Fine Stage Research Articles

The Influence of Crystal Orientation on Subsurface Damage of Mono-Crystalline Silicon by Bound-Abrasive Grinding.

Subsurface damage (SSD) produced in a grinding process will affect the performance and operational duration of single-crystal silicon. In order to reduce the subsurface damage depth generated during the grinding process by adjusting the process parameters (added), experiments were designed to investigate the influence of machining factors on SSD. This included crystal orientation, diamond grit size in the grinding wheel, peripheral speed of the grinding wheel, and feeding with the intention to optimize the parameters affecting SSD. Compared with isotropic materials such as glass, we considered the impact of grinding along different crystal directions <100> and <110> on subsurface damage depth (added). The Magnetorheological Finishing (MRF) spot technique was used to detect the depth of SSD. The results showed that the depth of SSD in silicon increased with the size of diamond grit. SSD can be reduced by either increasing the peripheral speed of the grinding wheel or decreasing the feeding rate of the grinding wheel in the <100> crystal orientation, if the same size of diamond grit was employed. In addition, we proposed a modified model around surface roughness and subsurface crack depth, which considered plastic and brittle deformation mechanisms and material properties of different crystal orientations. When the surface roughness (RZ) exceeded the brittle-plastic transition’s critical value RZC (RZC<100> > 1.5 μm, RZC<110> > 0.8 μm), cracks appeared on the subsurface. The experimental results were consistent with the predicted model, which could be used to predict the subsurface cracks by measuring the surface roughness. However, the model only gives the approximate range of subsurface defects, such as dislocations. The morphology and precise depth of plastic deformation subsurface defects, such as dislocations generated in the fine grinding stage, needed to be inspected by transmission electron microscopy (TEM), which were further studied.

Restoration of Missing Patterns on Satellite Infrared Sea Surface Temperature Images Due to Cloud Coverage Using Deep Generative Inpainting Network

In this paper, we propose a novel deep generative inpainting network (GIN) trained under the framework of generative adversarial learning, which is optimized for the restoration of cloud-disturbed satellite sea surface temperature (SST) imagery. The proposed GIN architecture can achieve accurate and fast restoration results. The proposed GIN consists of rough and fine reconstruction stages to promote the details and textures of missing (clouded) regions in SST images. We also propose a nov el preprocessing strategy that replaces the land areas with the average value of daily oceanic surface temperatures for improving restoration accuracy. To learn the proposed GIN, we developed a novel approach that combines multiple loss functions well suited for improving the restoration quality over missing SST information. Our results show that the difference in temperature between restored and actual satellite image data was no larger than 0.7 °C in monthly average values, which suggests excellent resilience against the missing sea surface temperature data. The proposed GIN has a faster restoration time and is feasible for real-time ocean-related applications. Furthermore, the computational cost of restoring SST images is much lower than the popular interpolation methods.

초정밀 직선 스테이지의 6자유도 운동 오차 보정을 위한 미세 구동 스테이지 개발 및 성능 평가

In ultra-precision processes, such as aerospace parts and precision mold machining, the accuracy of a feed drive system should be secured to achieve sufficient form accuracy. Dual-Servo stages, which compensate for multi-DOF motion errors, are being developed depending on the applied processes. This paper deals with the fine stage of a dual-servo stage to compensate for 6-DOF motion errors of a linear stage. The proposed fine stage measured 6-DOF errors of the linear stage motion with capacitive sensors, a reference mirror, and an optical encoder. It compensated for the errors using the flexure hinge mechanism with piezo actuators. The error equations and the inverse kinematics were derived to calculate the 6- DOF errors and displacements of piezo actuators for 6-DOF motions, respectively. Performance evaluation was implemented to verify feasibility of the developed fine stage of the fabricated dual-servo stage. Through the step response test of the fine stage, compensation resolutions for the translational and the rotational motion were confirmed to be less than 10 nm and 1/3 arcsec, respectively. The 6-DOF motion errors in the verification test were reduced by 73% on average.

A relic sketch extraction framework based on detail-aware hierarchical deep network

As the first step of the restoration process of painted relics, sketch extraction plays an important role in cultural research. However, sketch extraction suffers from serious disease corrosion, which results in broken lines and noise. To overcome these problems, we propose a deep learning-based hierarchical sketch extraction framework for painted cultural relics. We design the sketch extraction process into two stages: coarse extraction and fine extraction. In the coarse extraction stage, we develop a novel detail-aware bi-directional cascade network that integrates flow-based difference-of-Gaussians (FDoG) edge detection and a bi-directional cascade network (BDCN) under a transfer learning framework. It not only uses the pre-trained strategy to extenuate the requirements of large datasets for deep network training but also guides the network to learn the detail characteristics by the prior knowledge from FDoG. For the fine extraction stage, we design a new multiscale U-Net (MSU-Net) to effectively remove disease noise and refine the sketch. Specifically, all the features extracted from multiple intermediate layers in the decoder of MSU-Net are fused for sketch predication. Experimental results showed that the proposed method outperforms the other seven state-of-the-art methods in terms of visual and quantitative metrics and can also deal with complex backgrounds.

A 3 mW 6-bit 4 GS/s Subranging ADC With Subrange-Dependent Embedded References

A subranging analog-to-digital converter (ADC) with reference-embedded comparators (RECs) is proposed. By adjusting the bias current and/or body voltage of the REC's input differential pair, the REC offset can be adjusted to a specific voltage equal to a reference voltage referred to henceforth as the embedded reference. For the ADC's coarse stage, RECs with wide-range embedded references are implemented by adjusting the bias currents using current source arrays to cover the full-scale input. By contrast, for the ADC's fine stage, RECs with narrow-range embedded references are implemented by adjusting the body voltage. In addition, the centers of the embedded references in the different ADC subranges are created by current source arrays, which are digitally scaled according to the coarse ADC's output codes. As a result, the reference-voltage-switching network used in conventional subranging ADCs is not required, and hence the speed of the ADC is increased. Moreover, to eliminate the effects of process variation, the bias currents and body voltages in the RECs are calibrated with an auxiliary resistor ladder. After the calibration, the resistor ladder is removed. Consequently, no resistor ladder is used during normal operation, which greatly saves power. A 3 mW 6-bit 4 GS/s REC-based subranging ADC is implemented in 28-nm CMOS technology. With a near Nyquist frequency input, the ADC achieves SNDRs of 31.8 dB and 30.7 dB at 3.6 GS/s and 4 GS/s, respectively. Moreover, at 3.6 GS/s, the ADC has a Walden Figure-of-Merit (FoM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">W</sub> ) of 22.7 fJ/conv-step, which is the best compared with prior state-of-the-art 6-bit high-speed ADCs.

A Fully Synthesizable Fractional-N MDLL With Zero-Order Interpolation-Based DTC Nonlinearity Calibration and Two-Step Hybrid Phase Offset Calibration

In this paper, a fully-synthesizable digital-to-time (DTC)-based fractional-Nmultiplying delay-locked loop(MDLL) is presented. Noise and linearity of synthesizable DTCs are analyzed, and a two-stage synthesizable DTC is proposed in which a path-selection DTC is used as the coarse stage and a variable-slope DTC is used as the fine stage. To calibrate the DTC nonlinearity, a highly robust zero-order interpolation based nonlinearity calibration is proposed. Besides, the static phase offsets (SPO) between bang-bang phase detector (BBPD) and multiplexer (MUX) are calibrated by a proposed hybrid analog/digital phase offset calibration, while the dynamic phase offsets (DPO) are removed by a proposed complementary switching scheme. The co-design of the analog circuits and digital calibrations enable excellent jitter and spur performance. The MDLL achieves 0.70 and 0.48ps root-mean-square (RMS) jitter in fractional-N and integer-N modes, respectively. The fractional spur is less than -59.0dBc, and the reference spur is -64.5dBc. The power consumptions are 1.85mW and 1.22mW, corresponding to figures of merit (FOM) of -240.4dB and -245.5dB.

Quality parameters in linseed (Linum usitatissimum L.) varieties grown under sodic soil

The present study entitled “Evaluation of fatty acid profile and nutritional quality of some linseed (Linum usitatissimum L.) varieties grown under sodic soil” was conducted at Agronomy Research Farm and analysis was carried out in the Laboratory of Department of Agricultural Biochemistry, A.N.D. University of Agriculture and Technology, Kumarganj, Ayodhya (U.P.) during Rabi season in the year 2017-18 and 2018-19. Ten varieties of linseed NDL-1 NDL-3 Parvati, Shikha, Shubhra, Shekhar, Chambal, T-397, Mukta, and Garima, were collected from department of Genetics and plant breeding of this University, and used as experimental material in the field trail. The seeds of linseed varieties were sown in Completely Randomized Design with three replications on 2017-18 and 2018-19. The row to row and plant to plant spacing was kept 10 cm and 30 cm, respectively. The seeds were sown at the rate of 30-40 kg per hectare. All agronomical practices were adopted to achieve a good crop. After harvesting the seeds of each variety were collected separately and used for the analysis various physical parameters. The seeds of each varieties were ground into fine powder stage and stored in air tight container for further physical parameters such as number of pods per plant, 1000-seed weight, colour of seeds, yield per plant,. The data obtained in the experiment showed the Morphological characters such as number of pods per plant 59.80 to 45.50 pods and 59.88 to 45.53 pods, 1000 seed weight 08.23 to 06.50 g and 8.27 to 06.53 g, in variety NDL-3 and Garima, seed yield per plant 8.70 to 06.80 (g) and 08.70 to 06.83(g) during year 2017-18 and 2018-19, respectively. Variety Garima and NDL-3, recorded maximum and yield parameter was superior in physical parameters.

A Sample Weight and AdaBoost CNN-Based Coarse to Fine Classification of Fruit and Vegetables at a Supermarket Self-Checkout

The physical features of fruit and vegetables make the task of vision-based classification of fruit and vegetables challenging. The classification of fruit and vegetables at a supermarket self-checkout poses even more challenges due to variable lighting conditions and human factors arising from customer interactions with the system along with the challenges associated with the colour, texture, shape, and size of a fruit or vegetable. Considering this complex application, we have proposed a progressive coarse to fine classification technique to classify fruit and vegetables at supermarket checkouts. The image and weight of fruit and vegetables have been obtained using a prototype designed to simulate the supermarket environment, including the lighting conditions. The weight information is used to change the coarse classification of 15 classes down to three, which are further used in AdaBoost-based Convolutional Neural Network (CNN) optimisation for fine classification. The training samples for each coarse class are weighted based on AdaBoost optimisation, which are updated on each iteration of a training phase. Multi-class likelihood distribution obtained by the fine classification stage is used to estimate a final classification with a softmax classifier. GoogleNet, MobileNet, and a custom CNN have been used for AdaBoost optimisation, with promising classification results.

The influence of the cross-modal emotional pre-preparation effect on audiovisual integration.

Previous studies have shown that the cross-modal pre-preparation effect is an important factor for audiovisual integration. However, the facilitating influence of the pre-preparation effect on the integration of emotional cues remains unclear. Therefore, this study examined the emotional pre-preparation effect during the multistage process of audiovisual integration. Event-related potentials (ERPs) were recorded while participants performed a synchronous or asynchronous integration task with fearful or neutral stimuli. The results indicated that, compared with the sum of the unisensory presentation of visual (V) and auditory (A) stimuli (A+V), only fearful audiovisual stimuli induced a decreased N1 and an enhanced P2; this was not found for the neutral stimuli. Moreover, the fearful stimuli triggered a larger P2 than the neutral stimuli in the audiovisual condition, but not in the sum of the combined (A+V) waveforms. Our findings imply that, in the early perceptual processing stage and perceptual fine processing stage, fear improves the processing efficiency of the emotional audiovisual integration. In the last cognitively assessing stage, the fearful audiovisual induced a larger late positive component (LPC) than the neutral audiovisual. Moreover, the asynchronous-audiovisual induced a greater LPC than the synchronous-audiovisual during the 400-550 ms period. The different integration effects between the fearful and neutral stimuli may reflect the existence of distinct mechanisms of the pre-preparation in terms of the emotional dimension. In light of these results, we present a cross-modal emotional pre-preparation effect involving a three-phase emotional audiovisual integration.

Deep Learning-Based Generic Automatic Surface Defect Inspection (ASDI) With Pixelwise Segmentation

Automatic surface defect inspection (ASDI) is a crucial and challenging problem in industry because it affects the quality and efficiency of production greatly. Deep learning-based methods achieve promising improvements for surface detection, but they rely on a massive training data set that is impractical in industry. In this study, we propose a generic method that works effectively even on the small size training data set. Especially, we introduce a ResMask generative adversarial network (GAN) framework that is a residual GAN to expand the insufficient defect data sets. Meanwhile, the existing inspection data sets are so much easier to detect than the data sets in the real industrial scenarios that new industrial surface defect data sets containing more diverse and challenging images are established. Then, a coarse-to-fine module (CFM) that consists of a coarse detection subnetwork and a fine segmentation subnetwork is proposed for the needs of fast detection for high-resolution images. In the coarse detection stage, spatial pyramid pooling (SPP) is utilized to increase the receptive field of the network, reduce the false detection rate, and determine the approximate location of defects. In the fine segmentation stage, the receptive field of the network is enlarged by atrous SPP (ASPP), and skip links that incorporate low-level with high-level features achieve pixelwise precision on defect segmentation. Finally, our algorithm has achieved state-of-the-art results in DAGM, HR, and WB data sets (0.859, 0.761, and 0.805, respectively) according to MIoU. It achieves an average processing time of 44.4 ms on the test images with a resolution of <inline-formula> <tex-math notation="LaTeX">$512 \,\,\times \,\, 512$ </tex-math></inline-formula> and 131.6 ms for <inline-formula> <tex-math notation="LaTeX">$2048 \times 2048$ </tex-math></inline-formula> images. On the DAGM data set, the detection accuracy of mean intersection over union (MIoU) reaches 0.859.

A Panoramic Localizer Based on Coarse-to-Fine Descriptors for Navigation Assistance.

Visual Place Recognition (VPR) addresses visual instance retrieval tasks against discrepant scenes and gives precise localization. During a traverse, the captured images (query images) would be traced back to the already existing positions in the database images, rendering vehicles or pedestrian navigation devices distinguish ambient environments. Unfortunately, diverse appearance variations can bring about huge challenges for VPR, such as illumination changing, viewpoint varying, seasonal cycling, disparate traverses (forward and backward), and so on. In addition, the majority of current VPR algorithms are designed for forward-facing images, which can only provide with narrow Field of View (FoV) and come with severe viewpoint influences. In this paper, we propose a panoramic localizer, which is based on coarse-to-fine descriptors, leveraging panoramas for omnidirectional perception and sufficient FoV up to 360. We adopt NetVLAD descriptors in the coarse matching in a panorama-to-panorama way, for their robust performances in distinguishing different appearances, utilizing Geodesc keypoint descriptors in the fine stage in the meantime, for their capacity of detecting detailed information, formatting powerful coarse-to-fine descriptors. A comprehensive set of experiments is conducted on several datasets including both public benchmarks and our real-world campus scenes. Our system is proved to be with high recall and strong generalization capacity across various appearances. The proposed panoramic localizer can be integrated into mobile navigation devices, available for a variety of localization application scenarios.

A Coarse-to-Fine Framework for Multiple Pedestrian Crossing Detection.

When providing route guidance to pedestrians, one of the major safety considerations is to ensure that streets are crossed at places with pedestrian crossings. As a result, map service providers are keen to gather the location information about pedestrian crossings in the road network. Most, if not all, literature in this field focuses on detecting the pedestrian crossing immediately in front of the camera, while leaving the other pedestrian crossings in the same image undetected. This causes an under-utilization of the information in the video images, because not all pedestrian crossings captured by the camera are detected. In this research, we propose a coarse-to-fine framework to detect pedestrian crossings from probe vehicle videos, which can then be combined with the GPS traces of the corresponding vehicles to determine the exact locations of pedestrian crossings. At the coarse stage of our approach, we identify vanishing points and straight lines associated with the stripes of pedestrian crossings, and partition the edges to obtain rough candidate regions of interest (ROIs). At the fine stage, we determine whether these candidate ROIs are indeed pedestrian crossings by exploring their prior constraint information. Field experiments in Beijing and Shanghai cities show that the proposed approach can produce satisfactory results under a wide variety of situations.

Multimodal affine registration for ICGA and MCSL fundus images of high myopia.

The registration between indocyanine green angiography (ICGA) and multi-color scanning laser (MCSL) imaging fundus images is vital for the joint linear lesion segmentation in ICGA and MCSL and the evaluation whether MCSL can replace ICGA as a non-invasive diagnosis for linear lesion. To our best knowledge, there are no studies focusing on the image registration between these two modalities. In this paper, we propose a framework based on convolutional neural networks for the multimodal affine registration between ICGA and MCSL images, which contains two parts: coarse registration stage and fine registration stage. In the coarse registration stage, the optic disc is segmented and its centroid is used as a matching point to perform coarse registration. The fine registration stage regresses affine parameters directly using jointly supervised and weakly-supervised loss function. Experimental results show the effectiveness of the proposed method, which lays a sound foundation for further evaluation of non-invasive diagnosis of linear lesion based on MCSL.

Analysis, Design, and Implementation of Junction Temperature Fluctuation Tracking Suppression Strategy for SiC MOSFETs in Wireless High-Power Transfer

In order to improve system reliability and reduce device thermal fatigue failure in multiload wireless power transfer (WPT) systems for electric vehicles, the implementation of the junction temperature fluctuation suppression strategy for SiC MOSFETs is necessary. However, current methods are relatively lacking, and active thermal management has not been used in WPT systems. In this article, the relationship between circuit parameters and junction temperature of SiC MOSFET is analyzed. In particular, a junction temperature fluctuation tracking suppression strategy for SiC MOSFETs is proposed, which consists of a coarse and a fine adjustment stage. As for the former, a shunt capacitor bank switching method is implemented. In order to compensate for the coarse adjustment defects in the small adjustment range and poor effect, in fine adjustment stage, a changing driving voltage method is used. Finally, a 5-kW multiload WPT system is built for verification. Experimental results show that the proposed strategy has obvious effect on the suppression of junction temperature fluctuation and keeps temperature near the target temperature. Benefiting from this, the maximum 13.9 °C junction temperature fluctuation is completely eliminated when the power fluctuations are within 36.1% of the rated power, and the heat load of each SiC MOSFET can be independently adjusted.

Two-stage sentiment classification based on user-product interactive information

Document-level review sentiment classification aims to predict the sentiment category for given review documents written by users for products. Most of the existing methods focus on generating a good review document representation and classifying the review document directly. However, on the one hand, as document-level review sentiment classification usually includes many sentiment categories and the difference between these sentiment categories is not obvious, it may be difficult to obtain satisfying result by direct classification. On the other hand, this once classification process with review representation may fail to well interpret how the results are achieved. In addition, although some information such as user preference and product characteristics are incorporated when building models, the interactive information between user and product are usually ignored. In this paper, inspired by the deductive reasoning strategy of human doing multiple choice questions, we are motivated to propose a Two-Stage Sentiment Classification (TSSC) model to classify review documents in two stages: (1) Coarse classification stage, where model mainly adopts user-product interactive information to pre-judge the sentiment tendency of the review document without considering the review information; (2) Fine classification stage, where model uses text information of the review document for further classification based on the sentiment tendency obtained in coarse classification stage. Finally, the sentiment classification task is accomplished by combining both the results of coarse classification and fine classification. The experimental results demonstrate that our TSSC model significantly outperforms most of the related models (e.g., Trigram and NSC+UPA) on IMDB and Yelp datasets in terms of classification accuracy. When compared with the state-of-the-art HUAPA model, our TSSC model not only achieves slightly more accurate performance, but also has lower time complexity and stronger interpretability.

Physical activity level and participation in strength training (ST) activities among undergraduate female students

Background: Insufficient physical activity has been identified as a leading risk factor for Non-Communicable Diseases and global mortality. Healthy adults should be involved in Strength Training (ST) activities to be in Health Enhancing Physical Activities (HEPA) level. Undergraduates are physically fit and in finest age to understand importance of physical activities. However, most of undergraduates in worldwide were found to be inactive and highest were women. Therefore, this study aimed to describe the physical activity level, participation in ST activities and perceived benefits and barriers for ST activities among female undergraduates. Methodology: Two hundred female undergraduates of Faculty of Science, University of Colombo were randomly selected to conduct this descriptive cross-sectional study. Interviewer administered questionnaire was used to collect the data. Physical activity level was assessed by the first part of the questionnaire. Section one of the second part of the questionnaire focused on individual perception of benefits and barriers to ST activities while section two was used to determine barrier scores. Data was analyzed using the Statistical Package for the Social Sciences (SPSS), version 21.0. Results: The mean age of the participants was 22.66. Majority (63.0%; n= 121) was physically inactive, 33.3% (n=64) were minimally active and only 3.6% (n=7) were in HEPA category. Fifty one participants (26.6%) showed zero MET min/week for walking, 64 participants (33.3%) for moderate activities and more than half (68.8%; n=132) for vigorous activities. Majority (89.1%; n= 171) were not ST participants and 87.5% (n= 168) perceived ST as an important activity. Health benefits were perceived as the most important benefit of ST by 44.8% (n=86). Lack of time was the mostly (30.7%; n=59) perceived restricting barrier for ST activities. Barrier scores were different in at least one physical activity level (p&lt; 0.01). Conclusion: These undergraduates appeared to be having a sedentary life style. Improved awareness on importance of ST and HEPA and time management skills will increase participation in ST which results increase the physical level of undergraduates.

Near-Duplicate Image Detection System Using Coarse-to-Fine Matching Scheme Based on Global and Local CNN Features

Due to the great success of convolutional neural networks (CNNs) in the area of computer vision, the existing methods tend to match the global or local CNN features between images for near-duplicate image detection. However, global CNN features are not robust enough to combat background clutter and partial occlusion, while local CNN features lead to high computational complexity in the step of feature matching. To achieve high efficiency while maintaining good accuracy, we propose a coarse-to-fine feature matching scheme using both global and local CNN features for real-time near-duplicate image detection. In the coarse matching stage, we implement the sum-pooling operation on convolutional feature maps (CFMs) to generate the global CNN features, and match these global CNN features between a given query image and database images to efficiently filter most of irrelevant images of the query. In the fine matching stage, the local CNN features are extracted by using maximum values of the CFMs and the saliency map generated by the graph-based visual saliency detection (GBVS) algorithm. These local CNN features are then matched between images to detect the near-duplicate versions of the query. Experimental results demonstrate that our proposed method not only achieves a real-time detection, but also provides higher accuracy than the state-of-the-art methods.

A 9-Bit 70-MS/s Two-Stage SAR ADC With Passive Residue Transfer

A two-stage successive approximation register analog-to-digital converter (ADC) is reported. The power consumption is reduced by a technique, which combines the charge sharing passive residue transfer and the V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">cm</sub> -based switching methods. It is shown this technique reduces the transfer and the fine stage switching energy by 75%. In addition, the noise, the linearity, and the settling time are analyzed. To verify the performance throughout PVT variations, the full circuit is simulated in a 180-nm CMOS process. According to the simulation results, the ADC achieves a signal-to-noise and distortion ratio of 51.27 dB and a spurious-free dynamic range of 64.36 dB. The total power consumption of the calibration-free ADC is 2.99 mW, and the achieved figure of merit is 143 fJ/conversion-step, at a sampling rate of 70 MS/s.

Coarse-to-Fine Adversarial Networks and Zone-Based Uncertainty Analysis for NK/T-Cell Lymphoma Segmentation in CT/PET Images.

Extranodal natural killer/T cell lymphoma (ENKL), nasal type is a kind of rare disease with a low survival rate that primarily affects Asian and South American populations. Segmentation of ENKL lesions is crucial for clinical decision support and treatment planning. This paper is the first study on computer-aided diagnosis systems for the ENKL segmentation problem. We propose an automatic, coarse-to-fine approach for ENKL segmentation using adversarial networks. In the coarse stage, we extract the region of interest bounding the lesions utilizing a segmentation neural network. In the fine stage, we use an adversarial segmentation network and further introduce a multi-scale L1 loss function to drive the network to learn both global and local features. The generator and discriminator are alternately trained by backpropagation in an adversarial fashion in a min-max game. Furthermore, we present the first exploration of zone-based uncertainty estimates based on Monte Carlo dropout technique in the context of deep networks for medical image segmentation. Specifically, we propose the uncertainty criteria based on the lesion and the background, and then linearly normalize them to a specific interval. This is not only the crucial criterion for evaluating the superiority of the algorithm, but also permits subsequent optimization by engineers and revision by clinicians after quantitatively understanding the main source of uncertainty from the background or the lesion zone. Experimental results demonstrate that the proposed method is more effective and lesion-zone stable than state-of-the-art deep-learning based segmentation model.

Influencing factors and application prospects of CO2 flooding in heterogeneous glutenite reservoirs

In recent years, large amounts of CO2 emissions have led to many environmental problems. Developing a reasonable approach to carbon dioxide emissions is one of the most important issues for the sustainable development of human civilization. Compared with CO2 geological storage, CO2 flooding has greater advantages and a higher utilization efficiency. CO2 flooding technology has been successfully applied to many types of reservoirs, such as conventional sandstone reservoirs, low-permeability reservoirs, and carbonates. In this paper, the feasibility of CO2 flooding in glutenite reservoirs is studied. First, the glutenite reservoir is divided into fine stages, and reasonable vertical development units are determined. On this basis, the distribution scale, genetic environment and formation mode of the glutenite reservoir are characterized in detail and their spatial distribution rules are depicted in three dimensions. Then, the influencing factors of CO2 flooding in the glutenite reservoir are analysed via reservoir numerical simulation technology. The effects of reservoir thickness, reservoir heterogeneity, macropores, dominant channels and fracturing on CO2 flooding efficiency are evaluated individually, and a set of reasonable parameters constituting an evaluation system for CO2 flooding in the glutenite reservoir is established. Finally, based on parameter optimization, the disadvantaged conditions of CO2 flooding in glutenite reservoirs are optimized, and their effects are gradually eliminated. In view of the characteristics of glutenite reservoirs, a unique three-dimensional well pattern arrangement is designed, different injection modes are optimized, and suitable injection agents are tested. These methods are conducive to improving the effect of reservoir parameters on CO2 flooding efficiency, allowing these reservoirs to be used for CO2 flooding. Based on these new methods, the displacement effect of the Yanjia-Yongan glutenite reservoir is predicted. Our results show that the reservoir has been developed efficiently and achieved a high recovery rate, the displacement front of the CO2 has become more uniform, and the sweep range has become wider. After extending the successful application of CO2 flooding in this reservoir to all of the glutenite reservoirs in the entirety of the Bohaiwan Basin, we predict that the oil recovery rate may reach 40%, with a cumulative oil recovery rate of approximately 3.04 × 108 t and a total CO2 consumption of 1.672 × 108 t. Thus, the proposed approach not only can improve the atmospheric environment but can also greatly improve the efficiency of oil displacement.