Support-vector-regression machine technology for total organic carbon content prediction from wireline logs in organic shale: A comparative study

Tan, Maojin; Song, Xiaodong; Yang, Xuan; Wu, Qingzhao

doi:10.1016/j.jngse.2015.07.008

Cited by 128 publications

(40 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fortunately, such limitations nowadays could be overcome by machine learning approaches. The Machine learning approach, such as Neural Network method has been applied for shale TOC estimation (Alizadeh et al 2012, Khoshnoodkia et al 2011, Tan et al 2015. The method utilizes laboratory measurement of shale organic matter, and combines the measured TOC with well logs, followed by the data calibration for machine learning.…”

Section: Introductionmentioning

confidence: 99%

“…The method utilizes laboratory measurement of shale organic matter, and combines the measured TOC with well logs, followed by the data calibration for machine learning. Recently, Tan et al (2015) used Support Vector Regression Machine approach to estimate TOC in various organic shales using a variety of Kernel Functions. However, Neural Network ignores the generalization and always results in overfitting.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new method for TOC estimation in tight shale gas reservoirs

Rezaee

Wang

et al. 2017

International Journal of Coal Geology

View full text Add to dashboard Cite

Total organic carbon (TOC) estimation is significantly crucial for shale reservoir characterization. Traditional TOC estimation methods (such as Passey and Schmoker method) do not provide accurate TOC predictions in shale gas reservoirs especially for the self-generated and self-stored reservoirs. This study proposes, for the first time, a new TOC prediction method based on Gaussian Process Regression (GPR) bridging geostatistics and machine learning technique. The method utilizes a non-parametric regression approach in shale TOC predictions, and not only provides the expert solutions in high-dimension processing, small samples and non-linear problems, but also has a better adaptation and generalization ability compared with other machine learning methods. The approach accounts for all the well logging attributes and chooses the relevant logs to build TOC estimation model, and 7 different kernel functions and 5 attributes groups are analyzed to get the optimized hyperparameters in practice. Application of the developed model to two shale gas reservoirs showed that the model predicted TOC matched well with that from the laboratory measurements. The proposed model based on GPR method provides an accurate way for the TOC prediction in the tight shale gas reservoirs.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A new method for TOC estimation in tight shale gas reservoirs

Rezaee

Wang

et al. 2017

International Journal of Coal Geology

View full text Add to dashboard Cite

show abstract

“…However, neural networks are complex and involve many parameters which is difficult to choose. Therefore, in most cases, TOC estimation is achieved by constructing simple or multivariate regression methods [16]. Tan et al (2015) applied supportvector-regression machine technology in TOC estimation and acquired good results, and this technique needs multiple inputs [16].…”

Section: Introductionmentioning

confidence: 99%

Dual-shale-content method for total organic carbon content evaluation from wireline logs in organic shale

2017

View full text Add to dashboard Cite

Organic shale is one of the most important unconventional resources all around the world. Total organic carbon (TOC) content is an important evaluation parameter of reservoir hydrocarbon source quality. The regular evaluation methods have higher requirements of well logs and core experiment data for statistical regression. Through analyzing the resistivity and gamma ray logging response characteristics of shale content and organic matters, combined with digital rock physics experiment simulation, we put forward and improve the dual-shale-content method for TOC content logging evaluation. The accuracy of this method is verified by actual data processing. The result shows the dual-shale-content method is simple to use and the error is small. And by comparing with the calculation results by using the ∆LogR method, it is revealed that the trend of the TOC content calculated by our method agrees with the core results better. This new method is suitable for the evaluation of TOC content in the area or interval where there is only conventional logs.

show abstract

“…Considering high uncertainty in brittleness evaluation with simple regression, artificial intelligence technology is a powerful tool to model complex systems that seek to simulate human brain behavior by processing data on a trial-and-error basis. Because its advantages in recognize, cluster and organize complicated nonlinear relationships between parameters, the application of artificial intelligence approaches have been successfully applied in many well logging fields, including formation permeability, porosity and total carbon content (TOC) prediction (Baneshi et al, 2013;Tan et al, 2015). However, little research has been done on shale brittleness index prediction using any artificial intelligence approaches.…”

Section: Introductionmentioning

confidence: 99%