DGSA: A Matlab toolbox for distance-based generalized sensitivity analysis of geoscientific computer experiments

Park, Jihoon; Yang, Guang; Satija, Addy; Scheidt, Céline; Caers, Jef

doi:10.1016/j.cageo.2016.08.021

Cited by 57 publications

(36 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If not, an alternative data set can be proposed. Here, distancebased global sensitivity analysis (DGSA) was used to identify the most sensitive parameters (Park et al, 2016;Fenwick et al, 2014). It is worth noticing that these 2 first steps in BEL are field data independent, i.e.…”

Section: Bayesian Evidential Learningmentioning

confidence: 99%

See 1 more Smart Citation

Bayesian evidential learning: validation à partir de tests expérimentaux d’injection-pompage

Hermans

Lesparre

Schepper³

et al. 2019

Hydrogeol J

View full text Add to dashboard Cite

Recent developments in uncertainty quantification show that a full inversion of model parameters is not always necessary to forecast the range of uncertainty of a specific prediction in Earth Sciences. Instead, Bayesian evidential learning (BEL) uses a set of prior models to derive a direct relationship between data and prediction. This recent technique has been mostly demonstrated for synthetic cases. This paper demonstrates the ability of BEL to predict the posterior distribution of temperature in an alluvial aquifer during a cyclic heat tracer push-pull test. The data set corresponds to another push-pull experiment with different characteristics (amplitude, duration, number of cycles). This experiment constitutes the first demonstration of BEL on real data in a hydrogeological context. It should open the range of future applications of the framework for both scientists and practitioners.

show abstract

Section: Bayesian Evidential Learningmentioning

confidence: 99%

“…Then the sensitivity analysis for parameter A is repeated for each bin. If the response between bins is different, then a conditional effect or interaction is identified (Park et al, 2016).…”

Section: Sensitivity Analysismentioning

confidence: 99%

Bayesian evidential learning: validation à partir de tests expérimentaux d’injection-pompage

Hermans

Lesparre

Schepper³

et al. 2019

Hydrogeol J

View full text Add to dashboard Cite

show abstract

“…This is because CCA requires the sum of input data and model variable dimensions to be smaller than the Monte Carlo samples size L: L > dim(d) + dim(m). Otherwise it will always produce perfect correlations (correlation coefficients be 1) (Pezeshki et al, 2004). Although PCA can significantly reduce the dimensionality of m from L × P to L × L, where P is the number of model parameters and L P , this requirement is still difficult to meet.…”

Section: Reviewmentioning

confidence: 99%

“…Both the sensitive and non-sensitive variables will be used for posterior reconstruction in step 6. In this paper, we use a distance-based generalized sensitivity analysis (DGSA) method (Fenwick et al, 2014;Park et al, 2016) to perform sensitivity analysis. Compared to the other global sensitivity analyses, such as variance-based methods (e.g., Sobol, 2001Sobol, , 1993, regionalized methods (e.g., Pappenberger et al, 2008;Spear and Hornberger, 1980), or treebased method (e.g., Wei et al, 2015), DGSA has its specific advantages for high-dimensional problems while requiring no functional form between model responses and model parameters.…”

Section: Reviewmentioning

confidence: 99%

Automated Monte Carlo-based quantification and updating of geological uncertainty with borehole data (AutoBEL v1.0)

2020

Self Cite

View full text Add to dashboard Cite

Abstract. Geological uncertainty quantification is critical to subsurface modeling and prediction, such as groundwater, oil or gas, and geothermal resources, and needs to be continuously updated with new data. We provide an automated method for uncertainty quantification and the updating of geological models using borehole data for subsurface developments within a Bayesian framework. Our methodologies are developed with the Bayesian evidential learning protocol for uncertainty quantification. Under such a framework, newly acquired borehole data directly and jointly update geological models (structure, lithology, petrophysics, and fluids), globally and spatially, without time-consuming model rebuilding. To address the above matters, an ensemble of prior geological models is first constructed by Monte Carlo simulation from prior distribution. Once the prior model is tested by means of a falsification process, a sequential direct forecasting is designed to perform the joint uncertainty quantification. The direct forecasting is a statistical learning method that learns from a series of bijective operations to establish “Bayes–linear-Gauss” statistical relationships between model and data variables. Such statistical relationships, once conditioned to actual borehole measurements, allow for fast-computation posterior geological models. The proposed framework is completely automated in an open-source project. We demonstrate its application by applying it to a generic gas reservoir dataset. The posterior results show significant uncertainty reduction in both spatial geological model and gas volume prediction and cannot be falsified by new borehole observations. Furthermore, our automated framework completes the entire uncertainty quantification process efficiently for such large models.

show abstract

“…We use Distance-based Global Sensitivity Analysis (Park et al, 2016) in order to study the sensitivity of wireline logs to drilling data and select the optimum features for predicting each of the density, porosity, and sonic logs. Figure 3 shows the sensitivity of porosity to drilling parameters and other wireline logs such as density, gamma ray and sonic.…”

Section: Feature Engineeringmentioning

confidence: 99%

Real-Time Well Log Prediction From Drilling Data Using Deep Learning

Kanfar

Shaikh

Yousefzadeh

et al. 2020

Day 3 Wed, January 15, 2020

View full text Add to dashboard Cite

The objective is to study the feasibility of predicting subsurface rock properties in wells from real-time drilling data. Geophysical logs, namely, density, porosity and sonic logs are of paramount importance for subsurface resource estimation and exploitation. These wireline petro-physical measurements are selectively deployed as they are expensive to acquire; meanwhile, drilling information is recorded in every drilled well. Hence a predictive tool for wireline log prediction from drilling data can help management make decisions about data acquisition, especially for delineation and production wells. This problem is non-linear with strong ineractions between drilling parameters; hence the potential for deep learning to address this problem is explored. We present a workflow for data augmentation and feature engineering using Distance-based Global Sensitivity Analysis. We propose an Inception-based Convolutional Neural Network combined with a Temporal Convolutional Network as the deep learning model. The model is designed to learn both low and high frequency content of the data. 12 wells from the Equinor dataset for the Volve field in the North Sea are used for learning. The model predictions not only capture trends but are also physically consistent across density, porosity, and sonic logs. On the test data, the mean square error reaches a low value of 0.04 but the correlation coefficient plateaus around 0.6. The model is able however to differentiate between different types of rocks such as cemented sandstone, unconsolidated sands, and shale.

show abstract

DGSA: A Matlab toolbox for distance-based generalized sensitivity analysis of geoscientific computer experiments

Cited by 57 publications

References 45 publications

Bayesian evidential learning: validation à partir de tests expérimentaux d’injection-pompage

Bayesian evidential learning: validation à partir de tests expérimentaux d’injection-pompage

Automated Monte Carlo-based quantification and updating of geological uncertainty with borehole data (AutoBEL v1.0)

Real-Time Well Log Prediction From Drilling Data Using Deep Learning

Contact Info

Product

Resources

About