Interwell Saturation Prediction by Artificial Intelligence Analysis of Well Logs

Well log analysis, through deploying advanced artificial intelligence (AI) algorithms, is key for wellbore geological studies. By analyzing different well characteristics with modern AI tools it becomes possible to estimate interwell saturation with improved accuracy, outlining primary fluid channels and saturation propagations in the reservoirs interwell region. The development of modern deep learning and artificial intelligence methods allows analysts to predict interwell saturation as a function of observed data in the near wellbore logged geological layers. This work addresses the use of deep neural network architectures as well as tensor regression models for predicting interwell saturation from other well characteristics, such as resistivity and porosity, as well as local near-well saturation. Several algorithms are compared in terms of both accuracy and computational efficiency. Sensitivity analysis for model parameters is carried out, which is based on the wells’ geometry, radius, and multiple sampling techniques. Additionally, the impact of local saturation prior knowledge on the model accuracy is analyzed. A reservoir box model encompassing volumetric interwell porosity, resistivity and saturation data was utilized for the validating and testing of the AI algorithms. A prototype is developed with Python 3.6 programming language.

Investigation of waterflood front digitations during immiscible displacements in porous media

In this work, unstable displacements were conducted using special equipment designed to run in-situ CT-scanner experiments. All the displacements were conducted on a homogeneous Bentheimer sandstone plug, of 10 cm in diameter and 40 cm in length. Digitations (or fingering) have been observed under varying conditions of injection flowrate, displaced fluid viscosity, and core wettability. They have been characterized at both the core scale, using the core average oil saturation and the water breakthrough; and at the local scale, using the local saturations and had-hoc image processing analysis. It was found that the effect of the different flowing conditions on the front digitations could not be interpreted independently. The oil recovery at brine breakthrough showed a good correlation with the viscous fingering number for the water-wet case. However, a different scaling was observed for the oil-wet case. The interplay of the different flowing conditions mitigates the possibility of constructing a unique scaling number to account for all experimental condition. The local saturation monitoring has provided a new insight to characterize the finger shapes and analyze the production mechanisms. It allowed to distinguish two independent contributions to early breakthrough: viscous dominated digitations and capillary dominated digitations. A two-phases diagram has been constructed to plot and compare these contributions for all flowing conditions. Their evolutions show the main production mechanisms during the flooding. We observed that the viscous digitations were not causing phase trapping at core scale: the core is completely swept after breakthrough. For the water-wet case, we found that the local oil recovery of swept zone remained constant before and after breakthrough while for the oil-wet case it is improving during all the water flooding process.

Local saturation and square everywhere

We show that it is consistent relative to a huge cardinal that for all infinite cardinals [Formula: see text], [Formula: see text] holds and there is a stationary [Formula: see text] such that [Formula: see text] is [Formula: see text]-saturated.

A fast FFT method for 3D pore-scale rock-typing of heterogeneous rock samples via Minkowski functionals and hydraulic attributes

The integration of numerical simulation and physical measurements, e.g. digital and conventional core analysis, requires the consideration of significant sample sizes when heterogeneous core samples are considered. In such case a hierarchical upscaling of properties may be achieved through a workflow of partitioning the sample into homogeneous regions followed by characterization of these homogeneous regions and upscaling of properties. Examples of such heterogeneities are e.g. fine laminations in core samples or different micro-porosity types as consequence of source rock components and diagenesis. In this work we utilize regional measures based on the Minkowski functionals as well as local saturation information derived through a morphological capillary drainage transform as a basis for such a classification/partitioning. An important consideration is the size of the measurement elements utilized, which could be considerable in the case of larger heterogeneities; in such case the calculation of the regional measures can be computationally very expensive. Here we introduce an FFT approach to calculate these measures locally, utilizing their additivity. The algorithms are compared against direct summation techniques and shift-overlap approaches for a selection of different averaging supports to illustrate their speed and practical applicability. We consider a range of artificial Boolean models to illustrate the effect of including hydraulic information on the resulting classifications scheme. This allows the determination of bias, since for these model systems local classes are known ab-initio. The classification framework is tested by comparing to the known initial micro-structure distribution and relative bias quantified in terms of choice of averaging elements (size and shape). Importantly, depending on the actual morphological transition between micro-type partitions, partitions including hydraulic attributes differ from pure morphological partitions with applications to electrofacies and hydraulic unit definitions.

Assessing blue orchard bee (Osmia lignaria) propagation and pollination services in the presence of honey bees (Apis mellifera) in Utah tart cherries

Osmia lignaria is a commercially available, native solitary bee species recognized for its propensity to forage upon and pollinate tree fruit crops such as apple, almond and cherry. This study evaluated the implementation of O. lignaria co-pollination with honey bees in central Utah commercial tart cherry orchards during 2017 and 2018 bloom. Three paired 1.2 ha sites were selected for evaluation of cherry fruit set and yield with and without managed O. lignaria releases alongside the standard honey bee hive stocking rate of 2.5 hives/ha. Osmia lignaria supplementation did not measurably increase cherry fruit set, fruit per limb cross-sectional area or fruit weight. The lack of differences in yield is likely a consequence of local saturation of pollinator services supplied by managed honey bees throughout experimental orchards, such that no additive benefit of managed O. lignaria releases were measurable. An increase in managed O. lignaria populations was achieved in 2017 but not 2018, possibly due to unknown changes to orchard management or environmental factors. While flying O. lignaria in Utah tart cherries may support sustainable in-field bee propagation, their subsequent impacts on tart cherry yield were not detected when paired with standard stocking densities of honey bees.

Elementary subdomain technique for magnetic field calculation in rotating electrical machines with local saturation effect

Purpose The most significant point to be introduced in the subdomain technique (i.e. based on the formal resolution of Maxwell’s equations applied in subdomain) is the local saturation effect. This paper aims to present a novel contribution on the improvement in the two-dimensional (2-D) technique in polar coordinates by focusing on the local saturation. Design/methodology/approach The rotor and stator regions are divided into elementary subdomains (E-SDs) which are characterized by general solutions to the first harmonic of magnetostatic Maxwell equations. These E-SDs are connected in the two directions (i.e. r- and θ-edges). Newton–Raphson (NR) iterative algorithm is used for nonlinear magnetic field analysis. Findings The proposed model is relevant for different types of rotating electrical machines; as an example, the semi-analytical model has been implemented for spoke-type permanent-magnet (PM) machines (STPMMs). The magnetic field calculations have been performed for nonlinear B(H) curve and compared to nonlinear finite element method (FEM) predictions. The semi-analytic results are in good agreement with those obtained numerically, considering both amplitude and waveform. Originality/value A new model for full prediction of magnetic field in the rotating electrical machines with the local saturation effect is presented.

Simulating the boiling fluid outflow taking into account external condensation

This paper considers the axisymmetric problem of condensation at a boiling liquid outflow in a closed region filled with steam. It has been found that the unsteady condensation is due to the formation of a complex structure of lateral pressure surges, Mach disk, barrel shocks and a wall, on which the pressure exceeds the local saturation pressure.

On Numerical Investigation of Water Injection to Screw Compressors

Oil injection is widely used in screw compressors for lubrication, sealing and cooling purposes. More recently other, mainly lower viscosity fluids are used for the purpose, for example water. Water introduces new phenomena into the screw compressor process, one among them is evaporation. 3D numerical modelling is employed and presented in this paper for the detailed analysis of flow and thermodynamics process during injection of water in screw compressors. The advantage of such simulations is that realistic geometry of the rotors and the ports can be captured. In addition, the physical effects of fluid thermal interactions and leakage are directly taken into account by these models. Recent studies have shown that for oil free and oil injected air compressors a good agreement is achieved with measurements, in prediction of performance parameters. In these simulations the Eulerian-Eulerian multiphase modelling has been applied. To implement the same model for water injected compressors presents an additional challenge as the liquid water injected into the compression chamber changes phase and evaporates depending on the local saturation and thermodynamic conditions. Water also forms liquid film on the rotors and housing and thereby influences thermal changes. In this paper a numerical model for water injected screw compressor that accounts for evaporation effects has been presented. Empirical form of the Lee (9) evaporation-condensation model for phase change has been applied in the compression chamber using the phase specific mass and energy sources. Calculation of the amount of water required to just saturate the compressed air at delivery pressure is used to set the mass flow rate of water at two operating speeds. The effect of the suction air temperature and relative humidity is studied. Evaporation inside compression chamber has two important physical effects, one is that the latent heat of evaporating water lowers the gas temperature and the other is the change of state from water to vapour. Including vapour as a third phase adds complexity to already challenging deforming grids required for screw domains. Hence a mass and energy source formulation is proposed in the presented study to account for the vapour phase change and evaporation effects, thus limiting the number of phases to be modelled. Local drop in gas temperature, distribution of water and regions of evaporation were identified by the simulations. Thermal hot spots on the rotor were located. Reduction in the leakage of gas and its exit temperature was well predicted by the model. Such simplified evaporation model can be further used in the design of water injected screw compressors and extended to predict thermal deformation of the rotors and the housing.