Correlations of Rock-Physics Model Parameters From Bayesian Analysis: Pressure- and Porosity-Dependent Models Applied to Unconsolidated Sands

Correlations of rock-physics model inputs are important to know to help design informative prior models within integrated reservoir-characterization workflows. A Bayesian framework is optimal to determine such correlations. Within that framework, we use velocity and porosity measurements on unconsolidated, dry, and clean sands. Three pressure- and three porosity-dependent rock-physics models are applied to the data to examine relationships among the inputs. As with any Bayesian formulation, we define a prior model and calculate the likelihood in order to evaluate the posterior. With relatively few inputs to consider for each rock-physics model, we found that sampling the posterior exhaustively to be convenient. The results of the Bayesian analyses are multivariate histograms that indicate most likely values of the input parameters given the data to which the rock-physics model was fit. When the Bayesian procedure is repeated many times for the same data, but with different prior models, correlations emerged among the input parameters in a rock-physics model. These correlations were not known previously. Implications, for the pressure- and porosity-dependent models examined here, are that these correlations should be utilized when applying these models to other relevant data sets. Furthermore, additional rock-physics models should be examined similarly to determine any potential correlations in their inputs. These correlations can then be taken advantage of in forward and inverse problems posed in reservoir characterization.

Monitoring geological storage of CO2 using a new rock physics model

To mitigate the global warming crisis, one of the effective ways is to capture CO2 at an emitting source and inject it underground in saline aquifers, depleted oil and gas reservoirs, or in coal beds. This process is known as carbon capture and storage (CCS). With CCS, CO2 is considered a waste product that has to be disposed of properly, like sewage and other pollutants. While and after CO2 injection, monitoring of the CO2 storage site is necessary to observe CO2 plume movement and detect potential leakage. For CO2 monitoring, various physical property changes are employed to delineate the plume area and migration pathways with their pros and cons. We introduce a new rock physics model to facilitate the time-lapse estimation of CO2 saturation and possible pressure changes within a CO2 storage reservoir based on physical properties obtained from the prestack seismic inversion. We demonstrate that the CO2 plume delineation, saturation, and pressure changes estimations are possible using a combination of Acoustic Impedance (AI) and P- to S-wave velocity ratio (Vp/Vs) inverted from time-lapse or four-dimensional (4D) seismic. We assumed a scenario over a period of 40 years comprising an initial 25 year injection period. Our results show that monitoring the CO2 plume in terms of extent and saturation can be carried out using our rock physics-derived method. The suggested method, without going into the elastic moduli level, handles the elastic property cubes, which are commonly obtained from the prestack seismic inversion. Pressure changes quantification is also possible within un-cemented sands; however, the stress/cementation coefficient in our proposed model needs further study to relate that with effective stress in various types of sandstones. The three-dimensional (3D) seismic usually covers the area from the reservoir's base to the surface making it possible to detect the CO2 plume's lateral and vertical migration. However, the comparatively low resolution of seismic, the inversion uncertainties, lateral mineral, and shale property variations are some limitations, which warrant consideration. This method can also be applied for the exploration and monitoring of hydrocarbon production.

Estimating adsorption isotherm parameters in chromatography via a virtual injection promoting double feed-forward neural network

The means to obtain the adsorption isotherms is a fundamental open problem in competitive chromatography. A modern technique of estimating adsorption isotherms is to solve a nonlinear inverse problem in a partial differential equation so that the simulated batch separation coincides with actual experimental results. However, this identification process is usually ill-posed in the sense that the uniqueness of adsorption isotherms cannot be guaranteed, and moreover, the small noise in the measured response can lead to a large fluctuation in the traditional estimation of adsorption isotherms. The conventional mathematical method of solving this problem is the variational regularization, which is formulated as a non-convex minimization problem with a regularized objective functional. However, in this method, the choice of regularization parameter and the design of a convergent solution algorithm are quite difficult in practice. Moreover, due to the restricted number of injection profiles in experiments, the types of measured data are extremely limited, which may lead to a biased estimation. In order to overcome these difficulties, in this paper, we develop a new inversion method – the virtual injection promoting double feed-forward neural network (VIP-DFNN). In this approach, the training data contain various types of artificial injections and synthetic noisy measurement at outlet, generated by a conventional physics model – a time-dependent convection-diffusion system. Numerical experiments with both artificial and real data from laboratory experiments show that the proposed VIP-DFNN is an efficient and robust algorithm.

Tumor Hypoxia Heterogeneity Affects Radiotherapy: Inverse-Percolation Shell-Model Monte Carlo Simulations

Tumor hypoxia was discovered a century ago, and the interference of hypoxia with all radiotherapies is well known. Here, we demonstrate the potentially extreme effects of hypoxia heterogeneity on radiotherapy and combination radiochemotherapy. We observe that there is a decrease in hypoxia from tumor periphery to tumor center, due to oxygen diffusion, resulting in a gradient of radiative cell-kill probability, mathematically expressed as a probability gradient of occupied space removal. The radiotherapy-induced break-up of the tumor/TME network is modeled by the physics model of inverse percolation in a shell-like medium, using Monte Carlo simulations. The different shells now have different probabilities of space removal, spanning from higher probability in the periphery to lower probability in the center of the tumor. Mathematical results regarding the variability of the critical percolation concentration show an increase in the critical threshold with the applied increase in the probability of space removal. Such an observation will have an important medical implication: a much larger than expected radiation dose is needed for a tumor breakup enabling successful follow-up chemotherapy. Information on the TME’s hypoxia heterogeneity, as shown here with the numerical percolation model, may enable personalized precision radiation oncology therapy.

The Genomic Physics of COVID-19 Pathogenesis and Spread

Coronavirus disease (COVID-19) spreads mainly through close contact of infected persons, but the molecular mechanisms underlying its pathogenesis and transmission remain unknown. Here, we propose a statistical physics model to coalesce all molecular entities into a cohesive network in which the roadmap of how each entity mediates the disease can be characterized. We argue that the process of how a transmitter transforms the virus into a recipient constitutes a triad unit that propagates COVID-19 along reticulate paths. Intrinsically, person-to-person transmissibility may be mediated by how genes interact transversely across transmitter, recipient, and viral genomes. We integrate quantitative genetic theory into hypergraph theory to code the main effects of the three genomes as nodes, pairwise cross-genome epistasis as edges, and high-order cross-genome epistasis as hyperedges in a series of mobile hypergraphs. Charting a genome-wide atlas of horizontally epistatic hypergraphs can facilitate the systematic characterization of the community genetic mechanisms underlying COVID-19 spread. This atlas can typically help design effective containment and mitigation strategies and screen and triage those more susceptible persons and those asymptomatic carriers who are incubation virus transmitters.

Outstanding Performance of a New Exfoliated Clay Impregnated with Rutile TiO2 Nanoparticles Composite for Dyes Adsorption: Experimental and Theoretical Studies

Pure rutile TiO2 nanoparticles (Rt) were combined with exfoliated black clay (BC) to prepare a new composite for water decontamination, in particular, for the uptake of methylene blue (MB) and methyl orange (MO) dyes. The as-prepared Rt/BC was characterized by X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and Field emission scanning electron microscopy (FESEM) techniques, and the dyes’ adsorption isotherms at three temperatures (i.e., 25, 40, and 50 °C) were studied. The results indicated that Rt/BC displayed a high removal performance for MO (96.7%) and MB (91.4%) at pH 3.0 and 8.0, respectively. Adsorption data of MB and MO were adjusted by a double layer model at all temperatures. The theoretical parameters of this statistical physics model were interpreted to understand the MO and MB adsorption mechanisms at the molecular level. The removed molecules per active site (n) of Rt/BC ranged from 1.12 to 1.29 for MB and 1.47 to 1.85 for MO, thus representing parallel orientation and multi-interactions mechanisms (i.e., van der Waals forces, hydrogen bonding, and electrostatic interactions were involved). The Rt/BC composite had a density of surface adsorption sites of 100 mg/g. The aggregation of MO molecules was high and increased their adsorption capacities (Qsat = 294–370 mg/g) compared to that of MB (Qsat = 214–249 mg/g). Adsorption energies were 9.70–20.15 kJ/mol, and these values indicated that MO and MB adsorption processes were endothermic and occurred via physical interactions. Overall, the low cost, high regeneration performance, and stability of Rt/BC support its application as a promising adsorbent for organic pollutants from wastewaters.

Towards the development of an automated electrical self-potential sensor of melt and rainwater flow in snow

To understand snow structure and snowmelt timing, information about flows of liquid water within the snowpack is essential. Models can make predictions using explicit representations of physical processes, or through parameterization, but it is difficult to verify simulations. In situ observations generally measure bulk quantities. Where internal snowpack measurements are made, they tend to be destructive and unsuitable for continuous monitoring. Here, we present a novel method for in situ monitoring of water flow in seasonal snow using the electrical self-potential (SP) geophysical method. A prototype geophysical array was installed at Col de Porte (France) in October 2018. Snow hydrological and meteorological observations were also collected. Results for two periods of hydrological interest during winter 2018–19 (a marked period of diurnal melting and refreezing, and a rain-on-snow event) show that the electrical SP method is sensitive to internal water flow. Water flow was detected by SP signals before it was measured in conventional snowmelt lysimeters at the base of the snowpack. This initial feasibility study shows the utility of the SP method as a non-destructive snow sensor. Future development should include combining SP measurements with a high-resolution snow physics model to improve prediction of melt timing.

Neural Network-Based Electric Vehicle Range Prediction for Smart Charging Optimization

Range prediction is a standard feature in most modern road vehicles, allowing drivers to make informed decisions about when to refuel. Most vehicles make range predictions through data- or model-driven means, monitoring the average fuel consumption rate or using a tuned vehicle model to predict fuel consumption. The uncertainty of future driving conditions makes the range prediction problem challenging, particularly for less pervasive battery electric vehicles (BEV). Most contemporary machine learning-based methods attempt to forecast the battery SOC discharge profile to predict vehicle range. In this work, we propose a novel approach using two recurrent neural networks (RNNs) to predict the remaining range of BEVs and the minimum charge required to safely complete a trip. Each RNN has two outputs which can be used for statistical analysis to account for uncertainties; the first loss function leads to mean and variance estimation (MVE), while the second results in bounded interval estimation (BIE). These outputs of the proposed RNNs are then used to predict the probability of a vehicle completing a given trip without charging, or if charging is needed, the remaining range and minimum charging required to finish the trip with high probability. Training data was generated using a low-order physics model to estimate vehicle energy consumption from historical drive cycle data collected from medium-duty last-mile delivery vehicles. The proposed method demonstrated high accuracy in the presence of day-to-day route variability, with the root-mean-square error (RMSE) below 6% for both RNN models.