of this research is to present a unique process that incorporates machine learning neural networks in order to predict the fracture conductivity of multi-stage fractured horizontal well in shale gas reservoirs. To accurately predict fracture conductivity using fracture parameters such as width, height, length and orientation, a robust model is necessary. In this study, predictive ability of Multilayer Perceptron algorithm was used in forecasting fracture conductivity. The findings revealed the R-squared value of 0.82, which show a good correlation of these values with the previously conducted researches. Secondly, during validation of algorithm, CMG calculated fracture conductivity at 4.6 md.ft, although the machine learning model came closest at 4.43. Overall, values and other input variable parameters are near, indicating good model performance. Lastly, enhancing the cumulative gas output has been shown to be significantly aided by the process of fine-tuning fracture parameters, which are fracture length, height, and width, inside the CMG program. The obtained results can be used as references in the future examination of parameters that affect fracture conductivity.

Description

Keywords

Type of access: Restricted, prediction, artificial neural networks, fracture conductivity, multilayer perceptron neural network

Citation

Yegeubayeva, A. (2024). Predictive Modeling of Propped Fracture Conductivity in Shale Gas Reservoirs. Nazarbayev University School of Mining and Geosciences