Publications by authors named "Shaoyang Geng"
Article Synopsis
- Accurately predicting the Z-factor of natural gas is essential for assessing gas reserves and transport, but traditional machine learning methods often struggle with different gas conditions and components.
- To improve prediction accuracy, the authors propose a new framework that uses signal decomposition techniques, allowing them to break down the Z-factor into multiple components before applying machine learning algorithms.
- Their approach results in significantly better predictive performance, achieving high correlation coefficients and low error rates across diverse datasets, indicating it can effectively handle varying natural gas conditions.
The field production profile over the yearly horizon is planned for a balance between economy, security, and sustainability of energy. An optimal drilling schedule is required to achieve the planned production profile with minimized drilling frequency and summation. In this study, we treat each possible production process of each well as a dependent time series and the basic unit.
View Article and Find Full Text PDF