Constrained voxel inversion using the Cartesian cut cell method

This work presents examples from processing and interpretation of 2576 km of SkyTEM data in the Ringsted-Suså groundwater mapping area located southeast of Denmark. The mapping area is approx. 422 km2 and includes approx. 75,000 SkyTEM soundings. Additionally, diverse ground based geophysical data (>260 km Multi-Electrode Profiling (MEP), > 130 km Pulled-Array Continuous Electrical Soundings (PACES) and 200 ground-based TEM soundings) were utilized in a pseudo-3D inversion. Furthermore prior information was added from 39 borehole logs and lithological information from approx. 735 boreholes in the final inversion. The objectives of the work were to map the near subsurface for units of hydrogeological significance and to evaluate the available inversion schemes. Three types of inversion were used and compared; Laterally Constrained Inversion (LCI), Spatially Constrained Inversion (SCI) and Spatial Mutually Constrained Inversion (SMCI). SMCI is seen to produce results more consistent with both geological data of other geophysical data. Furthermore, the SMCI is less influenced by the degree of lineaments compared with LCI and SCI. This is due to the use of different geophysical data sets represented in one inversion result, which in addition allows gathering the geophysical inversion into one result.

Sound generation from a vibrating circular piston is a classical acoustic problem. The goal of this paper is to simulate numerically the sound radiation produced by oscillating baffled pistons, using both linear and nonlinear model, and to consider the interplay between wave propagation and geometric complexities. The linear solution, based on the linear Euler equations, will be compared to the Rayleigh integral approximation. The nonlinear solution, based on the Navier-Stokes equations, will be compared against linear model for low speed (less than 0.01 of sound speed). A main practical interest in this problem is to capture the behavior of the waves resulting from the source pistons with other solid objects or waves. The wave's properties in terms of frequency, amplitude and wavenumber are influenced by the initial frequencies and coordinates of the pistons, and the geometry. The wave equations in Cartesian coordinate with cut-cell and local grid refinement technique are employed along with the Optimized Prefactored Compact finite volume (OPC-fv) scheme for spatial discretization, the Low-Dispersion Low-Dissipation Runge-Kutta (LDDRK) scheme for time discretization. Problems for the waves around different geometries, and with varied frequencies and amplitudes are considered and presented.

This paper presents a new wall model to compute turbulent boundary layers using the Reynolds-averaged Navier–Stokes equations in high Reynolds number flows. The model solves a two-point boundary value problem for a coupled set of equations for the streamwise velocity and the turbulent viscosity. Since it includes both the pressure gradient and the momentum balance of the full Navier–Stokes system, the ordinary differential equation is valid farther from the wall. We implement the model within a Cartesian cut-cell method and use one-dimensional linelets in each cut cell to avoid the excessive mesh refinement that would otherwise be needed. The linelets are coupled to the outer Cartesian grid in a fully conservative manner, with two-way interaction between the linelets and the background grid. Detailed comparisons of velocity and eddy viscosity with three well-studied examples from the Turbulence Modeling Resource website are presented to demonstrate the model’s performance in two space dimensions, including an example with smooth-body separation. The results show the new model gives excellent results even when the value of the first point is in the wake layer.

Analysis of the bias induced by voxel and unstructured mesh Monte Carlo models for the MCNP6 code in orthovoltage applications

The objective of this work is to assess the performance of a dimensionally-split Cartesian cut-cell method for the numerical solution of a model for air and water droplet flow around aerodynamic bodies. The solver is implemented within a structured hierarchical adaptive mesh refinement (AMR) framework. A brief analysis of the mathematical model is given and the solution to the Riemann problem for the droplet phase is discussed. The dimensionally-split cut-cell approach is extended to facilitate the droplet phase equations. In particular, a modified boundary condition which accounts for collection efficiency is introduced. Validation case studies for dry air flow and water droplet laden flow are presented and the numerical results are compared to experimental data.

OctSurf: Efficient hierarchical voxel-based molecular surface representation for protein-ligand affinity prediction

An explicit spline representation of smooth free-form surfaces is combined with a hierarchy of meshes to form the basis of an interactive sculpting environment. The environment offers localized hierarchical modeling at different levels of detail, direct surface manipulation, change of connectivity for extrusion and to form holes and bridges, and built-in tangent continuity across the surface where wanted. The free-form surface is represented and can be exported either in NURBS form or as cubic triangular Bezier patches. Key characteristics of the approach are: (1) mesh pieces and surface pieces are related by strictly local averaging rules; (2) refinement rules depend only on direct, coarser-level ancestors and not on adjacent submeshes or patches; (3) submeshes at different levels look alike. The underlying data structure is a single winged-edge structure with additional pointers to support the hierarchy. Multiply refined regions may be directly adjacent to unrefined regions, and mesh fragments at different levels of refinement can be connected. CR Categories: I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling—Curve, surface, solid, and object representations,Splines; I.3.6 [Computer Graphics]: Methodology and Techniques—Graphics data structures and data types 1 Motivation, results and prior work Conventional modeling systems require precise specification of final features, e.g., distance in datum planes, positioning of control vertices with respect to one another, etc. Besides requiring a certain mindset not always compatible with creative, exploratory design, this is modeling at the highest level of refinement: a change in the overall shape of the model entails modifying all detail, however small, associated with the area of change. Not only is this time-consuming, but repeated detail work puts an undue strain on creative, exploratory design. Parameterized modeling, which allows initially precisely instantiated specifications to be varied, has been used in commercial solid modeling systems. A complementary paradigm that supports changes at different levels in the design gonzalez@acm.org yjorg@cise.ufl.edu zSupported by NSF National Young Investigator award 9457806-CCR history is hierarchical refinement. Levels express the designer’s intent by encoding the structure and dependence of geometric features. Several recent modeling environments are based on hierarchical modeling [10, 7, 21, 13]. Localized-hierarchy Surface Splines (LeSS) extend this modeling paradigm by combining localized mesh hierarchies with the ability to change topology by adding a handle or punching a hole. LeSS automatically maintains tangent continuity across the surface where wanted and defines the surface in standard form either as bicubic NURBS or cubic triangular Bezier patches. Non-standard properties of LeSS are: 1. The surface is represented explicitly in piecewise polynomial or spline form, and not as surface patches offset from ancestor surfaces. Finer-level patches replace coarse-level patches. Changes are propagated from coarse to fine inside an underlying mesh hierarchy: fragments (pieces with boundary) of finer meshes are represented as offsets in local coordinate frames of coarser meshes. This has several implications: The evaluation of the spline patches does not require traversal of the hierarchy and is numerically stable. Explicit bounding boxes and convex hulls for the surface are directly available and the moments of the enclosed object can be efficiently computed. The influence of changing coarser levels is transmitted by local reevaluation of the mesh hierarchy. Local topology changes are possible because there is no direct 1-1 correspondence of the coarse to the finer-level surface. 2. LeSS is based on a strictly localized hierarchy: the default position of a finer level mesh fragment before local editing is determined by local averaging rules that depend only on those coarser-level nodes from which the mesh nodes of the fragment are offset and not on adjacent mesh or patch information. This simplifies the data structures since no searches for neighbors at different levels of refinement are necessary. Repeatedly-refined patches can be placed directly next to unrefined patches. (Since the surface representation is explicit, the resolution of ‘crack problems’ where refined and unrefined patches meet can be delegated to the spline display routine.) 3. The representation is homogeneous: control structures everywhere, even across levels, look the same. Together with locality this homogeneity allows connecting arbitrary submeshes. This is useful since not all parts of a complex object must enter a strict dependence relation with all other parts – the designer’s choice of dependence can be only local and later decisions, say to join parts or create a hole (Figure 9), are not prevented by a rigid, global hierarchy.

The Gulf of St. Lawrence is the largest embayment on the Atlantic Coast of North America between Florida and the Labrador Sea. Surface geology, geophysical data, and subsurface well control define four structural units and extend them from the land areas across the gulf. Beds having possibilities for oil and gas production are present in early Paleozoic formations overlying Precambrian basement in the Anticosti basin and western Newfoundland fault belt; in middle Paleozoic beds overlying Taconic basement in the Gaspe fold belt; and in late Paleozoic formations overlying Acadian basement in the Maritimes basin. The southward continuation of this Acadian basement is exposed on various uplands in Nova Scotia, and also underlies the Sydney, Fundy, and other basins, and extends beneath the great Mesozoic and Tertiary sedimentary wedge of the Scotian basin of the continental shelf. Production has not been obtained from wells drilled in the Anticosti basin—a relatively stable monocline. Many gas and oil shows have been found throughout Ordovician rocks; Silurian rocks may be productive in deeper parts of the basin. The Gaspe fold belt is thrust over the south edge of the Anticosti basin and is overlapped by the north edge of the Maritimes basin. Many seeps and shows of oil, but only insignificant recovery, have come from structurally complex Devonian beds in the likely part of the fold belt. In places, the Silurian beds may be productive, too. The allochthonous or klippe sequence of the western Newfoundland fault belt has been complexly thrust over the autochthonous sequence at the eastern margin of the Anticosti basin. Porosity and shows have been found in Ordovician formations in both structural sequences; very minor oil production has been obtained from shallow wells. The Maritimes basin, a large complex intermontane basin of Middle Devonian to Permian sedimentary rocks, has one small oil and gas field producing from Mississippian sandstones. The presence, at the southeast, of a thick sedimentary section, sizeable structures, porous sandstones and carbonate rocks, many shows and seeps, oil shales, and albertite encourage exploration, especially in the Mississippian Horton and Windsor Groups and the Pennsylvanian “coarse-fluvial facies.” Although estimated values for oil and gas in place in these four structural units may be quite large, the Paleozoic age of the sedimentary section, the low-grade metamorphism in some parts of the area, the generally poor porosity at depth, and the large area under water where operations are very expensive all reduce the commercial potential.

Understanding and imaging the real seismo-tectonic of SW-Cairo zone are essential to eliminate the risk of this hazardous seismic zone which is surrounded by the highest condense population province in Egypt. In this study, up to date earthquake catalogue and aeromagnetic data of the area under study were analyzed for tracing active subsurface faults that are responsible of earthquakes activity. Using 2D-power spectrum techniques, depth of structures (faults) was estimated and a shallow intrusion was detected at 550 m. The tectonic framework of SW-Cairo was evaluated and discussed in view of seismicity, surface geology, subsurface structures from RTP maps and boreholes information and the active faults were traced at shallow and deep depths.

In a recent paper, Christensen et al. [1969] interpreted the gravity and seismic data from Mono Basin, California, to indicate that the depth to the basement there is 1–1.5 km, whereas Pakiser et al. [1960] and Pakiser [1968] interpreted the same data to indicate a depth of 5–5.5 km, with an acknowledged uncertainty of 1–1.5 km. Christensen et al. were led to their drastic revision of our earlier interpretations because the surface geology indicates to them that the basin is a broad, shallow downwarp rather than a deeply downfaulted structural block. A detailed re-examination of the available evidence indicates that their interpretation is incompatible with the geophysical data, but also that our interpretations require revision. Details of the geophysical data can be obtained from Pakiser et al. [1960], Pakiser [1968], and Johnson [1965].

We demonstrate a robust workflow for time-lapse gravity modeling in reservoir sequestration/production monitoring applications. This systematic approach outlines a reliable methodology to understanding the value and limitations of 4D gravity at a particular site, for both pre-acquisition decision making, and as a guide for post-data acquisition interpretation. To demonstrate, we present a multifaceted feasibility study for monitoring CO2 injection into a reservoir at various injection times using 4D micro-gravity method. The simulations are performed for a currently active CO2-EOR site, the Louisiana Delhi Field in the United State. We construct an accurate representation of the field directly from current seismic data, followed by application of binary inversion technology adapted to the time-lapse gravity problem and tailored to the specific site. Finally, we illustrate a method of resolution analysis to demonstrate the decreased recoverability of fluid movement at the site in the presence of varying data noise.

Stochastic surface models are useful for analyzing in situ roughness profiles and synthetic aperture radar (SAR) images of geological terrain. In this paper, two different surface models are discussed: surfaces with a stationary random roughness (conventional model) and surfaces with a power-law roughness spectrum (fractal model). In the former case, it must be considered that for short profiles (L<200l/sub 0/), the measured values of rms-height s and correlation length l may be significantly smaller than the intrinsic values s/sub 0/ and l/sub 0/. In the latter case, rms-height and correlation length depend on the profile length L, and the surface is better characterized by slope and offset of the roughness spectrum (which are independent of L), The sensitivity of the SAR signature to variations in surface roughness parameters is evaluated by means of theoretical scattering models. For smoother geological surfaces such as most arid terrain types, single scattering is dominant, which means that the roughness parameters can be determined from SAR data using comparatively simple algorithms. Multiple scattering processes on rough surfaces such as a'a lava and variations of the local incidence angle due to large-scale terrain undulations make the retrieval of roughness parameters by means of inverse modeling much more complex. Field data of surface roughness indicate that rougher geological surfaces may be in the diffractal regime at higher radar frequencies, in which the scattering characteristics deviate significantly from the patterns observed for stationary surfaces. On the basis of surface and scattering models, recently published observations of roughness data and radar signatures from volcanic, alluvial, and arid surfaces are examined.

We present a general scheme for 3D geophysical inversion using a deep learning convolutional neural network that enables three-dimensional inversions of useful size to be solved on laptops and desktops. A priori constraints of maximum smoothness or compactness on model parameters used during conventional geophysical inversion are not necessary. In environments where a single stabilizing functional is not capable of adequately representing the subsurface variation (e.g., steel cased wells, buried pipes in smoothly varying geology), the method provides an attractive alternative. Projecting data to the Hilbert space of model parameters using the adjoint operator before training mitigates non uniqueness associated with the underdetermined nature of traditional 3D geophysical inverse problems. A field inversion example using magnetic data collected in Washington -on- Brazos state historic site to image buried pipe and other infrastructure is presented. While the training for a single inversion takes longer, the results show better resolution of the top of the structure and overall shape compared to conventional minimum structure inversion methods. For multiple inversions such as for systems with moving footprints, or time lapse surveys, neural networks can be substantially faster than traditional inversion as the network needs to be trained only once for several successive model predictions. The method can be applied to other geophysical data including seismic, electromagnetic and gravity, at various scales and resolution.

ABSTRACTGravity surveys of the past century established that mountains have roots, seismic refraction lines shot in the second half of this century confirmed the downbulge of the Moho under the Alps, and recent reflection traverses provided new details on the behaviour of crustal layers in the deep part of the Alps. However, geophysical data are ambiguous geologically. For models of the root in terms of rock distribution to be tectonophysically acceptable, they must be the retrodeformable result of kinematic sequence that fits the geological surface data. For a cross‐section through the Swiss Alps based on refraction data and somewhat modified by the recent reflection traverses, a kinematic model compatible with large‐scale geological data may be obtained by the superposition of three Neogene phases with alternating vergence. Although Alpine collision is largely dextrally compressive in the central Alps, the N‐S component may be discussed in a cross‐section. Particularly puzzling geophysical features include a high‐velocity body in the middle crust and the disappearance of the layered foreland crust in the root. In order to account for these phenomena, it is proposed that the crustal root is interpreted as the result of complex reshuffling of middle and lower crustal masses as well as large‐scale phase transformations. The mid‐crustal highvelocity body is interpreted as a delaminated section of the lower crust of the Adria plate that was wedged into the middle crust of the Alps in the middle Miocene. The disappearance of the foreland lower crust is attributed to eclogitization attendant on the subduction of continental crust. Material balance estimates suggest that during Alpine collision large volumes of continental crust have disappeared through subduction.

A highly accurate, single-pass, unbiased frictional contact algorithm for higher-order elements based on the concept of midplane is presented. Higher-order elements offer a lucrative choice for contact problems as they can better represent the curvature of original geometries compared to the first-order elements. Compressive and frictional contact constraints are applied over the contact pairs of sub-segments obtained by the subdivision of higher-order segments. The normal traction depends upon the penalisation of true interpenetration, and frictional traction depends upon relative sliding between sub-segments over their shared patches. The midplane constructed by linearised subfacets can be corrected to account for the local curvature of interacting physical surfaces. Demonstrated through multiple tests, the use of higher-order elements surpasses the accuracy of first-order elements for curved geometries. Its versatility extends from static to dynamic conditions for flat and curved interfaces, including frictional contact. The presented examples include contact patch test, Hertzian contact, elastic collision, rotation of concentric surfaces, frictional sliding, self-contact and inelastic collision problems. Here, the contact patch test matches the accuracy of finite elements, and Hertzian contact shows a smoother solution compared to first-order meshes. The elastic collision problem highlights the utility of the algorithm in accurate prediction of configuration changes in multibody systems. The frictional sliding demonstrates the ability to represent the expected nonlinear distribution of nodal forces for the higher-order elements. The large deformation problems, e.g. self-contact and inelastic collision, specifically benefit from the accuracy in surface representation using higher-order discretisation and continuous contact constraint imposition on such surfaces during deformation.