Where to drill next? A dual-weighted approach to adaptive optimal design of groundwater surveys

Lykkegaard, Mikkel B.; Dodwell, Tim

doi:10.1016/j.advwatres.2022.104219

Cited by 7 publications

(8 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Future research will investigate the ability of the framework to sequential BOED, and solve the problem of allocating new sensors placement adaptively such as, for example, Haan et al. (2021), Lykkegaard and Dodwell (2022), Wang et al. (2022).…”

Section: Discussionmentioning

confidence: 99%

“…Because this process is typically costly and/or time‐consuming, it is critical to optimize resource utilization and information gain by controlling the experimental conditions for data acquisition (Attia et al., 2018). This is referred to as optimal experimental design (OED), which is commonly viewed as an optimization problem (Haan et al., 2021; Kleinegesse & Gutmann, 2020; Lykkegaard & Dodwell, 2022; Ryan et al., 2016).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Comparing Well and Geophysical Data for Temperature Monitoring Within a Bayesian Experimental Design Framework

Thibaut

Compaire

Lesparre

et al. 2022

Water Resources Research

View full text Add to dashboard Cite

Geothermal systems, including borehole thermal energy storage (BTES) and shallow aquifer thermal energy storage systems (ATES) are becoming more popular as the world looks for ways to reduce greenhouse gas emissions. Such systems use thermal energy extracted from the ground or groundwater to heat or cool buildings, which necessitates some electrical energy input for the heat pump, while storing the excess heat or cold underground. The goal is to re-use this thermal energy during the next season in a cyclic utilization (Bayer et al., 2013;Duijff et al., 2021;Saner et al., 2010;Vanhoudt et al., 2011). The performance of BTES and ATES strongly depends on the subsurface properties. Many variables are involved in geothermal processes, including porosity, hydraulic conductivity, thermal conductivity, and heat capacity. Subsurface temperature fluctuations are strongly influenced by the spatial distribution of these parameters, the boundary conditions, and the aquifer's hydraulic gradient

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Comparing Well and Geophysical Data for Temperature Monitoring Within a Bayesian Experimental Design Framework

Thibaut

Compaire

Lesparre

et al. 2022

Water Resources Research

View full text Add to dashboard Cite

show abstract

“…This topic was beyond the scope of Chapter 3 and Chapter 4, since the forward model was so expensive to run. Sequential BOED, solving the problem of allocating new sensors placement adaptively such as Haan et al (2021), Lykkegaard and Dodwell (2022), Wang et al (2022c), was developed in Chapter 5.…”

Section: General Limitations Of Experimental Designmentioning

confidence: 99%

Machine Learning for Bayesian Experimental Design in the Subsurface

Thibaut¹

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…We use the Python implementation of MLDA by Lykkegaard (2022) until level ℓ = 0 is reached. We note that different subchain lengths may be used on different levels but the analysis here is restricted to the same J ℓ = J on all levels.…”

Section: Multilevel Markov-chain Monte Carlomentioning

confidence: 99%

“…While multilevel methods have previously been used to accelerate MCMC algorithms (Dodwell et al, 2019;Lykkegaard & Dodwell, 2022; in a formal Bayesian framework, it has not yet been attempted for GLUE. In this study, we utilize ideas from multilevel Monte Carlo strategies to accelerate statistical inversion of hydrological models with the GLUE methodology.…”

Section: Introductionmentioning

confidence: 99%

Extending GLUE with Multilevel Methods to Accelerate Statistical Inversion of Hydrological Models

Rudolph,

Wöhling,

Wagener

et al. 2023

Preprint

View full text Add to dashboard Cite

Inverse problems are ubiquitous in hydrological modelling for parameter estimation, system understanding, sustainable water resources management, and the operation of digital twins. While statistical inversion is especially popular, its sampling-based nature often inhibits the inversion of computationally costly models, which has compromised the use of the Generalized Likelihood Uncertainty Estimation (GLUE) methodology, e.g., for spatially distributed (partial) differential equation based models. In this study we introduce multilevel GLUE (MLGLUE), which alleviates the computational burden of statistical inversion by utilizing a hierarchy of model resolutions. Inspired by multilevel Monte Carlo, most parameter samples are evaluated on lower levels with computationally cheap low-resolution models and only samples associated with a likelihood above a certain threshold are subsequently passed to higher levels with costly high-resolution models for evaluation. Inferences are made at the level of the highest-resolution model but substantial computational savings are achieved by discarding samples with low likelihood already on levels with low resolution and low computational cost. Two test problems demonstrate the similarity of inferred parameter posteriors and uncertainty estimates of MLGLUE and GLUE as well as increased computational efficiency. Findings are furthermore compared to inversion results from Markov-chain Monte Carlo (MCMC) and from multilevel delayed acceptance MCMC. The computation time of inversion of a groundwater flow model was decreases by ≈45% and ≈57% when using MLGLUE instead of conventional formulations of GLUE and MCMC, respectively.

show abstract

Where to drill next? A dual-weighted approach to adaptive optimal design of groundwater surveys

Cited by 7 publications

References 21 publications

Comparing Well and Geophysical Data for Temperature Monitoring Within a Bayesian Experimental Design Framework

Comparing Well and Geophysical Data for Temperature Monitoring Within a Bayesian Experimental Design Framework

Machine Learning for Bayesian Experimental Design in the Subsurface

Extending GLUE with Multilevel Methods to Accelerate Statistical Inversion of Hydrological Models

Contact Info

Product

Resources

About