Taiwan Association of Engineering and Technology Innovation: E-Journals
Not a member yet
887 research outputs found
Sort by
Quantitative Analysis of Green H2 Production Costs: A Comparison between Domestic Developed and Imported Electrolyzers
Get PDFThis study aims to present a quantitative cost analysis of hydrogen utilizing a developed alkaline electrolyzer, a similar-capacity imported alkaline electrolyzer, and a similar-capacity PEM electrolyzer.The research also finds the key parameters that can reduce or increase the production cost. One of the subjected electrolyzers is a locally developed Alkaline Electrolyzer (AE); the other two are similar-capacity imported AE and Polymer Electrolytic Membrane (PEM) electrolyzers. The study uses the Hybrid Optimization of Multiple Energy Resources (HOMER) software for estimating the Levelized Cost of Electricity (LCOE) and the Life Cycle Cost (LCC) method for hydrogen production cost estimation. Results show that the imported electrolyzers have higher production costs due to import duty, fees, and taxes. The estimated cost is 88.4% (AE) and 110.3% (PEM), higher than the locally developed electrolyzer. The economic changes also significantly impact production costs. Government policies can reduce the cost by rescheduling the hydrogen components taxes
A Novel MCDM-Based Framework to Recommend Machine Learning Techniques for Diabetes Prediction
Get PDFEarly detection of diabetes is crucial because of its incurable nature. Several diabetes prediction models have been developed using machine learning techniques (MLTs). The performance of MLTs varies for different accuracy measures. Thus, selecting appropriate MLTs for diabetes prediction is challenging. This paper proposes a multi-criteria decision-making (MCDM) based framework for evaluating MLTs applied to diabetes prediction. Initially, three MCDM methods—WSM, TOPSIS, and VIKOR—are used to determine the individual ranks of MLTs for diabetes prediction performance by using various comparable performance measures (PMs). Next, a fusion approach is used to determine the final rank of the MLTs. The proposed method is validated by assessing the performance of 10 MLTs on the Pima Indian diabetes dataset using eight evaluation metrics for diabetes prediction. Based on the final MCDM rankings, logistic regression is recommended for diabetes prediction modeling
A Domain Generalized Face Anti-Spoofing System Using Domain Adversarial Learning
Get PDFThis research addresses the enhancement of face anti-spoofing (FAS) in facial recognition systems (FRS) against sophisticated fraudulent activities. Prior methods primarily focus on extracting facial features like color, texture, and dynamic variations, yet these methods struggle to accurately identify common characteristics of forged faces, thereby limiting generalization in practical scenarios. This study aims to propose a novel representation learning framework incorporating adversarial learning algorithms, to segregate features into liveness-specific and domain-specific categories, emphasizing liveness-specific features for FAS advancement. Feature disentanglement is central to this approach, enabling the deep learning models to effectively discern separable latent generating factors, such as identity, liveness, appearance, and texture. This methodology enhances model interpretability, explainability, and generalization. Additionally, Grad-CAM is employed to elucidate the basis for classifications made by the architecture, increasing explainability and trustworthiness. Empirical evaluation across panoply available FAS datasets confirms the superiority, significantly improving performance and robustness over existing technologies
BAT Algorithm-Based Multi-Class Crop Leaf Disease Prediction Bootstrap Model
Get PDFIn the task of identification of infected agriculture plants, the leaf-based disease identification technique is especially effective in better understand crop disease among various techniques to detect infection. Recognition of an infected leaf image from healthy images gets encumbered when the model is required to detect the type of leaf disease. This paper presents a BAT-based crop disease prediction bootstrap model (BCDPBM) that identifies the health of the leaf and performs disease prediction. The BAT algorithm in the proposed model increases the capability of the Gaussian mixture model for foreground region detection. Furthermore, in the work, the co-occurrence matrix feature and histogram feature are extracted for the training of the bootstrap model. Hence, leaf foreground detection by the BAT algorithm with the Gaussian mixture improves the feature extraction quality for bootstrap learning. The proposed model utilizes a dataset of real leaf images for conducting experiments. The results of the model are compared with different existing models across various parameters. The results show the prediction accuracy enhancement of multiclass leaf disease using the BCDPBM model
Integrating Gamification Elements into a Personalized Cognitive Mobile-Learning LINE Bot
Get PDFIn recent years, chatbots gains widespread popularity across various industries, and LINE becomes an indispensable and widely utilized application. Human beings acquire knowledge through cognitive learning. Asynchronous digital drills and practice learning systems that require students to practice questions repeatedly can bore students and lack online monitoring by a teacher. In this study, the cognitive mobile-learning LINE bot provides digital drill and practice learning functions, enabling students to read questions and their answers from a Q&A database, take a postlearning self-test on these questions, and practice questions they originally answered incorrectly. Moreover, learners can ask open-ended questions. The LINE bot is used to substitute for a teacher in one-on-one synchronous interactive learning, and the post-hoc analyses of the interactions between the LINE bot and each student are performed and provided to teachers on time, enabling them to offer counseling and assistance as appropriate
A 20-GHz On-Chip Six-Port Reflectometer Using Simple Lumped Passive Devices and Bipolar Junction Transistors
Get PDFThis paper proposes an on-chip six-port reflectometer (SPR) fabricated in the 0.13-μm IBM BiCMOS-8HP technology. The SPR enjoys a compact circuit structure, with only four amplitude detectors as active devices, one resistive power divider, and one lumped phase shifter as passive devices. The power divider and phase shifter are responsible for manipulating the radio-frequency (RF) signals appropriately, whereas the detectors are responsible for sensing the processed signals. The chip area, which can be further reduced, is 1.25 mm in width and 1 mm in height. The SPR can perform in-situ measurement of reflection coefficients of devices under test (DUTs) and reduce testing costs of RF chips by using vector network analyzers (VNAs). The SPR demonstrates excellent performance in measuring the reflection coefficients of DUTs at around 20 GHz. The experimental results indicate that the maximum error of the measured reflection coefficients in absolute value is about -26 dB
Simulation and Measurement Analysis of an Integrated Flow Battery Energy-Storage System with Hybrid Wind/Wave Power Generation
Get PDFThis study aims to evaluate the power-system stability and the mitigation of fluctuations in a hybrid wind/wave power-generation system (HWWPGS) under different operating and disturbance conditions. This evaluation is performed by employing a vanadium redox flow battery-based energy storage system (VRFB-ESS) as proposed. The measurement results obtained from a laboratory-scale HWWPGS platform integrated with the VRFB-ESS, operating under specific conditions, are used to develop the laboratory-scale simulation model. The capacity rating of this laboratory-scale simulation model is then enlarged to develop an MW-scale power-system model of the HWWPGS. Both operating characteristics and power-system stability of the MW-scale HWWPGS power system model are evaluated through frequency-domain analysis (based on eigenvalue) and time-domain analysis (based on nonlinear-model simulations) under various operating conditions and disturbance conditions. The simulation results demonstrate that the fluctuations and stability of the studied HWWPGS under different operating and disturbance conditions can be effectively smoothed and stabilized by the proposed VRFB-ESS
Optimization of SM4 Encryption Algorithm for Power Metering Data Transmission
Get PDFThis study focuses on enhancing the security of the SM4 encryption algorithm for power metering data transmission by employing hybrid algorithms to optimize its substitution box (S-box). A multi-objective fitness function is constructed to evaluate the S-box structure, aiming to identify design solutions that satisfy differential probability, linear probability, and non-linearity balance. To achieve global optimization and local search for the S-box, a hybrid algorithm model that combines genetic algorithm and simulated annealing is introduced. This approach yields significant improvements in optimization effects and increased non-linearity. Experimental results demonstrate that the optimized S-box significantly reduces differential probability and linear probability while increasing non-linearity to 112. Furthermore, a comparison of the ciphertext entropy demonstrates enhanced encryption security with the optimized S-box. This research provides an effective method for improving the performance of the SM4 encryption algorithm
Selection of Elevation Models for Flood Inundation Map Generation in Small Urban Stream: Case Study of Anyang Stream
Get PDFTo reduce flood damages, the Ministry of Environment in Korea has provided a flood inundation map so that people can expediently identify flood-prone areas. However, the current flood inundation maps have been produced based on the DEM which makes it difficult to represent realistic situations due to the lack of reproduction of land surface conditions. This study aims to provide more accurate and detailed flood inundation maps for flooding events due to river overflow in small urban areas. In this study, flood inundation analysis is performed using the river analysis system, HEC-RAS 2D, with the DSM and the DEM of urban areas in the Anyang Stream Basin, Korea to examine the differences in terms of terrain data and flooded area. Finally, for urban areas with dense buildings and congested road networks, the flood inundation analysis based on DSM can represent a more realistic flood situation and create an appropriate flood inundation map
Food Waste Management Utilizing Black Soldier Fly Larvae
Get PDFFood waste is a growing concern in developing countries. This study aims to implement food waste bioconversion by utilizing black solider fly larvae for two eateries' food waste. The bioconversion process used 0.5 g of black solider fly eggs for 14 days in the six bio ponds. After 14 days, the waste, larvae, and compost are separated using sieves to measure the larvae and compost production. The bioconversion process is evaluated based on bioconversion characteristics and black soldier fly larvae and compost produced. Waste Reduction Index, Fresh Matter Consumption Rate, Dry Matter Consumption Rate, Dry Matter Rate, and Efficiency of Conversion of Digested Feed evaluated the bioconversion characteristics for reduction. According to the experimental results, utilizing BSFL is adequate for food waste management, effectively reducing up to 62.6%. Simultaneously, the fresh larvae and compost are produced within a 14-day bioconversion process. The compost meets standards for the nitrogen, C/N ratio, phosphorus, potassium, zinc, and iron content (SNI 19-7030-2004)