... training, and result interpretation. The focus of paper is on discussing the effectiveness of machine learning algorithms in predicting well’s production performance, including the ML model accuracy, interpretability, and prediction uncertainty quantification. The workflow is successfully applied...

... and water production rate. In addition, complex geological conditions often induce downhole problems, which fluctuate or even suspend production. Therefore, it is hard to accurately predict the production potential for new wells and analyze the key factors of production. This study aims to apply artificial...

... from corroborating to other methods. In our approach, Machine learning models are developed to predict EUR for comparison against other methodologies to assess alignment of predictions. Once confidence is established on production forecasts and EURs, statistical tools can be used to derive some key...

... Abstract Predicting production behavior plays an important role in oil and gas production and aids engineers to perform field management. However, this can be challenging in heterogeneous reservoirs using traditional models that require expensive computational time and various types...

... Abstract Accurate pore pressure and stress estimation are extremely important for safe and efficient drilling. The objective of this study is to establish robust models for an unconventional shale play to predict pore pressure and stress magnitudes from elastic properties derived from seismic...

... Abstract Application of a Convolutional neural network (CNN) is presented for an unconventional reservoir to simultaneously predict elastic, mineral volumes, geomechanical and reservoir properties from seismic and well data. In the Barnett Shale, success requires identifying brittle...

... relationship (IPR). Transient PI as the forecasting variable normalizes both surface pressure effects and takes phase behavior into account, reducing noise. The PI-based forecasting method is used to predict future IPRs and, as a result, oil, water, and gas rates for any bottom hole pressure or operating...

... units (DSUs) containing parent wells is hindered by the large degree of uncertainty around child well performance. Infill child wells have been observed with up to a 40% degradation from their unbounded parents in the Midland Basin. We present an AI enabled workflow to predict infill well performance...

.... The developed data were used to train different Artificial Neural Network (ANN) algorithms. ANN algorithms showed high capabilities to develop a surrogate model of the high-fidelity model. The ANN models provided robust predictions on the monthly production level for oil recovery factors. The robustness...

... Abstract Flow regime changes are important to capture and predict when estimating the future production from tight oil wells. Due to the nature of hydrofracture completions and reservoir heterogeneity, flow regimes are difficult to predict before they are observed. Also, while some wells...

... tree-based algorithms (random forest and gradient boosting), feed-forward neural networks, and recurrent neural networks (simple RNN, LSTM, and GRU), were implemented and trained using historic data of a number of lateral wells that liquid loaded. The best performing model was selected to predict...

.... The authors have developed an innovative approach to predict short-term cumulative oil production and long-term oil EUR using hourly flowback data, theory-based calculations, and empirical correlations. Unconventional wells produce in the linear flow regime in early time. During this period, the production...

... Abstract The novel Fractional Dimension RTA (FD-RTA) was applied to a recent multi-pad development campaign in the Haynesville to test the suitability of the model to represent well performance and its ability to predict production with a very limited data history calibration. Additionally...

... provide poor predictions for the near-critical/volatile fluids encountered in many unconventional reservoirs ( e.g. , highly volatile oils, retrograde gas condensates, and wet gases). In this work we present a series of customized PVT correlations to address these deficiencies for unconventional...

... erosion profile. With the help of offset-well fiber data and mechanistic models, significant impacts of inter-stage stress shadow effects were observed on far-field cluster efficiency. Using physics-based models, observed trends were captured with a maximum of 10% error in the prediction...

... Abstract This study demonstrates that machine learning models trained on manually performed petrophysical analyses (n = 1542) can generate predictions with accuracy that is sufficient to make business decisions. We evaluated multiple machine learning algorithms to establish a benchmark...

... learning and deep neural network to estimate and predict the geomechanical properties of the Permian Basin. The log-derived prediction algorithm includes (a) Single-Well prediction, 75% of log data of a single well is used as a specimen for training the Bi-LSTM, and the rest 25% of data of the same well...

... not be so straightforward and additional information about the pumping system may be needed to estimate the BHP. The goal of this work is to build a Machine Learning data-driven model that can predict the BHP for multi-fractured horizontal wells of the Vaca Muerta Formation in Argentina. Input variables...

... extrapolation ability and require sufficient training data, where training an under-determined neural network predictive model with limited data can result in overfitting and poor prediction performance. Unlike statistical models, physics-based models impose causal relations that can provide reliable...

... and fractures network scale. This work presents an empirical correlation for water-hydrocarbon relative permeability prediction in fractured reservoirs, as a function of stress state and capillary number variations usually relevant in the stimulated reservoir volume – SRV after hydraulic fracturing...