Developing a New Model for Drilling Rate of Penetration Prediction Using Convolutional Neural Network

Matinkia, Morteza; Sheykhinasab, Amirhossein; Shojaei, Soroush; Kand, Ali Vojdani Tazeh; Elmi, Arad; Bajolvand, Mahdi; Mehrad, Mohammad

doi:10.1007/s13369-022-06765-x

Cited by 17 publications

(6 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, the lower and upper quartiles of each log will be first determined to compute an interquartile range (IQR) and then a lower inner fence (LIF) and a upper inner fence (UIF) will be figured out, and the last two fences will form a range to detect the outliers varying outside. A simplified equation set describing Tukey's method is mentioned below (Anemangely et al., 2017; Matinkia et al., 2022a, 2022b, 2022c; Mehrad et al., 2022; Tukey, 1975):

\begin{equation}\left\{ { \def\eqcellsep{&}\begin{array}{@{}*{1}{l}@{}} {{\rm{IQR}} = 1.5 \times ({Q}_3 - {Q}_1)}\\ {{\rm{LIF}} = {Q}_1 - {\rm{IQR}}}\\ {{\rm{UIF}} = {Q}_3 + {\rm{IQR}}} \end{array} } \right.,\end{equation}

where Q 1 and Q 3 are the lower and the upper quartiles, respectively.…”

Section: Methodsmentioning

confidence: 99%

“…Since the variation of a noisy sample normally will exceed a rational observing range, such type of sample also can be regarded as an outlier. Currently, there exist a great deal of outlier-detecting methods, while among them Tukey's method is outstanding because of its more robust nature and a relatively easier implementation (Anemangely et al, 2017;Matinkia et al, 2022aMatinkia et al, , 2022bMatinkia et al, , 2022cMehrad et al, 2022;Tukey, 1975). Hence, it is selected as a remover for the outliers in this study.…”

Section: Preprocessmentioning

confidence: 99%

“…Specifically, the lower and upper quartiles of each log will be first determined to compute an interquartile range (IQR) and then a lower inner fence (LIF) and a upper inner fence (UIF) will be figured out, and the last two fences will form a range to detect the outliers varying outside. A simplified equation set describing Tukey's method is mentioned below (Anemangely et al, 2017;Matinkia et al, 2022aMatinkia et al, , 2022bMatinkia et al, , 2022cMehrad et al, 2022;Tukey, 1975):…”

Section: Preprocessmentioning

confidence: 99%

“…Based on the observations of each crossplot within the lower map and each kernel density along the diagonal, it is understood that no matter what type of subplot is applied, the distribution of any classifying target always appears a large overlap with the other and then there exist no satisfactory lithofacies discriminations. The causation of this undesirable classification can be attributed to the similar responses of three types of lithofacies on some exampled logs, while for a deep reason the explanation can be a colinearity issue of the used logs (Matinkia et al, 2022a(Matinkia et al, , 2022b(Matinkia et al, , 2022cMehrad et al, 2022). Specifically, as some logs such as RXO, RI and RT will demonstrate the similar geophysical properties of the measuring reservoirs, the data of them undoubtedly will vary similarly, which accordingly in F I G U R E 4 Pairwise plots used to illustrate capability of employed logs on the identification of lithofacies.…”

Section: Data Source Problem and Experimental Designmentioning

confidence: 99%

See 3 more Smart Citations

Lithofacies prediction driven by logging‐based Bayesian‐optimized ensemble learning: A case study of lacustrine carbonate reservoirs

Zhang

Wang³

et al. 2022

Geophysical Prospecting

View full text Add to dashboard Cite

Although lithofacies routinely is featured by distinct logging responses from each other, many types of lithofacies in practical cases show similar measuring characteristics on logs, and then to achieve a desirable solution from logging‐based lithofacies prediction actually is challengeable. Since the mathematical essence of lithofacies prediction can be explained as an issue of pattern recognition, a light gradient boosting machine, a state‐of‐the‐art ensemble learning, specifically developed to address supervised classification, could be a potential solver. Nonetheless, due to an incompatibility of inherent exclusive feature bundling algorithm for logs and usage of a great deal of hyper‐parameters, a raw light gradient boosting machine might not be suitable or functional to predict lithofacies. Thus, continuous restricted Boltzmann machine and Bayesian optimization, respectively, employed to process original logging data and optimize hyper‐parameters, are adopted as technical assistants for the light gradient boosting machine, and accordingly a new ensemble learning‐based predictor called continuous restricted Boltzmann machine – Bayesian optimization – the light gradient boosting machine, is proposed for lithofacies. To validate classifying capability and robust nature of new predictor, three experiments are designed purposefully based on the application of a dataset collected from the wells located within pre‐salt lacustrine carbonate reservoirs of the Santos Basin. Simultaneously, to highlight validating effect, other three sophisticated classifiers are introduced as competitors in all experiments, including supper vector machine, random forest and extreme gradient boosting. According to the analysis and comparison of all experimental results, overall four points are verified: (1) the integration of continuous restricted Boltzmann machine and Bayesian optimization indeed is beneficial for the light gradient boosting machine in data preprocess and modelling stage; (2) the light gradient boosting machine cored predictor shows more capability of producing reliable predicted lithofacies information compared to other three competitors; (3) training more learning samples is a simple and effective approach to enhance computing capability of any validated predictor, and under this circumstance the light gradient boosting machine cored predictor still performs relatively better; (4) the light gradient boosting machine cored predictor is characterized with a better robustness as its predicting outcomes, even derived from a sparse‐condition dataset, are also proved more qualified. Consequently, the new proposed predictor is demonstrated as a high‐efficient and robust solver for the prediction of lithofacies and deserves a widespread application in the geological, geophysical and petrophysical research.

show abstract

\begin{equation}\left\{ { \def\eqcellsep{&}\begin{array}{@{}*{1}{l}@{}} {{\rm{IQR}} = 1.5 \times ({Q}_3 - {Q}_1)}\\ {{\rm{LIF}} = {Q}_1 - {\rm{IQR}}}\\ {{\rm{UIF}} = {Q}_3 + {\rm{IQR}}} \end{array} } \right.,\end{equation}

where Q 1 and Q 3 are the lower and the upper quartiles, respectively.…”

Section: Methodsmentioning

confidence: 99%

Section: Preprocessmentioning

confidence: 99%

“…Specifically, the lower and upper quartiles of each log will be first determined to compute an interquartile range (IQR) and then a lower inner fence (LIF) and a upper inner fence (UIF) will be figured out, and the last two fences will form a range to detect the outliers varying outside. A simplified equation set describing Tukey's method is mentioned below (Anemangely et al, 2017;Matinkia et al, 2022aMatinkia et al, , 2022bMatinkia et al, , 2022cMehrad et al, 2022;Tukey, 1975):…”

Section: Preprocessmentioning

confidence: 99%

Section: Data Source Problem and Experimental Designmentioning

confidence: 99%

See 2 more Smart Citations

Lithofacies prediction driven by logging‐based Bayesian‐optimized ensemble learning: A case study of lacustrine carbonate reservoirs

Zhang

Wang³

et al. 2022

Geophysical Prospecting

View full text Add to dashboard Cite

show abstract

“…Convolution neural network is proposed to process image information. Its principle is to convolve an image with a convolution kernel [18]. For a point on the image, let the origin of the convolution kernel coincide with the point.…”

Section: Machine Learning and Deep Learning Algorithmsmentioning

confidence: 99%

A comparative study of deep learning methods for drilling performance prediction

Jiao¹,

Liu²,

Cao³

et al. 2023

International Conference on Computer Graphics, Artificial Intelligence, and Data Processing (ICCAID 2022)

View full text Add to dashboard Cite

Rate of penetration is a key parameter to describe the efficiency of drilling through the formations and it has always been a measure of drilling efficiency. Previous research has demonstrated that Machine learning can be applied to improve the prediction of ROP from conventional correlation approaches. More recently, deep learning models have also been used toward that purpose. In this research, the typical used machine learning methods and deep learning methods are tested and compared in terms of ROP prediction. An open-source drilling dataset is used for single well and multiple well modeling and testing. The results demonstrate that deep learning models have more general and accurate results, and LSTM shows solid prediction performance for both single well and multiple well cases. The accurate prediction model for ROP can be further applied for the planning phase and real-time operation optimization.

show abstract

Automated neural network optimization for data-driven predictive models: an application to ROP in drilling

Khebouri,

Rechak,

Boulham

et al. 2024

Soft Comput

View full text Add to dashboard Cite

Developing a New Model for Drilling Rate of Penetration Prediction Using Convolutional Neural Network

Cited by 17 publications

References 49 publications

Lithofacies prediction driven by logging‐based Bayesian‐optimized ensemble learning: A case study of lacustrine carbonate reservoirs

Lithofacies prediction driven by logging‐based Bayesian‐optimized ensemble learning: A case study of lacustrine carbonate reservoirs

A comparative study of deep learning methods for drilling performance prediction

Automated neural network optimization for data-driven predictive models: an application to ROP in drilling

Contact Info

Product

Resources

About