Solar Radiation Estimation Based on a New Combined Approach of Artificial Neural Networks (ANN) and Genetic Algorithms (GA) in South Algeria

When designing solar systems and assessing the effectiveness of their many uses,estimating sun irradiance is a crucial first step.This study examined three approaches(ANN,GA-ANN,and ANFIS)for estimating daily global solar radiation(GSR)in the south of Algeria:Adrar,Ouargla,and Bechar.The proposed hybrid GA-ANN model,based on genetic algorithm-based optimization,was developed to improve the ANN model.The GA-ANN and ANFIS models performed better than the standalone ANN-based model,with GA-ANN being better suited for forecasting in all sites,and it performed the best with the best values in the testing phase of Coefficient of Determination(R=0.9005),Mean Absolute Percentage Error(MAPE=8.40%),and Relative Root Mean Square Error(rRMSE=12.56%).Nevertheless,the ANFIS model outperformed the GA-ANN model in forecasting daily GSR,with the best values of indicators when testing the model being R=0.9374,MAPE=7.78%,and rRMSE=10.54%.Generally,we may conclude that the initial ANN stand-alone model performance when forecasting solar radiation has been improved,and the results obtained after injecting the genetic algorithm into the ANN to optimize its weights were satisfactory.The model can be used to forecast daily GSR in dry climates and other climates and may also be helpful in selecting solar energy system installations and sizes.

Constructing low-dimensional perovskite network to assist efficient and stable perovskite solar cells

The use of low-dimensional(LD)perovskite materials is crucial for achieving high-performance perovskite solar cells(PSCs).However,LD perovskite films fabricated by conventional approaches give rise to full coverage of the underlying 3D perovskite films,which inevitably hinders the transport of charge carriers at the interface of PSCs.Here,we designed and fabricated LD perovskite structure that forms net-like morphology on top of the underlying three-dimensional(3D)perovskite bulk film.The net-like LD perovskite not only reduced the surface defects of 3D perovskite film,but also provided channels for the vertical transport of charge carriers,effectively enhancing the interfacial charge transfer at the LD/3D hetero-interface.The net-like morphological design comprising LD perovskite effectively resolves the contradiction between interfacial defect passivation and carrier extraction across the hetero-interfaces.Furthermore,the net-like LD perovskite morphology can enhance the stability of the underlying 3D perovskite film,which is attributed to the hydrophobic nature of LD perovskite.As a result,the net-like LD perovskite film morphology assists PSCs in achieving an excellent power conversion efficiency of up to 24.6%with over 1000 h long-term operational stability.

Understanding of the Relationship between the Properties of Cu(In,Ga)Se_(2) Solar Cells and the Structure of Ag Network Electrodes

The relation between the structure of the silver network electrodes and the properties of Cu(In,Ga)Se_(2)(CIGS)solar cells is systemically investigated.The Ag network electrode is deposited onto an Al:ZnO(AZO)thin film,employing a self-forming cracked template.Precise control over the cracked template's structure is achieved through careful adjustment of temperature and humidity.The Ag network electrodes with different coverage areas and network densities are systemically applied to the CIGS solar cells.It is revealed that predominant fill factor(FF)is influenced by the figure of merit of transparent conducting electrodes,rather than sheet resistance,particularly when the coverage area falls within the range of 1.3–5%.Furthermore,a higher network density corresponds to an enhanced FF when the coverage areas of the Ag networks are similar.When utilizing a thinner AZO film,CIGS solar cells with a surface area of 1.0609 cm^(2)exhibit a notable performance improvement,with efficiency increasing from 10.48%to 11.63%.This enhancement is primarily attributed to the increase in FF from 45%to 65%.These findings underscore the considerable potential for reducing the thickness of the transparent conductive oxide(TCO)in CIGS modules with implications for practical applications in photovoltaic technology.

PREPARATION AND PROPERTIES OF GEL POLYMER ELECTROLYTE BASED ON PVDF/ACRYLATE INTERPENETRATING POLYMER NETWORK FOR DYE-SENSITIZED SOLAR CELL

The gel polymer electrolytes（GPEs）based on poly（vinylidence fluoride）（PVDF）/acrylate interpenetrating polymer network（IPN）are prepared.The micro-phase separation type GPEs are characterized by Fourier transform infrared（FTIR）spectroscopy,scanning electron microscope（SEM）,respectively.Moreover,the conductivity and the voltage-current curves of the electrolytes are measured by electrochemical workstation.The higher porosity and electrolyte uptake are observed in the membranes prepared at lower crosslinker concentration.The suitable cross-linking acrylate monomer improves the porosity and the electrochemical behavior of GPE.A dye-sensitized solar cell（DSSC）employing PGE based on PVDF/poly（ethylene glycol dimethacrylate）（PEGDMA）IPN yields an open-circuit voltage of 0.674 V,short-circuit current of 8.476 mA·cm-2and the conversion efficiency of 2.710% under 100 mW·cm-2illumination.

MACDCGAN的发电机轴承故障诊断方法

在实际工况中,发电机中传感器采集到的故障样本数据有限,使用基于深度学习的方法进行故障诊断存在过拟合问题导致模型泛化能力较差以及诊断精度不高。为了解决这个问题,采用样本扩充的思路,提出了一种改进的辅助分类器条件深度卷积生成对抗网络(MACDCGAN)的故障诊断方法。通过对采集的一维时序信号进行小波变换增强特征,构建简化结构参数的条件深度卷积生成对抗网络模型生成样本,并在模型中采用Wasserstein距离优化损失函数解决训练过程中存在模式崩塌和梯度消失的缺点;通过添加一个独立的分类器来改进分类模型的兼容性,并在分类器中引入学习率衰减算法增加模型稳定性。试验结果表明,该方法可以有效地提高故障诊断的精度,并且验证了所提模型具有良好的泛化性能。

无线传感网络中一种基于RF-GAN模型的入侵检测算法

针对现有无线传感网络入侵检测算法存在的效率低、精度差等问题,提出基于RFGAN模型的无线传感网络入侵检测算法。首先,采集无线传感网络运行数据,通过去噪、缺失补偿等步骤,完成对原始数据的预处理;然后,利用RF-GAN模型提取无线传感网络运行特征;最后,通过提取特征与检测标准的匹配,得出网络入侵检测结果。理论分析及实验结果表明:优化设计方法的入侵类型误检率明显更低,入侵数据量检测误差为0.015GB,相较于现有检测算法具有一定优势。

Solar image deconvolution by generative adversarial network

With aperture synthesis(AS)technique,a number of small antennas can be assembled to form a large telescope whose spatial resolution is determined by the distance of two farthest antennas instead of the diameter of a single-dish antenna.In contrast from a direct imaging system,an AS telescope captures the Fourier coefficients of a spatial object,and then implement inverse Fourier transform to reconstruct the spatial image.Due to the limited number of antennas,the Fourier coefficients are extremely sparse in practice,resulting in a very blurry image.To remove/reduce blur,“CLEAN”deconvolution has been widely used in the literature.However,it was initially designed for a point source.For an extended source,like the Sun,its efficiency is unsatisfactory.In this study,a deep neural network,referring to Generative Adversarial Network(GAN),is proposed for solar image deconvolution.The experimental results demonstrate that the proposed model is markedly better than traditional CLEAN on solar images.The main purpose of this work is visual inspection instead of quantitative scientific computation.We believe that this will also help scientists to better understand solar phenomena with high quality images.

Novel model of a AlGaN/GaN high electron mobility transistor based on an artificial neural network

In this paper we present a novel approach to modeling AlGaN/GaN high electron mobility transistor （HEMT） with an artificial neural network （ANN）. The AlGaN/GaN HEMT device structure and its fabrication process are described. The circuit-based Neuro-space mapping （neuro-SM） technique is studied in detail. The EEHEMT model is implemented according to the measurement results of the designed device, which serves as a coarse model. An ANN is proposed to model AIGaN/CaN HEMT based on the coarse model. Its optimization is performed. The simulation results from the model are compared with the measurement results. It is shown that the simulation results obtained from the ANN model of A1GaN/GaN HEMT are more accurate than those obtained from the EEHEMT model.

基于Transformer和GAN的对抗样本生成算法

对抗攻击与防御是计算机安全领域的一个热门研究方向。针对现有基于梯度的对抗样本生成方法可视质量差、基于优化的方法生成效率低的问题,提出基于Transformer和生成对抗网络(GAN)的对抗样本生成算法Trans-GAN。首先利用Transformer强大的视觉表征能力,将其作为重构网络,用于接收干净图像并生成攻击噪声;其次将Transformer重构网络作为生成器,与基于深度卷积网络的鉴别器相结合组成GAN网络架构,提高生成图像的真实性并保证训练的稳定性,同时提出改进的注意力机制Targeted Self-Attention,在训练网络时引入目标标签作为先验知识,指导网络模型学习生成具有特定攻击目标的对抗扰动;最后利用跳转连接将对抗噪声施加在干净样本上,形成对抗样本,攻击目标分类网络。实验结果表明:Trans-GAN算法针对MNIST数据集中2种模型的攻击成功率都达到99.9%以上,针对CIFAR10数据集中2种模型的攻击成功率分别达到96.36%和98.47%,优于目前先进的基于生成式的对抗样本生成方法;相比快速梯度符号法和投影梯度下降法,Trans-GAN算法生成的对抗噪声扰动量更小,形成的对抗样本更加自然,满足人类视觉不易分辨的要求。

Enhanced performance of InGaN/GaN multiple quantum well solar cells with double indium content

The performance of a multiple quantum well （MQW） InGaN solar cell with double indium content is investigated. It is found that the adoption of a double indium structure can effectively broaden the spectral response of the external quantum efficiencies and optimize the overall performance of the solar cell. Under AM1.5G illumination, the short-circuit current density （Jsc） and conversion efficiency of the solar cell are enhanced by 65% and 13% compared with those of a normal single-indium-content MQW solar cell. These improvements are mainly attributed to the expansion of the absorption spectrum and better extraction efficiency of the photon-generated carriers induced by higher polarization.

Decision Technique of Solar Radiation Prediction Applying Recurrent Neural Network for Short-Term Ahead Power Output of Photovoltaic System

In recent years, introduction of a renewable energy source such as solar energy is expected. However, solar radiation is not constant and power output of photovoltaic (PV) system is influenced by weather conditions. It is difficult for getting to know accurate power output of PV system. In order to forecast the power output of PV system as accurate as possible, this paper proposes a decision technique of forecasting model for short-term-ahead power output of PV system based on solar radiation prediction. Application of Recurrent Neural Network (RNN) is shown for solar radiation prediction in this paper. The proposed method in this paper does not require complicated calculation, but mathematical model with only useful weather data. The validity of the proposed RNN is confirmed by comparing simulation results of solar radiation forecasting with that obtained from other

融合门控变换机制和GAN的低光照图像增强方法

针对低光照图像增强过程中存在的配对图像数据依赖、细节损失严重和噪声放大问题,提出结合门控通道变换机制和生成对抗网络(GAN)的低光照图像增强方法AGR-GAN,该方法可以在没有低/正常光图像对的情况下进行训练。首先,设计特征提取网络,该网络由多个基于门控通道变换单元的多尺度卷积残差模块构成,以提取输入图像的全局上下文特征和多尺度局部特征信息;然后,在特征融合网络中,采用卷积残差结构将提取的深浅层特征进行充分融合,再引入横向跳跃连接结构,最大程度保留细节特征信息,获得最终的增强图像;最后,引入联合损失函数指导网络训练过程,抑制图像噪声,使增强图像色彩更自然匀称。实验结果表明,该方法在主观视觉分析和客观指标评价方面相较其他算法均具有显著优势,其能有效提高低光照图像的亮度和对比度,减弱图像噪声,增强后的图像更清晰且色彩更真实,峰值信噪比、结构相似度和无参考图像质量评价指标平均可达16.48 dB、0.93和3.37。

A Network Pharmacology Study on the Effects of Ma Xing Shi Gan Decoction on Influenza

Objective Pharmacological methods were used to screen targets and signaling pathways of Ma Xing Shi Gan Decoction(MXSGD)during influenza treatments,and mechanisms underlying antiinfluenza effects were elucidated.Methods The Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform(TCMSP)and relevant literature were searched under predefined conditions to identify the main compounds and their targets.Interactions between the target proteins were predicted using the STRING database.Gene Ontology(GO)functional enrichment analyses and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway analyses were performed on the core targets involved in the influenza protein-protein interaction(PPI)network,using WebGestalt and the reactome database.iGEMDOCK was used for molecular docking of receptors and ligands to produce docking scores,and the results were visualized using Autodock and PyMOL.Results In total,126 major compounds and their respective targets were screened.355 influenza target proteins and 1221 influenza protein interactions were predicted using the STRING database.Influenza-related signaling pathways were strongly enriched in pharmacodynamic targets of MXSGD such as cytokine signaling in immune system and signaling by interleukin.The main biological process was response to the stimulates.Molecular docking results showed that RELALicochalcone A docking elicited by MXSGD,was superior to that of other target proteins and active compounds,suggesting that the docking site is also the main effector site of MXSGD during influenza treatments.Conclusions The results showed that MXSGD exerts antiinfluenza effects by interfering with virus adsorption,inhibiting virus proliferation,influencing immune functions and protecting host cells,which may prevent inflammation-induced tissue damage.

基于有效注意力和GAN结合的脑卒中EEG增强算法

在基于脑电的卒中分类诊断任务中,以卷积神经网络为基础的深度模型得到广泛应用,但由于卒中类别病患样本数量少,导致数据集类别不平衡,降低了分类精度。现有的少数类数据增强方法大多采用生成对抗网络(GAN),生成效果一般,虽然可通过引入缩放点乘注意力改善样本生成质量,但存储及运算代价往往较大。针对此问题,构建一种基于线性有效注意力的渐进式数据增强算法LESA-CGAN。首先,算法采用双层自编码条件生成对抗网络架构,分别进行脑电标签特征提取及脑电样本生成,并使生成过程逐层精细化;其次,通过在编码部分引入线性有效自注意力(LESA)模块,加强脑电的标签隐层特征提取,并降低网络整体的运算复杂度。消融与对比实验结果表明,在合理的编码层数与生成数据比例下,LESA-CGAN与其他基准方法相比计算资源占用较少,且在样本生成质量指标上实现了10%的性能提升,各频段生成的脑电特征样本均更加自然,同时将病患分类的准确率和敏感度提高到了98.85%和98.79%。

Application of Hopfield Neural Networks Approach in Solar Energy Product Conceptual Design

A new product conceptual design approach is put forward based on Hopfield neural networks models. By research on the mechanisms of Hopfield neural networks, the associative simulation approaches are proposed. The approach is given by Hebb learn- ing law, Hopfield neural networks and crossover and mutation. The calculating models and the calculating formulas for the concep- tual design are put forward. Finally, an example for the conceptual design of a solar energy lamp is given. The better results are ob- tained in the conceptual design.

面向舰船目标检测的SAR图像数据PCGAN生成方法

针对现有合成孔径雷达(SAR)图像数据生成方法大多无法同时生成舰船图像及其检测标签的问题,面向SAR舰船图像生成及目标检测任务,构建基于位置信息的条件生成对抗网络(PCGAN).首先,提出将舰船位置信息作为约束条件用于限制生成图像中舰船的位置,并将其作为舰船图像的检测标签;随后,引入Wasserstein距离稳定PCGAN的训练过程;最后,利用生成的SAR舰船图像及对应检测标签完成YOLOv3网络的端到端训练,实现舰船数据增强与目标检测的协同学习,进而获得更耦合目标检测实际应用的多样性数据.在HRSID(high resolution SAR image dataset)数据集上的实验结果表明,PCGAN方法能生成清晰、鲁棒的SAR舰船数据,舰船检测准确度最高提升1.01%,验证了所提出方法的有效性.

Carrier transport via V-shaped pits in InGaN/GaN MQW solar cells

Carrier transport via the V-shaped pits （V-pits） in InGaN/GaN multiple-quantum-well （MQW） solar cells is numer- ically investigated. By simulations, it is found that the V-pits can act as effective escape paths for the photo-generated carriers. Due to the thin barrier thickness and low indium composition of the MQW on V-pit sidewall, the carriers entered the sidewall QWs can easily escape and contribute to the photocurrent. This forms a parallel escape route for the carries generated in the fiat quantum wells. As the barrier thickness of the fiat MQW increases, more carriers would transport via the V-pits. Furthermore, it is found that the V-pits may reduce the recombination losses of carriers due to their screening effect to the dislocations. These discoveries are not only helpful for understanding the carrier transport mechanism in the InGaN/GaN MQW, but also important in design of the structure of solar cells.

融合情感分析和GAN-TrellisNet的股价预测方法

将时序深度神经网络应用于股票价格预测,已成为量化金融领域的重要研究方向。时序神经网络具有很好的序列数据捕捉能力和学习记忆能力,在股票预测上有一定适用性。但是现有的模型大多存在预测准确度不高、模型结构复杂导致训练时间较长等问题.为了解决以上问题,提出了一种基于情感分析和GAN-TrellisNet的股价预测方法。提出了一个基于LSTM-CNN的情感分析模型,用于分析爬虫获取的主流金融论坛股票评论,并获得股票情感指数。为了提高预测准确度,将情感指数和百度搜索指数加入股票交易数据中作为训练集,提出了一个基于TrellisNet和CNN的改进型GAN股价预测模型,利用TrellisNet生成器的卷积特性来捕捉数据的局部特征,选取特征提取能力较强的CNN作为判别器来区别预测结果和真实股价。通过选取10只代表性股票和三种大盘指数的不同时段数据进行算法验证,结果表明,与ConvLSTM和GAN-LSTM预测模型相比,GAN-TrellisNet模型能有效缩短训练时间,提高预测准确率。

Advantages of InGaN/GaN multiple quantum well solar cells with stepped-thickness quantum wells

InGaN/GaN multiple quantum well （MQW） solar cells with stepped-thickness quantum wells （SQW） are designed and grown by metal-organic chemical vapor deposition. The stepped-thickness quantum wells structure, in which the well thickness becomes smaller and smaller along the growth direction, reveals better crystalline quality and better spectral overlap with the solar spectrum. Consequently, the short-circuit current density （Jsc） and conversion efficiency of the solar cell are enhanced by 27.12% and 56.41% compared with the conventional structure under illumination of AM1.5G （100 mW/cm2）. In addition, approaches to further promote the performance of InGaN/GaN multiple quantum well solar cells are discussed and presented.

High Concentration InGaN/GaN Multi-Quantum Well Solar Cells with a Peak Open-Circuit Voltage of 2.45V

We report InGaN/GaN multi-quantum well (MQW) solar cells with a comparatively high open-circuit voltage and good concentration properties.Thc open circuit voltage (Voc) keeps increasing logarithmically with concentration ratio until 60suns.The peak Voc of InGaN/GaN MQW solar cells,which has a predominant peak wavelength of 456nm from electroluminescence measurements,is found to be 2.45 V when the concentration ratio reaches 333×.Furthermore,the dcpendence of conversion efficiency and fill factor on concentration ratio are analyzed.