Conditioning complex subsurface flow models on nonlinear data is complicated by the need to preserve the expected geological connectivity patterns to maintain solution plausibility. Generative adversarial networks (GANs) have recently been proposed as a promising approach for low-dimensional representation of complex high-dimensional images. The method has also been adopted for low-rank parameterization of complex geologic models to facilitate uncertainty quantification workflows. A difficulty in adopting these methods for subsurface flow modeling is the complexity associated with nonlinear flow data conditioning. While conditional GAN (CGAN) can condition simulated images on labels, application to subsurface problems requires efficient conditioning workflows for nonlinear data, which is far more complex. We present two approaches for generating flow-conditioned models with complex spatial patterns using GAN. The first method is through conditional GAN, whereby a production response label is used as an auxiliary input during the training stage of GAN. The production label is derived from clustering of the flow responses of the prior model realizations (i.e., training data). The underlying assumption of this approach is that GAN can learn the association between the spatial features corresponding to the production responses within each cluster. An alternative method is to use a subset of samples from the training data that are within a certain distance from the observed flow responses and use them as training data within GAN to generate new model realizations. In this case, GAN is not required to learn the nonlinear relation between production responses and spatial patterns. Instead, it is tasked to learn the patterns in the selected realizations that provide a close match to the observed data. The conditional low-dimensional parameterization for complex geologic models with diverse spatial features (i.e., when multiple geologic scenarios are plausible) performed by GAN allows for exploring the spatial variability in the conditional realizations, which can be critical for decision-making. We present and discuss the important properties of GAN for data conditioning using several examples with increasing complexity.
Ocean Dynamics - In this paper, a two-phase coastal liquid-sediment system and a wave breaker were numerically simulated using a particle-based smooth particle hydrodynamics (SPH). To verify the... 相似文献
This study examines the impact of income inequality and economic growth on environmental degradation in Pakistan using ARDL bounds testing approach for the period 1966–2011. Empirical results for the aggregate CO2 emissions and its four sources such as CO2 emissions from solid fuel, liquid fuel and gaseous fuel consumption as well as electricity and heat production confirm the existence of long run co-integrated relationship between income inequality, economic growth and environment degradation. The estimated results indicate that carbon emissions increase as the income gap expands in Pakistan. Besides the negative impact of industrial share and population density on CO2 emissions, we also confirm that economic growth in Pakistan comes up with higher emissions. Hence, the hypothesis of EKC is not valid for Pakistan during the study period. Our empirical findings are robust as evidenced by dynamic ordinary least squared and the U-tests. Overall, this study suggests that the distribution of income matters to aggregate carbon emissions and focus should be made on sustained economic growth to reduce pollutants and hence CO2 emission in the study area. 相似文献
The crustal structure beneath three seismic stations over Malaysia has been investigated with the application of the group velocity dispersion analysis of the northern Sumatra earthquake data which occurred on 06 April 2010. Eighteen crustal layer models are constructed to assess the structure. Group velocity dispersions have been computed for the recorded earthquake data using a graphical method and modified Haskell matrix method for the models. Both dispersions have been presented for the interpretation of crustal layers. Findings have shown four major crustal layers having thicknesses of 2.5–4.0, 2.0–5.5, 5.0–8.0, and 8.5–9.0 km, while in Terengganu, it has shown three layers. Density, shear, and compressional wave velocities used in models have suggested that the crustal structure of the northern part of Peninsular Malaysia is crystalline. Major crustal minerals are of quartz, plagioclase, and mica. Most layers seem to have upward directions toward Perak from Kedah and Terengganu. 相似文献
The force analogy method(FAM) is considered as one of the most time-saving and cost-efficient methods for analyzing frames. Through a set of assumptions and restoring forces, FAM analyzes nonlinear frames responses through Hooke's law. This study evaluates the effect of element type on FAM through numerical and experimental tests. The conventional Euler Bernoulli(EB) element is replaced by the Timoshenko(TS) beam-column element and the results are interpreted and compared with experimental findings. Three experimental tests were conducted for benchmarking and comparison purposes via 2 D aluminum frames. The results indicated that during the analysis, when the frame response is in the linear region, there is no difference between the element types in the response of the frames. When the frame entered the plastic region, the frames which were analyzed by the TS element revealed closer responses to the experimental outcomes. The gap between the results of the frame which was analyzed by EB and TS enlarged especially when the frame experienced a sharp or huge rotation of more than 0.2 rad. The final recorded deformations based on the TS element revealed an accuracy between 98.05% to 98.65%, while the EB element showed 14.66% to 45.14% for rotations of more than 0.2 rad at plastic hinge locations(PHLs). 相似文献
The Industrialized Building System (IBS) was recently introduced to minimize the time and cost of project construction. Accordingly, ensuring the integration of the connection of precast components in IBS structures is an important factor that ensures stability of buildings subjected to dynamic loads from earthquakes, vehicles, and machineries. However, structural engineers still lack knowledge on the proper connection and detailed joints of IBS structure construction. Therefore, this study proposes a special precast concrete wall-to-wall connection system for dynamic loads that resists multidirectional imposed loads and reduces vibration effects (PI2014701723). This system is designed to connect two adjacent precast wall panels by using two steel U-shaped channels (i.e., male and female joints). During casting, each joint is adapted for incorporation into a respective wall panel after considering the following conditions: one side of the steel channel opens into the thickness face of the panel; a U-shaped rubber is implemented between the two channels to dissipate the vibration effect; and bolts and nuts are used to create an extension between the two U-shaped male and female steel channels. The developed finite element model of the precast wall is subjected to cyclic loads to evaluate the performance of the proposed connection during an imposed dynamic load. Connection performance is then compared with conventional connections based on the energy dissipation, stress, deformation, and concrete damage in the plastic range. The proposed precast connection is capable of exceeding the energy absorption of precast walls subjected to dynamic load, thereby improving its resistance behavior in all principal directions. 相似文献
Watershed prioritization has gained importance in natural resources management, especially in the context of watershed management.
Morphometric analysis has been commonly applied to prioritization of watersheds. The present study makes an attempt to prioritize
sub-watersheds based on morphometric and land use characteristics using remote sensing and GIS techniques in Kanera watershed
of Guna district, Madhya Pradesh. Various morphometric parameters, namely linear and shape have been determined for each sub-watershed
and assigned ranks on the basis of value/relationship so as to arrive at a computed value for a final ranking of the sub-watersheds.
Land use/land cover change analysis of the sub-watersheds has been carried out using multi-temporal data of IRS LISS II of
1989 and IRS LISS III of 2001. The study demonstrates the significant land use changes especially in cultivated lands, open
scrub, open forest, water bodies and wastelands from 1989 to 2001. Based on morphometric and land use/land cover analysis,
the sub-watersheds have been classified into three categories as high, medium and low in terms of priority for conservation
and management of natural resources. Out of the seven sub-watersheds, two sub-watersheds viz., SW1 and SW6 qualify for high
priority, whereas SW7 has been categorised as medium priority based on the integration of morphometric and land use change
analysis. 相似文献
