According to the characteristics of overpressure carbonate reservoirs Kenkiyak pre-salt oil field in littoral Caspian Basin, cores were made artificially and stress sensibility matrix with non-packed, semi-packed fully packed fractures analyzed. The method gas measurement was used experiment. confining pressure sample first increased then decreased. When became stable, porosity permeability data from each point obtained analyze cores. results showed that these four types a descending order...

The carboniferous carbonate reservoirs in the North Truva Oilfield have undergone complex sedimentation, diagenesis and tectonic transformation. Various reservoir spaces of pores, caves fractures, with strong heterogeneity diverse pore structures, been developed. As a result, quantitative description structure is difficult, accuracy logging identification prediction low. These pose lot challenges to classification evaluation as well efficient development reservoirs. This study based on...

The 3D geological model of dual porosity reservoirs in the Kenkiyak pre-salt oilfield is built step by based on detailed core observation, employing data, log and seismic data. Firstly, structural built. Then sedimentary facies basis analyzing electro-facies, models reservoir matrix properties controlled are structure analysis methods optimization. Secondly, fracture distribution established applying results property description. Thirdly, three-dimensional integrated set up combining with...

Zearalenone (ZEN) is a mycotoxin produced by Fusarium spp., which commonly and severely contaminate food/feed. ZEN affects food/feed safety reduces economic losses owing to its carcinogenicity, genotoxicity, reproductive toxicity, endocrine effects, immunotoxicity. To explore efficient methods detoxify ZEN, we identified characterized an ZEN-detoxifying microbiota from the culturable microbiome of Pseudostellaria heterophylla rhizosphere soil, designated Bacillus amyloliquefaciens D-1. Its...

Taking the Kenjiyak Oilfield in Kazakhstan as an example, fracture characteristics and genetic mechanism of carbonate reservoirs with overpressure were discussed. The study showed that are formed under developed hydraulic fractures besides tectonic karstic fractures. Carboniferous sub-salt Kenkiyak mainly Mesozoic Cenozoic, on a macroscopic level they cut rocks into brecciated texture or packed by calcite veins, microscopic radiant distribution around pores. formation evolution system region...

Traditional supervised voice activity detection (VAD) methods work well in clean and controlled scenarios, with performance severely degrading real-world applications.One possible bottleneck is that speech the wild contains unpredictable noise types, hence frame-level label prediction difficult, which required for traditional VAD training.In contrast, we propose a general-purpose (GPVAD) framework, can be easily trained from noisy data weakly fashion, requiring only clip-level labels.We...

The aim of this paper is to illustrate the impact stress-sensitive permeability on productivity in naturally fracture carbonate reservoirs. First, stress sensitivity evaluation experiments were conducted artificial cores with matrix, uncemented fracture, partially- cemented and fully-cemented by measuring their process increasing confining pressure. Then, a new mathematical model for prediction fractured reservoir consideration developed. Comparisons have been made between results field...

This paper focuses on fast eye localization method. According to the priori proportional relationships of face features, we firstly set an appropriate candidate window from region detected. Secondly, histogram equalization is applied eliminate illumination effects. presents a new haar-like feature generating confidence throughout in order locate eyeball accurately and rapidly. Our method proved be simple robust against disturbance caused by glasses, eyebrow hair. The process training...

To accelerate and compress deep neural networks (DNNs), many network quantization algorithms have been proposed. Although the strategy of any algorithm from state-of-the-arts may outperform others in some architectures, it is hard to prove always better than others, even cannot judge that best choice for all layers a network. In other words, existing are suboptimal as they ignore different characteristics quantize by uniform strategy. solve issue, this paper, we propose differentiable search...

Abstract The genesis of the overpressure carbonate reservoir is not controlled only by lithofacies, karstification and tectonic movement, but also overpressure. hydraulic fractures formed under its connection expansion to early solution vugs structural are key factors high productive zones. Kenkiyak Pre-salt Oilfield with a burial depth more than 4300 meters. matrix permeability 0.98md. It was considered as very low porous-type defined oilfield without economic benefit. According this study,...

Dictionary learning and deep both have played important roles in computer vision. strives to learn best representative atoms reconstruct the source images or signals. Alternating iterative methods are developed solve such problems. On other hand, performs an end-to-end mode, where feature extraction classification achieved simultaneously. Obviously, scheme of is different from that traditional dictionary which shallow focuses more on data reconstruction. However, studies building a...

Convolutional Neural Networks (CNNs) are getting larger and deeper, thus becoming harder to be deployed on systems with limited resources. Though convolutional filters benefit from the concept of receptive field, they still take up lots resources store these parameters in large amounts filters. Therefore, a compression method pre-trained CNN models using "Linear Representation" kernels is introduced this paper. First, codebook template "K_t". generated by conducting unsupervised...

Traditional supervised voice activity detection (VAD) methods work well in clean and controlled scenarios, with performance severely degrading real-world applications. One possible bottleneck is that speech the wild contains unpredictable noise types, hence frame-level label prediction difficult, which required for traditional VAD training. In contrast, we propose a general-purpose (GPVAD) framework, can be easily trained from noisy data weakly fashion, requiring only clip-level labels. We...

Electromagnetic wave simulation is of pivotal importance in the design and implementation photonic nano-structures. In this study, we developed a lattice Boltzmann model with single extended force term (LBM-SEF) to simulate propagation electromagnetic waves dispersive media. By reconstructing solution macroscopic Maxwell equations using equation, final form only involves an equilibrium non-equilibrium term. The two terms are evaluated variables effect, respectively. LBM-SEF scheme capable...

Jurassic strata in the ST gas field of northern West Siberia Basin have been regarded as a potential exploration target with undiscovered hydrocarbon resources. However limited research has performed on sequence stratigraphy strata, well its sandstone distribution controlled by variable sea level change and sediment input. In this paper, four third-order sequences (SQ1, SQ2, SQ3, SQ4) nine fourth-order for are interpreted based seismic facies analysis lithology stacking patterns seven wells....

Voice activity detection (VAD) is an essential pre-processing step for tasks such as automatic speech recognition (ASR) and speaker recognition. A basic goal to remove silent segments within audio, while a more general VAD system could all the irrelevant noise even unwanted from non-target speakers. We define task, which only detects target speaker, speaker-dependent voice (SDVAD). This task quite common in real applications usually implemented by performing verification (SV) on audio...