Reduced graphene oxide (rGO) composites demonstrate favorable energy storage characteristics, including adjustable porosity, superior conductivity, chemical stability, and remarkable charge storing capacity. Nonetheless, the intrinsic rigidity of rGO constrains its application in contemporary disposable and flexible energy storage systems. This paper details manufacture of flexible composites based on reduced graphene oxide and iron sulfide, utilizing natural fibers derived from discarded bioresources, specifically pineapple leaf fiber, as a binder. The novelty of this work lies in the sustainable utilization of pineapple leaf fiber (PALF) as a bio-binder combined with microwave-assisted rapid synthesis and electrodeposition-controlled tuning of Fe₃S₄ to fabricate highly flexible and eco-friendly paper electrodes. The rGO and iron sulfide (Fe₃S₄) nanoparticles are produced by a rapid microwave-assisted method. Additionally, Fe₃S₄ nanoparticles are electrochemically coated on synthesized rGO- paper electrodes to improve energy storage and electronic conductivity properties. Highly flexible paper electrodes were analyzed using several characterization techniques, including Fourier transform infrared (FTIR) spectroscopy, Raman spectroscopy and scanning electron microscopy (SEM) examined their chemical bonding and morphology. Electrochemical assessments, comprising Galvanostatic Charge/Discharge (GCD), Electrochemical Impedance Spectroscopy (EIS), and Cyclic Voltammetry (CV) were conducted to analyze capacitive and kinetics characteristics of electrodes. LC/rGO/Fe₃S₄ at 2400 s demonstrates a specific capacitance 69.79 F/g, rGO/LC, which were 60 F/g. The value of 69.79 F/g was obtained in a three-electrode configuration, whereas the 39.8 F/g value corresponds to the assembled symmetric device configuration. The synthesized tertiary composite (rGO/LC/Fe₃S₄) exhibits exceptional charge-discharge performance. The LC/rGO/Fe₃S₄ configuration has a power density 4.6 W/kg on 12 Wh/kg a specific capacitance 39.8 F/g over 2400 s. The rGO/LC/Fe₃S₄ (2400 s) electrode has superior electrochemical properties, evidenced by lower Rs (1.8 Ω) and Rct (0.3 Ω) values in comparison to the binary composite (rGO/LC). These composites offer profound understanding of the fabrication of electrodes exhibiting strong ionic and electrical conductivity for rapid energy storage systems. The electrochemical performance obtained in this study is comparable to or exceeds previously reported rGO/iron sulfide-based flexible electrodes, highlighting the effectiveness of the proposed fabrication strategy.
Evapotranspiration is a crucial process in hydrology, agriculture, and climate studies, significantly impacting crop yield, water resources, and climate modeling, and necessitating accurate forecasting for effective management. This research evaluates the effectiveness of deep learning and machine learning techniques in forecasting evapotranspiration (ET) by utilizing climatic factors, including Rainfall, Temperature, and Sunshine hours. The Long Short-Term Memory (LSTM) model exhibited exceptional capabilities in recognizing temporal relationships and seasonal patterns within the data, achieving a test R² of 0.66, an RMSE of 0.11, and an MAE of 0.07. Statistical significance testing showed that LSTM significantly outperformed RNN (p = 0.0091), and RNN outperformed GRU (p = 0.0324), while the performance difference between LSTM and GRU was not statistically significant (p = 0.2694). The bootstrap analysis revealed low standard errors for LSTM (SE = 0.0118), RNN (SE = 0.0113), and GRU (SE = 0.0114), with 95
Diabetic Retinopathy (DR) is a leading cause of preventable blindness worldwide, affecting an estimated 103 million persons with diabetes. Timely screening is critical, yet a severe global shortage of ophthalmologists renders manual fundus examination impractical at scale. This study conducts a systematic comparative evaluation of six ImageNet-pretrained convolutional neural network (CNN) architectures - DenseNet121, DenseNet169, InceptionV3, ResNet50, VGG19, and Xception - for binary DR classification from retinal fundus images using transfer learning. A dataset of 2,838 images sourced from Kaggle was used with an 80/10/10 train/validation/test split, standard augmentation, and 25 training epochs under identical protocols. All metrics are reported with DR as the positive class, which is the clinically appropriate convention. InceptionV3 achieves the highest sensitivity (94.69%) and F1 score (0.9596), identifying it as the most suitable architecture for population-level screening where minimising missed DR cases is paramount. DenseNet121 attains the highest precision (99.06%), advantageous for confirmatory diagnostic roles. ResNet50 underperforms across all metrics and is not recommended for clinical screening. Perimage inference time is additionally reported, with InceptionV3 offering the most favourable accuracy-latency trade-off among the top-performing models. These findings provide evidencebased, clinically framed guidance for CNN architecture selection in automated DR screening systems.
Efficient cooling is a key factor influencing the performance of compact electronic systems, which requires advanced thermal management. Microchannel Heat Exchangers (MCHEs) are increasingly becoming popular due to their high surface area-to-volume ratio and ability to dissipate high heat flux. The current study aims to explore the performance of MCHEs using a water-based, multiwalled carbon nanotube-blended nanofluid as coolant under different channel counts and geometries. A 3D CFD model was developed in ANSYS Fluent using the finite volume method, assuming constant thermophysical properties and steady-state, laminar flow, with uniform heat flux boundary conditions. An aluminum MCHE of dimensions 45 mm × 45 mm × 7 mm was analyzed with five channel geometries (square, circular, sawtooth, cross, and curved sawtooth) at three channel counts (5, 8, and 11). Two concentrations of multi-walled carbon nanotube (MWCNT)–water nanofluid (0.1 and 0.2%) were considered. Results exhibited that the baseline configuration (Square channel, 5 channels, water as coolant) achieved a convective heat transfer coefficient of 2008.24 W/m²K, overall heat transfer coefficient (U) of 1258.75 W/m²K, and thermal effectiveness (ε) of 0.102. The optimum configuration, curved sawtooth geometry with 5 channels and 0.2% MWCNT nanofluid, yielded h = 8271 W/m²K, U = 6590 W/m²K, and ε = 0.4765. ANOVA analysis confirmed channel geometry as the most influential factor, contributing 70.6% to effectiveness and 78.2% to the overall heat transfer coefficient.
Holmium doped lithium strontium borate glasses are synthesized by melt-quenching technique. XRD, infrared, Raman, photoluminescence, UV-Visible spectroscopic and NLO studies are carried out. Optical band gap, refractive index, optical basicity and Urbach energy are in good correlation with the glass structure. Furthermore, Judd-Ofelt (J-O) intensity (Omega(lambda) with lambda = 2, 4 and 6) parameters are estimated and suggest Omega(2) > Omega(6) > Omega(4) trend. Among the LSBH glasses, the glass with highest Ho(2)O(3 )concentration has the largest value for X. Thus, it may be suitable for laser gain medium. Bonding parameters, beta and delta suggests improvement in covalent nature of Ho3+ ion and ligand bonds. Four emission bands centred 550 nm (green), 594 nm (yellow), 645 nm (red) and 750 nm (red, Near IR) are observed in photoluminescence spectra corresponding to the I-5(8)-> F-5(4), I-5(8)-> S-5(2),F-5(4), I-5(8)-> F-5(4) and I-5(7)-> F-5(4) transitions respectively. At 645 nm, an intense emission peak is seen which corresponds to the I-5(8)-> F-5(5) transition. CIE and CCT studies suggest that the investigated glasses emit red light at 2955 K 532 nm ns pulsed laser Z-scan technique is employed to investigate non-linear absorption (NLA) properties. Non-linear absorption coefficient 'beta' is in good agreement with linear optical characteristics. The observed trends are related to the structural variations supported by Raman and FTIR spectroscopic studies. Diborate B4O5 & Oslash;(2-)(4 )pentaborate B5O6 & Oslash;(-)(4), dipentaborate B5O6 & Oslash;(2-)(5), metaborate chain [B & Oslash;O-2(-)](n) and ring type B(3)O(6 )(3-)confirmed by Raman and FTIR spectroscopy. M and W patterns observed in normalized transmittance spectra of Z-scan studies reveal saturable absorption as well as reverse saturable absorption behaviour which are sensitive to the formation of non-bridging oxygens in the glass structure. This is further supported by the conversion of dipentaborate B5O6 & Oslash;(2-)(5) and diborate B4O5 & Oslash;(2-)(4) units (which contain [B & Oslash;(-)(4)]) into metaborate B3O63- ring and( )chain type [B & Oslash;O-2(-)](n) units (which contain B & Oslash;(2)O(- )units). These glasses are suitable for red-light lasers, optical switches and optical limiters.
