Calibration Of Flow Models Research Articles

Improving calibration of groundwater flow models using headwater streamflow intermittence

AbstractNon‐perennial streams play a crucial role in ecological communities and the hydrological cycle. However, the key parameters and processes involved in stream intermittency remain poorly understood. While climatic conditions, geology and land use are well identified, the assessment and modelling of groundwater controls on streamflow intermittence remain a challenge. In this study, we explore new opportunities to calibrate process‐based 3D groundwater flow models designed to simulate hydrographic network dynamics in groundwater‐fed headwaters. Streamflow measurements and stream network maps are considered together to constrain the effective hydraulic properties of the aquifer in hydrogeological models. The simulations were then validated using visual observations of water presence/absence, provided by a national monitoring network in France (ONDE). We tested the methodology on two pilot unconfined shallow crystalline aquifer catchments, the Canut and Nançon catchments (Brittany, France). We found that both streamflow and stream network expansion/contraction dynamics are required to calibrate models that simultaneously estimate hydraulic conductivity and porosity with low uncertainties. The calibration allowed good prediction of stream intermittency, both in terms of flow and spatial extent. For the two catchments studied, Canut and Nançon, the hydraulic conductivity is close reaching 1.5 × 10−5 m/s and 4.5 × 10−5 m/s, respectively. However, they differ more in their storage capacity, with porosity estimated at 0.1% and 2.2%, respectively. Lower storage capacity leads to higher groundwater level fluctuations, shorter aquifer response times, an increase in the proportion of intermittent streams and a reduction in perennial flow. This new modelling framework for predicting headwater streamflow intermittence can be deployed to improve our understanding of groundwater controls in different geomorphological, geological and climatic contexts. It will benefit from advances in remote sensing and crowdsourcing approaches that generate new observational data products with high spatial and temporal resolution.

An approach for topology optimization of heat sinks for two-phase flow boiling: Part 2 – Model calibration and experimental validation

Flow boiling offers increased heat transfer coefficients and effective heat capacity rates that can be leveraged by topology optimization (TO) to generate high-performance microchannel heat sinks. Topology optimization of heat sinks having two-phase flows can be accomplished using the homogenization approach wherein designs are represented as spatially varying distributions of microstructures, as demonstrated in Part 1 of this two-part study. Flow and heat transfer within the microstructures are captured using a two-phase mixture model featuring an effective porous medium formulation. However, closure of these governing equations requires empirical correlations for pressure drop and heat transfer that are specific to the operating conditions, geometry, and surface finish. Therefore, it must be demonstrated these available correlations can be successfully calibrated over a range of microstructural variations present within the homogenization framework, so as to attain the required prediction generality and accuracy needed to ensure the resulting designs achieve Pareto-optimality. This Part 2 of our study successfully demonstrates a comprehensive end-to-end process for two-phase flow model calibration, topology optimized design generation, and experimental validation of optimized performance for flow boiling heat sinks. To this end, a TO algorithm with the homogenization approach and a two-phase mixture model is implemented using available flow boiling correlations for microchannels. A set of uniform pin fin calibration samples are additively manufactured and experimentally tested under flow boiling at various flow rates and heat inputs for model calibration. All of the unknown/free coefficients in the adopted correlations are determined by minimizing the error between the model predictions and the experimental measurements using gradient-based optimization. The calibrated topology optimization algorithm is then used to generate a Pareto-optimal set of heat sinks optimized for minimum pressure drop and thermal resistance during flow boiling. Experimental characterization of these additively manufactured heat sinks, unseen during the model coefficient calibration process, reveals that the measured Pareto optimality curve matches that predicted by the topology optimization algorithm. Lastly, a heat sink design is generated for a design space involving multiple hot spots and background heating to showcase the capability of the experimentally calibrated two-phase topology optimization algorithm at handling complex boundary conditions. The optimized heat sink intelligently distributes an adequate amount of coolant flow to each of the heated regions to avoid local dry-out. This work demonstrates a complete framework for two-phase topology optimization of heat sinks through experimental calibration of flow boiling correlations to the porous medium used by the homogenization approach.

Groundwater Flow Model Calibration Using Variable Density Modeling for Coastal Aquifer Management

The paper investigates the mechanism of seawater intrusion and the performance of free and open-source codes for the simulation of variable density flow problems in coastal aquifers. For this purpose, the research focused on the Marathon Watershed, located in the northeastern tip of Attica, Greece. For the simulation of the groundwater system, MODFLOW, MT3DMS and SEAWAT codes were implemented, while sensitivity analysis and calibration processes were carried out with UCODE. Hydraulic head calibration was performed on the MODFLOW model, and TDS concentration was validated in the SEAWAT model. The calibrated parameters of the MODFLOW model were obtained for the variable density flow simulation with SEAWAT. The MODFLOW and SEAWAT hydraulic head outputs were analyzed and compared to one another. The outcome of this analysis is that SEAWAT produced slightly better results in terms of the hydraulic heads, concluding that parameter transferability can take place between the two models. For the purpose of the seawater intrusion assessment, the use of the SEAWAT code revealed that the aquifer is subjected to passive and passive–active seawater intrusion during wet and dry seasons, respectively. Finally, an irregular shape of a saltwater wedge is developed at a specific area associated with the hydraulic parameters of the aquifer.

Seasonal biodegradation of the artificial sweetener acesulfame enhances its use as a transient wastewater tracer

The persistence of the artificial sweetener acesulfame potassium (ACE) during wastewater treatment and subsequently in the aquatic environment has made it a widely used tracer of wastewater inputs to both surface water and groundwater. However, the recently observed biodegradation of ACE during wastewater treatment has questioned the validity of this application. In this study, we assessed the use of ACE not only as a marker of wastewater, but also as a transient wastewater tracer that allows both the calculation of mixing ratios and travel times through the aquifer as well as the calibration of transient groundwater flow and mass transport models. Our analysis was based on data obtained in a nearly 8-year river water and groundwater sampling campaign along a confirmed wastewater-receiving riverbank filtration site located close to a drinking water supply system. We provide evidence that temperature controls ACE concentration and thus its seasonal oscillation. River water data showed that ACE loads decreased from 1.5–4 mg·s−1 in the cold season (December to June; T<10 °C) to 0–0.5 mg·s−1 in the warm season (July to November; T>10 °C). This seasonal variability of >600% was detectable in the aquifer and preserved >3 km, with ACE concentrations oscillating between <LOQ in the warm season up to 1 μg·L−1 in the cold season. The large seasonal variation in ACE concentrations during wastewater treatment, compared to the other sweeteners (sucralose, cyclamate, and saccharin) and chloride enables its use as a transient tracer of wastewater inflows and riverbank filtration. In addition, the arrival time of the ACE concentration peak can be used to estimate groundwater flow velocity and mixing ratios, thereby demonstrating its potential in the calibration of groundwater numerical models.

Using helium-4, tritium, carbon-14 and other hydrogeochemical evidence to evaluate the groundwater age distribution: The case of the Neogene aquifer, Belgium

Apparent groundwater age dating has been proven useful and robust in understanding water origin and mixing processes, particularly when multiple tracers are considered. However, even though now extensively used, the age tracers have not been widely applied in the general practice of flow and transport model calibration. A multi tracer-study was carried out in the Neogene aquifer in Flanders to quantify the apparent age and construct a joint interpretation for the delineation of different groundwater flow systems. This understanding is critical as part of the safety and feasibility studies for the underlying Boom Clay Formation that has been considered as a potential host rock for the geological disposal of radioactive waste. In this study, we combine evidence from tritium/helium-3 (3H/3He), helium-4 (4He) and radiocarbon (14C) dating as well as stable isotopic (δ18O, δ2H) and hydrochemical signatures in combination with particle tracking-based age distributions from the 3D groundwater flow model. The results of the study indicate that mixing of groundwater with young and old fractions occurs predominantly in the central part of the aquifer which is made evident by the coexistence of 3H (pre and post-bomb pulse Era), 14C and 4He in several groundwater samples. The mixing between water of different origin is also supported by the sampled stable isotopic and hydrochemical composition of groundwater. Particle tracking residence time results show an acceptable agreement with apparent ages derived from age tracers for young (≤100 years) and old (>1000 years) groundwater. Groundwater with ages between 100 and 1000 years is likely a mixture of water with young/old fractions and shows the strongest discrepancies between advective model ages and age tracer based apparent ages. On the basis of our findings, we distinguish between three groundwater flow systems in the Neogene aquifer: i) a shallow/local flow system, with groundwater originating from modern meteoric water; ii) a deep/semi-regional flow system, characterized by old groundwater where the presence of 4Herad is significant; iii) a mixed zone of groundwater flow where the recently infiltrated meteoric water mixes with discharging old groundwater. These results have helped us to refine previously proposed conceptual models for the study area and will in the end reduce uncertainties relevant to the potential future geological disposal of radioactive waste.

R2ogs5: Calibration of Numerical Groundwater Flow Models with Bayesian Optimization in R.

Large-scale and high-resolution groundwater models are currently becoming increasingly important in order to clarify the extent to which climate trends and extreme weather affect the groundwater balance regionally. As a result, the parameterization of groundwater models is becoming more detailed and more complex, making conventional calibration methods too time-consuming. Moderating the computational demand to find optimal solutions for the resulting potentially multi-modal objective function requires intelligent and efficient global optimization methods. Moreover, the increasing use of modern scripting languages R and Python to craft environmental analysis workflows calls to integrate groundwater flow simulators in such. Here we introduce and exemplify r2ogs5, a tool that integrates version 5 of the open-source simulation software OpenGeoSys into the programming language R. r2ogs5 allows for calibration of numerical groundwater flow models with a sequential model-based optimization approach that combines Bayesian optimization (BO) with surrogate modeling. Here, we describe the structure and function of r2ogs5 as well as the implemented calibration method. We then demonstrate the calibration method by calibrating 4 and 12 parameters of two simple groundwater flow models. The results indicate that this method needs fewer runs of the groundwater flow model than conventional gradient search and Latin hypercube sampling in case of the 12 parameter model. We believe that our method offers the potential to calibrate computationally expensive groundwater flow models. r2ogs5 supports groundwater flow modelers to access the statistical analysis and visualization capabilities of the R language and is a valuable tool for geoscientists already using R.

High spatial resolution prediction of tritium (3H) in contemporary global precipitation

Tritium (3H) in Earth’s precipitation is vigilantly monitored since historical nuclear bomb tests because of radiological protection considerations and its invaluable role as a tracer of the global water cycle in quantifying surface, groundwater, and oceanic fluxes. For hydrological applications, accurate knowledge of 3H in contemporary local precipitation is prerequisite for dating of critical zone water and calibrating hydrogeologic transport and groundwater protection models. However, local tritium input in precipitation is hard to constrain due to few 3H observation sites. We present new high-spatial resolution global prediction maps of multi-year mean 3H in contemporary “post-bomb” (2008–2018) precipitation by using a robust regression model based on environmental and geospatial covariates. The model accurately predicted the mean annual 3H in precipitation, which allowed us to produce global 3H input maps for applications in hydrological and climate modelling. The spatial patterns revealed natural 3H in contemporary precipitation sufficient for practical hydrological applications (1–25 TU) but variable across continental regions and higher latitudes due to cumulative influences of cyclical neutron fluxes, stratospheric inputs, and distance from tropospheric moisture sources. The new 3H maps provide a foundational resource for improved calibration of groundwater flow models and critical zone vulnerability assessment and provides an operational baseline for quantifying the potential impact of future anthropogenic nuclear activities and hydroclimatic changes.

Optimized Pilot Point Emplacement Based Groundwater Flow Calibration Method for Heterogeneous Small-Scale Area

Groundwater flow modeling in a small-scale area requires practical techniques to obtain high accuracy results. The effectiveness of the model calibration is the most challenging for simulating the hydraulic head. In pursuit of this, we proposed an optimized groundwater flow calibration method based on the pilot point emplacement technique for a 3D small-scale area in this work. Subsequently, two emplacement structures were tested during the experimentation, the regular pilot point placement, and the middle head measurement down gradient (MHMDG) placement with two different densities. The parameter estimation (PEST) numerical code applying the kriging interpolation was used to estimate the hydraulic conductivity field by MODFLOW. Moreover, geological SGrid models were chosen for the conceptual model. Thirty-seven observation wells were used for experimental simulations to test the proposed method in a heterogeneous confined aquifer. The result shows that the small-scale modeling was complicated, and the studying area presented a significant heterogeneity in horizontal hydraulic conductivity. The middle head measurement down gradient (MHMDG) pilot point case with the larger density gave the best R-squared 0.901 and minimum residual error of 0.0053 m compared to 0.880 and 0.078 m, respectively, for the regular placement. The calibration accuracy depended on the frequency and the emplacement of the pilot point. Therefore, the initial value should be technically selected to minimize the computation burden. The proposed techniques help to improve the groundwater flow model calibration based on the pilot point methodology for groundwater resources management.

USING OF LOW COST MOISTURE SENSORS IN LABORATORY EXPERIMENTS

&amp;lt;jats:p&amp;gt;Soil moisture is one of crucial parameters in modeling and analysis of groundwater flow. Measurement of soil water dynamics improves the understanding of processes and is vital for good groundwater flow model calibration. Determining soil moisture change during time by laboratory methods, as an integral part of practical classes, requires significant material and financial resources. Nowadays, relatively cheap sensors for measuring soil moisture are available. They could be combined with affordable data loggers to speed up the measurement procedure and to reduce cost price. In this paper, the possibility of using these sensors for measuring soil moisture and their usage for the educational experimental exercises in a field of hydraulic engineering was investigated. The paper first presents the procedure of sensor calibration for four different sensors used in two types of soil, and then their usage in a simple hydraulic engineering experiment.&amp;lt;/jats:p&amp;gt;

Conditioning generative adversarial networks on nonlinear data for subsurface flow model calibration and uncertainty quantification

Conditioning generative adversarial networks on nonlinear data for subsurface flow model calibration and uncertainty quantification

Deep Convolutional Autoencoders for Robust Flow Model Calibration Under Uncertainty in Geologic Continuity

AbstractSubsurface flow model calibration is commonly performed by assuming that a known conceptual model of geologic continuity is available and can be used to constrain the solution search space. In real applications, however, the knowledge about geologic continuity is far from certain and subjective interpretations can lead to multiple distinct plausible geologic scenarios. Conventional parameterization methods that are widely used in model calibration, such as the principal component analysis, encounter difficulty in capturing diverse spatial patterns from distinct geologic scenarios. We propose a deep learning architecture, known as variational auto‐encoder, for robust dimension‐reduced parameterization of spatially distributed aquifer properties, such as hydraulic conductivity, in solving model calibration problems under uncertain geostatistical models. We show that convolutional autoencoders offer the versatility and robustness required for nonlinear parameterization of complex subsurface flow property distributions when multiple distinct geologic scenarios are present. The robustness of these models results, in part, from the use of many convolutional filters that afford the redundancy needed to extract, classify and encode very diverse spatial patterns at different abstraction levels/scales and enable their mapping onto low‐dimensional variables in a learned latent space. The resulting low‐dimensional latent variables control the salient spatial patterns in different geologic continuity models and are effective for parameterization of model calibration problems under uncertainty in geologic continuity, a task that is not trivial to accomplish using traditional parameterization methods. Several numerical experiments are used to demonstrate the robustness of convolutional deep learning models for reduced‐order parameterization of flow model calibration problems when alternative plausible geologic continuity models are present.

Driving Behavior at Signalized Intersections Operating under Disordered Traffic Conditions

A higher degree of heterogeneity in vehicle class and drivers, coupled with non-lane-based driving habits, creates several challenges in traffic flow analysis. This study investigates vehicles’ microscopic driving behavior at signalized intersections operating under weak lane discipline with mixed traffic (disordered) conditions. For this purpose, a comprehensive vehicular trajectory data set is developed from field-recorded video footage using a semi-automated tool for data extraction. Microscopic parameters such as relative velocity, spacing between vehicles, following time, lane preference, longitudinal and lateral speed profile, hysteresis evidence, and lateral movement of different vehicle classes during different traffic phases are presented in the study. The data is then segregated into three flow conditions: stopped flow, saturated flow, and unaffected flow. It is found that smaller vehicles prefer near-side lanes over far-side lanes. Motorized three-wheeler (3W) and motorized two-wheeler (2W) vehicle classes exhibit the greatest lateral velocity, lateral movement, and aggressiveness. This results in several interactions between vehicles as a function of different leader–follower vehicle pairs. Signalized intersections with more heterogeneity in traffic composition, especially higher composition of 2W and 3W vehicle classes, exhibit higher levels of aggressive driving behavior that might lower safety standards. As a practical application, ranges of various driving behavior parameter values for different leader–follower combinations and traffic conditions are quantified in the study. The observations and results are expected to help better understand prevailing driving behavior in disordered traffic and contribute toward robust calibration of microscopic traffic flow models for better replicating disordered traffic conditions at signalized intersections.

Groundwater Flow Model Calibration of a Coastal Multilayer Aquifer System Based on Statistical Sensitivity Analysis

Groundwater Flow Model Calibration of a Coastal Multilayer Aquifer System Based on Statistical Sensitivity Analysis

Groundwater flow velocities in karst aquifers; importance of spatial observation scale and hydraulic testing for contaminant transport prediction.

We review scale dependence of hydraulic conductivities and effective porosities for prediction of contaminant transport in four UK karst aquifers. Approaches for obtaining hydraulic parameters include core plug, slug, pumping and pulse tests, calibration of groundwater flow models and spring recession curves. Core plug and slug tests are unsuitable because they do not characterize a large enough volume to include a representative fracture network. Pumping test values match regional-scale hydraulic conductivities from flow modelling for the less intensively karstified aquifers: Magnesian Limestone, Jurassic Limestone and Cretaceous Chalks. Reliable bulk hydraulic conductivities were not available for the intensively karstified Carboniferous Limestone due to dominance of flow through pipe conduits in Mendips. Here, the only hydraulic conductivity value found from spring recession is one order of magnitude higher than that indicated by pumping tests. For all four carbonate aquifers, effective porosities assumed for transport modelling are two orders of magnitude higher than those found from tracer and hydrogeophysical tests. Thus, a combination of low hydraulic conductivities and assumed flowing porosities resulted in underestimated flow velocities. The UK karst aquifers are characterized by a range of hydraulic behaviours that fit those of karst aquifers worldwide. Indeed, underestimation of flow velocity due to inappropriate parameter selection is common to intensively karstified aquifers of southern France, north-western Germany and Italy. Similar issues arise for the Canadian Silurian carbonates where the use of high effective porosities (e.g. 5%) in transport models leads to underestimation of groundwater velocities. We recommend values in the range of 0.01-1% for such aquifers.

GEOPOLIS software tools – Groundwater flow modeling at deep repository of liquid radioactive waste at Severny test site

This article discusses a three-dimensional groundwater flow model of a deep disposal facility at Severny test site. The three-dimensional model is a part of the certified software GEOPOLIS, based on the hydrogeological code GeRa (Geomigration of Radionuclides) serving as the calculation engine. This study describes the hydrogeological patterning of the groundwater flow model, as well as the results of calibration and verification of the model water heads with respect to the data of monitoring for more than 40 years of the deep repository exploitation. The article begins with a brief overview of the previously developed hydrogeological models of this object and continues with a description of the geological structure of the territory, and with a substantiation of the boundaries and parameters of the model. The results of groundwater flow modeling, model calibration, verification and estimation of discrepancy between the model results and monitoring data are shown. The comparison of the modeled and observed water heads in the stationary conditions (before the start of injection) and during operation of the deep repository allows making conclusion on the quality of calibration.

Inverting subsurface flow data for geologic scenarios selection with convolutional neural networks

Subsurface flow model calibration against dynamic response data is often constrained by a prior conceptual model of geologic scenario that specifies the expected spatial variability and geologic patterns in the solution. However, several sources of uncertainty exist in developing a conceptual geological model, including data limitation, process-based modeling assumptions, and subjective interpretations. Therefore, it is prudent to consider the uncertainty in geologic scenarios during model calibration and to utilize the flow response data in supporting or rejecting the proposed scenarios. We present a novel framework whereby dynamic response data is used to screen a set of proposed geologic scenarios by combining gradient-based inversion for feature extraction and convolutional neural networks for feature recognition. To compactly represent the salient features of each scenario, an extremely low-rank parameterization with the leading elements of the Principal Components Analysis (PCA) is adopted. The PCA basis elements of different scenarios are combined to define the search space for feature extraction, which is implemented via an iterative least-squares inversion formulation to match the observed data. The inversion solution is obtained by selectively combining the PCA representations from different scenarios, and contains the dominant spatial patterns that are needed to match the observed data. Feature recognition to identify the relevant geologic scenarios based on the reconstructed spatial solution of feature extraction inversion is performed using a pre-trained convolutional neural network (CNN). The method offers several advantages, including an efficient implementation that does not require extensive forward simulation runs, use of flow data to identify pertinent geologic scenarios, and the ability to combine geologic scenarios if supported by data. The performance of the proposed approach is evaluated using numerical experiments from pumping tests in groundwater aquifers.

Research on drinking-groundwater source safety management based on numerical simulation

A drinking-groundwater source protection technology system based on a three-dimensional finite-difference groundwater model was constructed and applied to the safe management of drinking groundwater in the first terrace of Fujiang River. In the study area, the main type of groundwater is the quaternary systemic alluvial deposit loose rock pore water and the aquifer thickness varies between 20 and 35 m. Groundwater is the main source of water and is used for various purposes through two exploitation wells. The water volumes of 1# exploitation well (1#) and 2# exploitation well (2#) are 10,000 m3/day and 5000 m3/day, respectively. An analysis of 22 indicators from 11 groundwater samples showed that a higher concentration of chemical-oxygen-demand (CODMn) and ammonia–nitrogen (NH3–N), and they had a high correlation with most of the other water-quality factors. Therefore, CODMn and NH3–N were selected as indicator factors for model calibration and prediction. Twenty-two hydraulic head observation wells were used for flow-model calibration. The flow model indicated that a drop funnel formed with a maximum depth of 12 m, and the particle-capture zone in the original downstream direction of the south side extended to 1100 m because of groundwater exploitation. The solute-transport model showed that industrial pollution sources were the main factors that led to a deterioration of water quality. To analyze the necessity and effectiveness of remediation measures for the safety of drinking-water sources, two scenarios were considered to predict the concentration of NH3-N and CODMn in groundwater exploitation wells over 20 years. Scenario I, which considered that current conditions were maintained, predicted that the NH3-N would exceed the drinking-water quality standard of 0.5 mg/L after 16 years. Scenario II, in which industrial sewage treatment plants were installed outside the particle-trapped zone of the exploitation wells and strict anti-seepage measures were implemented, predicted that the peak concentrations of NH3-N and CODMn in the exploitation wells would be 0.26 mg/L and 1.33 mg/L, respectively, after 3 years of model operation. This study provides a theoretical basis for drinking-groundwater source protection that can be applied to safety management practices.

A pattern-matching method for flow model calibration under training image constraint

Modern geostatistical modeling techniques are developed to simulate complex geologic connectivity patterns (e.g., curvilinear fluvial systems) at the grid-level using a training image (TI) that encodes multiple-point statistical (MPS) information. A challenging aspect of using MPS methods is conditioning the resulting models on nonlinear flow data. We develop a pattern-matching method for calibration of MPS-based facies models subject to the TI constraint. Since the exact statistical information in the TI can only be expressed empirically, flow data conditioning and pattern matching are carried out in two iterative steps, using an alternating-direction algorithm. Flow data integration is formulated through a regularized least-squares by taking advantage of learned k-SVD sparse parametrization and l1-norm sparsity-promoting regularization methods. The TI constraint is enforced through a MPS-based pattern-matching algorithm that uses the identified model calibration solution to generate a corresponding facies model that is consistent with the TI. The pattern-matching algorithm uses a local search template to scan the TI to find facies patterns with smallest distances from the corresponding local patterns in the parameterized approximate solution. The identified patterns for each location in the model are stored and used to estimate local conditional probabilities for assigning the facies types to each grid cell. The resulting solution is passed to the flow data conditioning step as a regularization term to perform the next iteration. The process is repeated until the MPS facies model provides an acceptable match to the data. Numerical experiments are presented to evaluate the performance of the pattern-matching method for calibration of complex facies models.

Experiments and sensitivity coefficients analysis for multiphase flow model calibration of enhanced DNAPL dissolution

Multiphase flow modeling is often used for the comparison and optimization of subsurface nonaqueous phase liquid (NAPL) remediation schemes. The calibration of such models is a challenging task due to the lack of detailed data describing the initial NAPL spatial distribution and the processes governing the fate and transport of NAPLs in porous media. In this study laboratory scale experiments were conducted to evaluate reagent-enhanced dense nonaqueous phase liquid (DNAPL) solubilization in saturated heterogeneous media. The DNAPL consisted of both pooled and residual saturation forms. To gain insight into the influence of various input parameters on effluent concentrations, the multiphase flow program was used to compute the sensitivity coefficients of key parameters, relating to the flow, flushing solution properties, soil parameters, NAPL distribution and mass transfer coefficient. The sensitivity coefficients were, in turn, used to aid in the model calibration and to underline the difficulties associated with the calibration of multiphase flow models, most notably the non-uniqueness of the calibration process when complete information is lacking. To alleviate this uncertainty and provide additional constraints, the conducted flushing experiments were jointly used to calibrate the multiphase flow model. The results of the model calibration suggest that the interphase mass transfer coefficient is dependent on the properties of the reagent aqueous solution used for DNAPL remediation, most notably the viscosity and interfacial tension.

Integration of soft data into multiple-point statistical simulation: re-assessing the probability conditioning method for facies model calibration

Conditioning multiple-point statistical (MPS) facies simulation on dynamic flow data is complicated by the complex relationship between discrete facies distribution and the flow response data. The process can be facilitated by using the probability conditioned method (PCM), in which an ensemble-based data assimilation approach is used to infer, from observed flow data, a facies probability map that can guide conditional simulation from a training image (TI). We discuss important aspects of flow model calibration based on facies probability maps and propose several improvements to the original PCM, including: (1) using a statistically consistent approach for inferring facies probability maps from updated ensemble of continuous hydraulic properties; (2) assigning spatially variable weights to simulation grids based on relative confidence in the information content of observed data; and (3) considering spatial patterns instead of marginal facies probabilities in constructing probability maps. We illustrate that, in simulating categorical facies, the pattern-imitating behavior of MPS can have a significant impact on soft data conditioning. In particular, in the Single Normal Equation SIMulation (SNESIM) algorithm, the initial points that are visited on the random path tend to determine the main global facies connectivity. Once these initial points are populated with simulated values, the conditional probabilities (inferred from the TI) for the subsequent cells on the random path tend to be very close to 0 and 1. Hence, the facies types in the remaining cells are known almost deterministically, making the soft data inconsequential at the later stages of simulation. This behavior suggests that, without compensating for this effect, the facies probability map is not effectively integrated into the SNESIM simulation, an issue that is reflected in the generated realizations. Several numerical experiments are presented and discussed to illustrate this behavior and its impact on soft-data conditioning.