Projection of Logging While Drilling Data at the Bit by Implementing Supervised Machine Learning Algorithm

Sharma, Ashutosh; Burak, Tunc; Nygaard, Runar; Hellvik, Svein; Hoel, Espen; Welmer, Morten

doi:10.2118/213070-ms

Cited by 3 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The workflow is visually represented in Figure 3, which illustrates the comprehensive steps involved: data collection, data processing, model training and testing, followed by efficient model selection. The comparative analysis of all regression models is performed based on the R-squared and error metrics as mentioned in [19][20][21] for efficient model selection.…”

Section: Methodsmentioning

confidence: 99%

Predicting Gas Separation Efficiency of a Downhole Separator Using Machine Learning

Sharma,

Osorio Ojeda,

Yuan

et al. 2024

Energies

Self Cite

View full text Add to dashboard Cite

Artificial lift systems, such as electrical submersible pumps and sucker rod pumps, frequently encounter operational challenges due to high gas–oil ratios, leading to premature tool failure and increased downtime. Effective upstream gas separation is critical to maintain continuous operation. This study aims to predict the efficiency of downhole gas separator using machine learning models trained on data from a centrifugal separator and tested on data from a gravity separator (blind test). A comprehensive experimental setup included a multiphase flow system with horizontal (31 ft. (9.4 m)) and vertical (27 ft. (8.2 m)) sections to facilitate the tests. Seven regression models—multilinear regression, random forest, support vector machine, ridge, lasso, k-nearest neighbor, and XGBoost—were evaluated using performance metrics like RMSE, MAPE, and R-squared. In-depth exploratory data analysis and data preprocessing identified inlet liquid and gas volume flows as key predictors for gas volume flow per minute at the outlet (GVFO). Among the models, random forest was most effective, exhibiting an R-squared of 96% and an RMSE of 112. This model, followed by KNN, showed great promise in accurately predicting gas separation efficiency, aided by rigorous hyperparameter tuning and cross-validation to prevent overfitting. This research offers a robust machine learning workflow for predicting gas separation efficiency across different types of downhole gas separators, providing valuable insights for optimizing the performance of artificial lift systems.

show abstract

Section: Methodsmentioning

confidence: 99%

Predicting Gas Separation Efficiency of a Downhole Separator Using Machine Learning

Sharma,

Osorio Ojeda,

Yuan

et al. 2024

Energies

Self Cite

View full text Add to dashboard Cite

show abstract

Evaluating PDC bit-rock interaction models to investigate torsional vibrations in geothermal drilling

Sharma,

Al Dushaishi,

Nygaard

2024

Geothermics

View full text Add to dashboard Cite

Projecting Petrophysical Logs at the Bit through Multi-Well Data Analysis with Machine Learning

Sharma,

Burak,

Nygaard

et al. 2023

Day 3 Thu, September 07, 2023

View full text Add to dashboard Cite

The vertical distance from logging while drilling (LWD) sensors to the bit is often more than 30m (98 ft), which leads to difficulty in performing real-time comparison of LWD and drilling data. This study aims to predict the petrophysical data at the drill bit with the objective of determining the best supervised machine learning algorithm to incorporate to reduce the sensor offset problem. The bulk density and porosity logs are predicted at the bit in this paper using petrophysical and drilling parameters. The results of the model will be used to perform lithology identification in real-time that can be used in real-time drilling analysis. To predict the bulk density and porosity logs at the bit, data from four different wells located in the Norwegian continental shelf in the North Sea was used as a training dataset. The data from a fifth well from the same field was used as a validation dataset. The prediction was based on input variables of the Gamma ray (GR) log data recorded close to the bit, along with other drilling parameters measured at the bit using Measurement while Drilling (MWD) sensors. The five regression models used for prediction and comparative analysis were: Multi-linear regression (MLR), K-nearest neighbor (KNN) regression, Random forest regression (RFR), Support vector machine (SVM) regression and Artificial neural network (ANN). All five models were tested for their accuracy in predicting porosity and bulk density, and it was determined that the KNN model was more effective for predicting both porosity and bulk density. The coefficient of determination (R2) value for the KNN model for porosity and bulk density predictions were 86% and 74% respectively with the least mean square error (MSE) calculated on the blind dataset (data from a well not included in model training). SVM was found to be the least effective model for predicting both porosity and bulk density, as it had the highest MSE value. Prediction of porosity and bulk density logs at the bit using multiple machine learning techniques to eliminate the sensor offset problem have not been performed extensively in the past. The developed machine learning model will improve real-time drilling analysis.

show abstract

Projection of Logging While Drilling Data at the Bit by Implementing Supervised Machine Learning Algorithm

Cited by 3 publications

References 11 publications

Predicting Gas Separation Efficiency of a Downhole Separator Using Machine Learning

Predicting Gas Separation Efficiency of a Downhole Separator Using Machine Learning

Evaluating PDC bit-rock interaction models to investigate torsional vibrations in geothermal drilling

Projecting Petrophysical Logs at the Bit through Multi-Well Data Analysis with Machine Learning

Contact Info

Product

Resources

About