Search

Knock is one of the main obstacles to improving the thermal efficiency of spark-ignition internal combustion engines. Although knock has been widely studied and accepted as a result of end-gas auto-ignition, the fundamentals regarding auto-ignition behaviors are still not fully revealed. In this study, the ignition behaviors of primary reference fuels were investigated in an optical rapid compression machine equipped with a quartz combustion chamber allowing for visualizing the combustion process from the lateral view. By combining both the lateral view and the top view photography, ignition behaviors under vortex-existing conditions with a creviced piston and vortex-minimized conditions with a flat piston were comprehensively analyzed to reveal the impact of the vortex on the ignition behaviors of PRF fuels. The influence of fuel reactivity was also investigated. The results showed that mild ignition was prevalent under large Da* numbers. The occurrence of mild ignition was closely related to the ignition delay time of the mixture, and the critical ignition delay time was not fixed but decreased with increasing initial temperature. The propensity of mild ignition could be boosted under vortex-existing conditions due to the increasing hotspot formation probability. Vortices were demonstrated to be capable of mitigating knock intensity via 1) the buffer effect of surrounding burned regions on the shock waves generated from surrounded unburnt pockets; 2) a larger burned mass fraction at the instant of the final ignition under vortex-existing conditions. The results also showed that strong ignition was more likely to occur under vortex-minimized conditions. Besides, higher fuel reactivity also could increase the probability of strong ignition occurrence. Compared with the creviced piston, the use of the flat piston could shift the ignition regime towards regions with higher Reₜ and lower Daₜ in the Daₜ-Reₜ diagram, where the strong ignition is less pronounced.

Breast cancer is a common life-threatening disease among women. Computer-aided methods can provide second opinion or decision support for early diagnosis in mammography images. However, the whole images classification is highly challenging due to small sizes of lesion and slow contrast between lesions and fibro-glandular tissue. In this paper, inspired by conventional machine learning methods, we present a Multi Frequency Attention Network (MFA-Net) to highlight the salient features. The network decomposes the features into low spatial frequency components and high spatial frequency components, and then recalibrates discriminating features based on two-dimensional Discrete Cosine Transform in two different frequency parts separately. Low spatial frequency features help determine if there is a tumor while high spatial frequency features help focus more on the margin of the tumor. Our studies empirically show that compared to traditional convolutional neural network (CNN), the proposed method mitigates the influence of the margin of pectoral muscle and breast in mammography, which brings significant improvement. For malignant and benign classification, by using transfer learning, the proposed MFA-Net achieves the AUC index 91.71% on the INbreast dataset.

U-Net is a commonly used deep learning model for mammogram segmentation. Despite outstanding overall performance in segmenting, U-Net still faces from two aspects of challenges: (1) the skip-connections in U-Net have limitations, which may not be able to effectively extract multi-scale features for breast masses with diverse shapes and sizes. (2) U-Net only merges low-level spatial information and high-level semantic information through concatenating, which neglects interdependencies between channels. To address these two problems, we propose the U-shape adaptive scale network (ASU-Net), which contains two modules: adaptive scale module (ASM) and feature refinement module (FRM). In each level of skip-connections, ASM is used to adaptively adjust the receptive fields according to the different scales of the mass, which makes the network adaptively capture multi-scale features. Besides, FRM is employed to allows the decoder to capture channel-wise dependencies, which make the network can selectively emphasize the feature representation of useful channels. Two commonly used mammogram databases including the DDSM-BCRP database and the INbreast database are used to evaluate the segmentation performance of ASU-Net. Finally, ASU-Net obtains the Dice Index (DI) of 91.41% and 93.55% in the DDSM-BCRP database and the INbreast database, respectively.

The combustion of combustible gaseous mixture with/without oil droplet is studied based on a rapid compression machine under conditions comparable with the state at the end of compression in SI-eng...

Based on visual cortical theory of Rybak, a new heterogeneous Rybak neural network (HRYNN) model is proposed for image enhancement. HRYNN is constructed with several Rybak neural network (RYNN) models proposed, which have different parameters corresponding to different neurons. We show that HRYNN can better represent prior information for edge detail enhancement than the logarithmic domain. To capture different resolution texture features of image, a novel receptive field model is proposed to solve the problem of detail enhancement. HRYNN model has excellent enhancement effect on the edge details of image based on the receptive field’s lateral inhibitory characteristics. Moreover, the experimental enhancement results of the colour images from Berkeley image Dataset show the validity and efficiency of the proposed enhancement method. Finally, three evaluation indicators are employed to measure the enhancement result.

Intra-class inconsistency and inter-class indistinction are intractable problems that commonly exist in breast mass segmentation from mammograms. In this work, a novel deep learning segmentation model is presented to address these problems. Firstly, we propose a simple yet effective aggregated pyramid attention module (APAM) for capturing intra-class dependencies, aiming at effectively aggregating contextual dependencies from different receptive fields to reinforce feature representations. Then, a novel aggregated pyramid attention network (APANet) is developed for further releasing the limitation of both intra-class inconsistency and inter-class indistinction. The APANet can combine low-level spatial details and high-level contextual information via encoder-decoder structure for further refining semantic representations. Finally, our proposed APANet is greatly demonstrated on two public mammographic databases including the DDSM-BCRP and INbreast, separately achieving the Dice Similarity Coefficient (DSC) of 91.04% and 94.02%.

Automated benign and malignant breast masses classification is a crucial yet challenging topic. Recently, many studies based on convolutional neural network (CNN) are presented to address this task, but most of these CNN-based methods neglect the effective global contextual information. Moreover, their methods do not further analyze the reliability and interpretability of CNN models, which does not correspond to the clinical diagnosis. In this work, we firstly propose a novel multi-level global-guided branch-attention network (MGBN) for mass classification, which aims to fully leverage the multi-level global contextual information to refine the feature representation. Specifically, the MGBN includes a stem module and a branch module. The former extracts the local information through standard local convolutional operations of ResNet-50. The latter embeds the global contextual information and establishes the relationships of different feature levels via global pooling and Multi-layer Perceptron (MLP). The final prediction is computed by local information and global information together. Then, we discuss the reliability and interpretability of our mass classification network by visualizing the coarse localization map through Gradient-weighted Class Activation Mapping (Grad-CAM), which is important in clinical diagnosis. Finally, our proposed MGBN is greatly demonstrated on two public mammographic mass classification databases including the DDSM and INbreast databases, resulting in AUC of 0.8375 and 0.9311, respectively.

Image enhancement plays an important role in improving image quality in the field of image processing, which is achieved by highlighting useful information and suppressing redundant information in the image. In this paper, the development of image enhancement algorithms is surveyed. The purpose of our review is to provide relevant researchers with a comprehensive and systematic analysis on image enhancement techniques and give them a valuable reference. Various image enhancement algorithms were mentioned and underlying difficulties, limitations, merits and disadvantages were discussed in applying these techniques in the past two decades with three aspects: supervised algorithm, unsupervised algorithm and quality evaluation, respectively. Further, we summarize some existing problems and analyze the future development trend of existing enhanced algorithms.

Breast cancer is currently the second most fatal cancer in women, but timely diagnosis and treatment can reduce its mortality. Breast masses are the most obvious means of cancer identification, and thus, accurate segmentation of masses is critical. In contrast to mass-centered patch segmentation, accurate segmentation of breast masses in full-field mammograms is always a challenging topic because of the extremely low signal-to-noise ratio and the uncertainty with respect to the shape, size, and location of the mass. In this study, we propose a novel adaptive channel and multiscale spatial context network for breast mass segmentation in full-field mammograms. A standard encoder-decoder structure is employed, and an elaborate adaptive channel and multiscale spatial context module (ACMSC module) is embedded in a multilevel manner in our network for accurate mass segmentation. The proposed ACMSC module utilizes the self-attention mechanism to adaptively capture discriminative contextual information among channel and spatial dimensions.The multilevel embedding of the ACMSC module enables the network to learn distinguishing features on multiple scales of feature maps. Our proposed model is evaluated on two public datasets, CBIS-DDSM and INbreast. The experimental results show that by adaptively capturing the context of the channel and spatial dimensions, our model can effectively remove false positives, predict difficult samples and achieve state-of-the-art results, with Dice coefficients of 82.81% for CBIS-DDSM and 84.11% for INbreast, respectively. We hope that our work will contribute to the CAD system for breast cancer diagnosis and ultimately improve clinical diagnosis.

A long-term research goal of pulse-coupled neural network (PCNN) is to control neuronal firing states at each iteration. Recently, we propose a fire-controlled MSPCNN model (FC-MSPCNN) and provide a parameter setting method to control firing and fired neurons within an effective pulse cycle. We firstly design the proposed model according to previous prevalent PCNN models. Secondly, the setting methods of the adaptive parameters α, β, V, and Rn are given to control neuronal firing time more effectively. Thirdly, a predetermined parameter P will determine the total iteration times of all the neurons. Fourthly, we also propose a color image quantization method and a gallbladder image location method based on the FC-MSPCNN. The evaluation experiments achieve good image processing performances compared to prevalent PCNN models and prove the effectiveness and robustness of the proposed methods.

To explore a suitable combustion strategy for natural gas engines using jet ignition, lean burn with air dilution, stoichiometric burn with EGR dilution and lean burn with EGR dilution were investigated in a single-cylinder natural gas engine, and the performances of two kinds of jet ignition technology, passive jet ignition (PJI) and active jet ignition (AJI), were compared. In the study of lean burn with air dilution strategy, the results showed that AJI could extend the lean limit of excess air ratio (λ) to 2.1, which was significantly higher than PJI’s 1.6. In addition, the highest indicated thermal efficiency (ITE) of AJI was shown 2% (in absolute value) more than that of PJI. Although a decrease of NOx emission was observed with increasing λ in the air dilution strategy, THC and CO emissions increased. Stoichiometric burn with EGR was proved to be less effective, which can only be applied in a limited operation range and had less flexibility. However, in contrast to the strategy of stoichiometric burn with EGR, the strategy of lean burn with EGR showed a much better applicability, and the highest ITE could achieve 45%, which was even higher than that of lean burn with air dilution. Compared with the most efficient points of lean burn with pure air dilution, the lean burn with EGR dilution could reduce 78% THC under IMEP = 1.2 MPa and 12% CO under IMEP = 0.4 MPa. From an overall view of the combustion and emission performances under both low and high loads, the optimum λ would be from 1.4 to 1.6 for the strategy of lean burn with EGR dilution.

Breast density is known as a significant indicator of breast cancer risk prediction and greatly reduces the digital mammograms sensitivity. In this work, based on the simple pulse coupled neural network (SPCNN), a novel Morph_SPCNN model is proposed for dealing with the limitations of over-segmentation that commonly existed in density segmentation of mammograms. To evaluate the proposed model, the segmentation result is employed as a feature map of the breast density classification system. In addtion, the texture features of mammogram calculated based on the gray level co-occurrence matrix (GLCM) and the statistical features (mean, skewness, kurtosis) are extracted and input to the support vector machine (SVM) for breast density classification. Finally, the performance of SVM classifier is evaluated based on the ten-fold cross-validation. Our method is verified both on the MIAS dataset, DDSM database and hybrid dataset (MIAS database and Gansu Provincial Academy of Medical Sciences (GPAMS) database), respectively achieving 87.80%, 94.89% and 95.37% accuracy for breast density classification. The experimental results indicate that our proposed method has greatly improved the performance of breast density segmentation and classification.

Spark assistance for homogeneous charge compression ignition (HCCI) can control combustion phasing, improve thermal efficiency, and reduce emissions in gasoline engines. As the characteristics of flame propagation determine the control authority of ignition timing, it is important and necessary to investigate pressure dependence of flame speed in the lean-premixed mixture relative to engine operating conditions. Experimental study in an optical rapid compression machine (RCM) and simulation work were carried out using two fuels comprising n-heptane/iso-octane/ethanol with varied octane sensitivity (S). The effective pressure ranged from 10 to 35 bar, temperature from 715 to 860 K, and equivalence ratios between 0.3 and 0.7 to cover the region of lean flammability limits of low and high S fuels with ethanol blended. Based on pressure profiles, flame speed extracted from images, and sensitivity analysis of flame speed, the dependence of flame speed on the effective pressure in low and high S fuels was discovered and the fundamental mechanism behind this phenomena became to be understood in the negative temperature coefficient (NTC) and non-NTC regions, respectively. In the studied temperature conditions, the flame speed of high S fuel has stronger dependence on the pressure than that of low S fuel does. In the NTC region, this phenomenon is attributed to the dependence of H radical concentration on pressure in the unburned mixture and flame structure. In the non-NTC region, promoting effect of dominant reactions varied with pressure can significantly influence pressure dependence of flame speed. Although quite limited data of laminar burning velocity for studied fuels were obtained in high pressures (>15 bar), the trend of flame speed's dependence on pressure was well predicted by two models with different but well-accepted core mechanisms, showing consistent results with the experimental ones in the RCM.

Spark assistance is an effective method to improve combustion stability of homogeneous charge compression ignition (HCCI) with lean mixture. Ethanol, a renewable energy, blended with commercial fuels is a promising method to deal with problems of energy safety and greenhouse gas (GHG) emission. However, the influence of ethanol blends on spark ignition to compression ignition is not clear. To this end, this study presents experimentally fundamental investigation on the compression ignition (with and without spark assistance) of ethanol-blended gasoline surrogate fuel in an optical rapid compression machine at the equivalence ratio of 0.5. Three fuels with different octane sensitivities (S) comprising n-heptane/iso-octane/ethanol were formulated and S is the difference between research octane number (RON) and motor octane number (MON). Effective temperature of experiments ranges from 730 K to 860 K, overlapping most of the temperature region with negative temperature coefficient (NTC) of iso-octane at the corresponding equivalence ratio, and the pressure is consistence with engine-relevant operating conditions. The results show that more fuel reactivity of ethanol than iso-octane at 860 K is not attributed to the NTC behavior of iso-octane but to the reactive species generation in ethanol oxidation. There is always a positive correlation between effective pressure and knock intensity (KI). At lower temperature (

Super-knock has been the main obstacle to improve power density and engine efficiency of modern highly turbocharged gasoline engines. Previous researches show that pre-ignition is the inducement of super-knock, while detonation is the root reason of how super-knock could damage engines dramatically. Lots of studies have been conducted to eliminate pre-ignition for suppressing super-knock indirectly. This work applies the fuel of methane to suppress detonation and then to suppress super-knock directly using a rapid compression machine (RCM). Furthermore, 1-D simulation model was set up to investigate the mechanism why methane could suppress detonation and super-knock. Finally, through single cylinder engine tests, this suppression strategy in engine practical usage was validated. The experiment and simulation results show that replacing the fuel from iso-octane or gasoline to methane while keeping other conditions identically could transfer detonation combustion mode to flame propagation. The peak pressure could be reduced dramatically and there is no pressure oscillation. Therefore, methane mixture could suppress detonation and then to suppress super-knock effectively, even if pre-ignition already exists. It could be an effective and practical control strategy to protect modern highly turbocharged engines.

Breast cancer poses a threat to the lives of many women. Breast density is a closely related indicator of breast cancer risk. The aim of this paper is to propose a classification system for breast density, which can appropriately segment the glandular tissue from the whole breast and to achieve a better classification result. A new threshold method is applied to segment the breast glandular tissue. The gray level co-occurrence matrix (GLCM) is implemented to extract the texture features of the glandular tissue. Meanwhile, we obtain three statistical features (mean, skewness, kurtosis). In addition, the calculated breast density that is served as a new feature is added to the feature vectors. The mixed feature vectors are classified by Support Vector Machine (SVM) and Ultimate Learning Machine (ELM). Ten-fold cross-validation is used to verify the classifier performance. The system using the SVM achieves 96.19% accuracy for three density types in the MIAS database and achieves 96.35% accuracy of four density types in the DDSM database. The accuracy in the database mixed with the local database was 95.01% and there are three density types in the mixed database. The experimental results indicate that the system proposed has a better performance in breast density classification. The system proposed in this paper can be considered to help the physician to classify breast density.

Super-knock has been the main obstacle to improve power density and engine efficiency of modern gasoline engines. Previous researches show that the key points of super-knock are composed of pre-ignition and detonation. The former is the inducement and the latter is the root reason why super-knock could damage engines dramatically. Lots of studies have been conducted to suppress super-knock by eliminating pre-ignition. Using a rapid compression machine, this work explores the lean mixture control strategy to eliminate detonation and thus suppressing super-knock. Furthermore, a one-dimensional simulation model was set up to investigate the mechanism of how lean mixture could suppress detonation and super-knock. The experimental and simulation results show that based on detonation and super-knock operating condition, adding 80% additional air while keeping the same fuel energy density could transfer combustion from detonation mode to super-sonic auto-ignitive deflagration mode. The peak pressure and pressure oscillation could be reduced significantly. This indicates that boosted lean burn could suppress detonation and thus suppress super-knock effectively, even if pre-ignition occurs. It could be an effective and practical control strategy to protect modern highly turbocharged gasoline engines.

Overall ignition delay time (OID) is used as indicator to characterize reactivity and anti-knock quality of fuel. In this work, both experimental measurement and modeling work have been conducted to evaluate OID properties for six ternary blends comprising n-heptane/iso-octane/ethanol. Experimental measurement was performed in a rapid compression machine, while modeling work was based on the gasoline surrogate mechanism. The decoupling study on research octane number (RON)/motor octane number (MON) and octane sensitivity (S) has been carried out by adjusting fuel composition. The thermal conditions of experiments cover low-to-medium temperatures from 640 K to 740 K and pressures from 9.5 bar to 21 bar. Oxygen is utilized as the oxidizer while nitrogen and argon are regarded as the buffer gas to adjust thermal conditions. Combined with chemical kinetics analysis, negative temperature coefficient (NTC) behavior in the experiment is entirely attributed to iso-octane and the timing of the first peak HO2 radical concentration has more significant influence on NTC behavior than peak value itself. Octane number and/or S cannot fully revealed fuel anti-knock quality, which is substantially influence by the property of fuels and thermal conditions. At low temperature (640 K), fuels with higher S exhibit better anti-knock performance. However, this advantage will disappear when temperature falls into NTC region. For engine operating in this area, high RON but low S fuels are recommended. Ethanol extends the beneficial area of high S contents due to its low temperature chemical inertness. The low temperature boundary of NTC for iso-octane can be regarded as the demarcation line to determine whether high S contents should be utilized, which is an important reference for fuel-engine co-optima research.

Super-knock is the main obstacle to improve power density and fuel efficiency of highly boosted gasoline engines. Previous investigations show that pre-ignition and detonation are the two key combustion processes of super-knock. The former is the inducement and the latter is the root reason of how super-knock could damage engines dramatically. Lots of studies have been conducted for suppressing super-knock through eliminating pre-ignition. Using a rapid compression machine, this study explores the stoichiometric propane or methanol mixture to suppress super-knock by eliminating detonation. Under the same pressure at the end of compression, the same fuel energy density and the same charged fresh air, the peak pressure of propane mixture could be reduced dramatically and the pressure oscillation could be eliminated, when compared to iso-octane mixture. The combustion could be transferred from detonation to flame propagation. For methanol mixture, the combustion process could be transferred from detonation to weak auto-ignition with low peak pressure and negligible pressure oscillation. These indicate that both propane and methanol mixture could suppress detonation and thus suppress super-knock effectively, even if pre-ignition occurs. It could be an effective and practical control strategy to protect modern highly turbocharged spark ignition engines.

PURPOSE Breast mass segmentation in mammograms remains a crucial yet challenging topic in computer-aided diagnosis systems. Existing algorithms mainly used mass-centered patches to achieve mass segmentation, which is time-consuming and unstable in clinical diagnosis. Therefore, we aim to directly perform fully automated mass segmentation in whole mammograms with deep learning solutions. METHODS In this work, we propose a novel dual contextual affinity network (a.k.a., DCANet) for mass segmentation in whole mammograms. Based on the encoder-decoder structure, two lightweight yet effective contextual affinity modules including the global-guided affinity module (GAM) and the local-guided affinity module (LAM) are proposed. The former aggregates the features integrated by all positions and captures long-range contextual dependencies, aiming to enhance the feature representations of homogeneous regions. The latter emphasizes semantic information around each position and exploits contextual affinity based on the local field-of-view, aiming to improve the indistinction among heterogeneous regions. RESULTS The proposed DCANet is greatly demonstrated on two public mammographic databases including the DDSM and the INbreast, achieving the Dice similarity coefficient (DSC) of 85.95% and 84.65%, respectively. Both segmentation performance and computational efficiency outperform the current state-of-the-art methods. CONCLUSION According to extensive qualitative and quantitative analyses, we believe that the proposed fully automated approach has sufficient robustness to provide fast and accurate diagnoses for possible clinical breast mass segmentation.

UNet adopting an encoder-decoder structure has been used widely in medical image segmentation tasks for its outstanding performance. However, in our work, we find that UNet has the worse segmentation performance of small masses. The reason behind this is that the sizes of receptive fields are limited. In this work, to address this issue, we develop a novel end-to-end model, Adaptive Receptive Field Network (ARF-Net), for the precise breast mass segmentation in whole mammographic images and ultrasound images. ARF-Net composes of an encoder network and a corresponding decoder network, followed by a pixel-wise classifier. In ARF-Net, a Selective Receptive Filed Module (SRFM) is proposed to allocate the suitable sizes of receptive fields to the breast masses of different sizes. SRFM consists of a Multiple Receptive Field Module (MRFM) for generating multiple receptive fields of different sizes and a Multi-Scale Selection Module (MSSM) for selecting the suitable sizes of receptive fields based on the objects’ size. The proposed ARF-Net achieves the dice index of 86.1 % , 85.75 % , and 88.12 % on the two mammographic databases (INbreast and CBIS-DDSM) and one ultrasonic database (UDIAT), respectively. Moreover, extensive ablation experiments show that ARF-Net transcends several state-of-the-art segmentation networks, and the developed MSSM exceeds several counterparts.

An experimental investigation on the ignition of a single lubricating oil droplet in combustible ambient gaseous mixture was performed on a rapid compression machine, under high-temperature...

The present study focuses on one mode of detonation initiation observed in a rapid compression machine (RCM). This mode is referred to as shock wave and flame front-induced detonation (SWFID). Experimental high-speed imaging and two-dimensional numerical simulations with skeletal chemistry are combined to unravel the dominant steps of detonation initiation under SWFID conditions. It is shown that the interaction between the shock wave generated by the end-gas auto-ignition and the spherical flame creates a region of high pressure and temperature which enables the acceleration of the flame front and the detonation onset. The experimental observation lacks adequate spatial and temporal resolution despite good reproducibility of the detonation onset. Based on the numerical results, phenomenological interpretation of the event within the framework of shock wave refraction indicates that the formation of a free-precursor shock wave at the transition between regular and irregular refraction may be responsible for detonation onset. The present results along with previous findings on shock wave reflection-induced detonation in the RCM indicate that super-knock occurs after the interaction of the shock wave generated by end-gas auto-ignition with the RCM walls, preignition flame, or another shock wave.

Gasoline direct injection (GDI) engines are currently the dominant powertrains for passenger cars. With the implementation of increasingly stringent fuel consumption and emission regulations worldwide, GDI engines are facing challenges owing to high particulate matter emissions and a tendency to knock, leading to a change in the research and design (R&D) issues compared with those in the twentieth century. This paper reviews the progress in research regarding GDI engine technologies over the past 20 years, focusing on combustion system configurations, and also highlights common issues in GDI R&D, including pre-ignition and deto-knock, soot formation and PM emissions, injector deposits and gasoline compression ignition (GCI). First, an overview of recent developments in the field as driven by regulations is provided, following which progress in injection and combustion systems is examined. Third, the review addresses the occurrence and mechanism of deto-knock and considers means of suppressing this phenomenon. The fourth section discusses soot formation mechanisms and particulate matter emission characteristics of GDI engines and describes the application of gasoline particulate filter (GPF) after-treatment. The subsequent section summarizes studies regarding injector deposit formation, as well as pioneering research into GCI combustion modes. Finally, a summary and future prospects for GDI engine technologies are provided.

Based on the fact that actual cerebral cortex has different structure, a new heterogeneous simplified pulse coupled neural network (HSPCNN) model is proposed in this paper for image segmentation. HSPCNN is constructed with several simplified pulse coupled neural network (SPCNN) models, which have different parameters corresponding to different neurons. An image is segmented by HSPCNN into several regions according to their gray levels. Moreover, the parameter of HSPCNN is set automatically in this paper, the experimental segmentation results of the gray natural images from the Berkeley Segmentation Dataset (BSD 300) show the validity and efficiency of the proposed segmentation method. Finally, an evaluation index is proposed to measure the segmentation result.

Dedicated exhaust gas recirculation (D-EGR) is to generate H 2 via fuel-rich combustion and viewed as a potential technique to meet future emission regulations without further after-treatment. In this study, firstly, the H 2 /CO formation through fuel-rich combustion in a single-cylinder natural gas spark ignition engine was quantitatively characterized by gas chromatography. Then, the effect of H 2 /CO addition on stoichiometric natural gas combustion performance and emission characteristics at 15% and 20% EGR levels was investigated. Finally, reaction path analysis and the brute-force sensitivity of ignition delay were conducted to evaluate the effect of H 2 addition on reaction process. The yields of H 2 and CO approximately linearly increased from ∼2% to ∼10% as the equivalence ratio varied from 1.1 to 1.5. The H 2 /CO addition accelerated the flame speed of mixture and significantly shortened the combustion duration, significantly improving the indicated thermal efficiency and the total unburned hydrocarbon with the acceptable penalty of increased NOx and CO emissions. Numerical results revealed that OH + H 2 = H + H 2 O and H + O 2 = O + OH were the most sensitive reactions with the presence of H 2 . This study delivered a quantitative basis for the optimization of D-EGR fueling strategies in natural gas engines.

Breast mass segmentation in mammograms is still a challenging and clinically valuable task. In this paper, we propose an effective and lightweight segmentation model based on convolutional neural networks to automatically segment breast masses in whole mammograms. Specifically, we first developed feature strengthening modules to enhance relevant information about masses and other tissues and improve the representation power of low-resolution feature layers with high-resolution feature maps. Second, we applied a parallel dilated convolution module to capture the features of different scales of masses and fully extract information about the edges and internal texture of the masses. Third, a mutual information loss function was employed to optimise the accuracy of the prediction results by maximising the mutual information between the prediction results and the ground truth. Finally, the proposed model was evaluated on both available INbreast and CBIS-DDSM datasets, and the experimental results indicated that our method achieved excellent segmentation performance in terms of dice coefficient, intersection over union, and sensitivity metrics.

Breast density is a significant element for breast cancer precaution. The existing mammographic density classification methods cannot achieve satisfactory classification accuracy while achieving end-to-end. In this paper, we present a novel bilateral adaptive spatial and channel attention network (BASCNet) which integrates the information of the left and right breasts and adaptively pays attention to the discriminative features in spatial and channel dimensions. The proposed BASCNet has been fully proved on the public Digital Database for Screening Mammography (DDSM) and INbreast dataset, and the classification accuracies of 85.10% and 90.51% were achieved with fivefold cross-validation, respectively. Our method is fully automatic and has achieved the classification performance superior to the existing breast density classification methods. Massive ablation experiments were conducted to demonstrate the effectiveness of the network structure. Moreover, we compared the effects of different views (CC and MLO) on breast density classification and verified the effectiveness of the contralateral breast information integration. Overall, the proposed BASCNet has the potential to be applied to clinical diagnosis.

For the breast mass segmentation in whole mammograms, in our studies, we observe that there is an enormous performance reduction in the case of considering the normal data during training. Therefore, the mammogram classification (normal vs. abnormal) is essential for boosting the breast mass segmentation performance in whole mammograms and is our research topic in this paper. Due to the breast lesions with a variety of sizes, the mammogram classification (normal vs. abnormal) is a challenging task. To improve the mammogram classification performance, we propose an end-to-end convolutional neural network, namely Multi-Scale Attention-Guided Network (MSANet). Specifically, MSANet can be constructed by stacking several Multi-Scale Attention (MSA) bottlenecks. Each MAS bottleneck consists of a Scale Aggregation (SA) unit and a Multi-Scale Attention Module (MSAM). The SA unit is used to generate multiple feature maps of different scales, and the MSAM is used to allocate the suitable size of receptive field for objects of different sizes. According to the extensive experiments, our proposed MSANet-50 achieves a fully automated classification AUC of 0.942 on the DDSM database, which outperforms several approaches.

This paper presents numerical study of detonation initiation induced by shock wave reflection which is photoed by high speed camera in a rapid compression machine. It is found that detonation initiates near the wall of combustion chamber. In the numerical study, the process of shock wave propagation, the interaction between shock wave and wall, and the detonation initiation by shock wave reflection are analysed with a CFD software. The numerical simulation demonstrates that mach reflection occurs near the wall during the propagation of the shock wave in the combustion chamber. The temperature of the region between the slip plane and the wall becomes higher. The auto-ignition of a hot spot in the high temperature region forms shock wave which is coupled with the reaction surface, resulting in the detonation finally.

This paper investigates using EGR instead of throttle to control the load of a stoichiometric dual-fuel dieseline (diesel and gasoline) compression ignition (SDCI) engine for reducing fuel consumption and emissions in future internal combustion engines. To investigate the combustion process in the SDCI mode, the flame development in a single-cylinder optical engine with a dual injection (PFI + DI system) has been recorded by high-speed imaging. The images are processed and analyzed, and the premixed blue flame is identified according to their RGB magnitudes. The effects of EGR and load conditions, injection pressure, injection timing, and split injection strategy on SDCI combustion are studied. An earlier injection strategy is found to be ideal for soot reduction; however, the ignition–injection decoupling problem results in difficulties in combustion control. A split injection strategy offers a flexible solution as simultaneous optimization for control of pressure rise rate, reduction of soot and increase of thermal efficiency. The premixed blue flame propagation speeds are correlated to the heat release data, and it is found that the flame propagation speeds are significantly higher than those in conventional spark ignition (SI) engines even with an EGR ratio higher than 40%.