Taiwan Association of Engineering and Technology Innovation: E-Journals
Not a member yet
887 research outputs found
Sort by
Preparation and Characterization of Carrot Nanocellulose and Ethylene/Vinyl Acetate Copolymer-Based Green Composites
Get PDFThis study aims to investigate the effect of nanocellulose on the properties and physical foaming of ethylene/vinyl acetate (EVA) copolymer. The nanocellulose is prepared from waste carrot residue using the 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) oxidation method (CT) and is further modified through suspension polymerization of methyl methacrylate (MMA) monomer (CM). The obtained nanocellulose samples (CT or CM) are added to EVA to create a series of nanocomposites. Moreover, the EVA and CM/EVA composite were further foamed using supercritical carbon dioxide physical foaming. TEM results show that the average diameters of CT and CM are 24.35 ± 3.15 nm and 30.45 ± 1.86 nm, respectively. The analysis of mechanical properties demonstrated that the tensile strength of pure EVA increased from 10.02 MPa to 13.01 MPa with the addition of only 0.2 wt% of CM. Furthermore, the addition of CM to EVA enhanced the melt strength of the polymer, leading to improvements in the physical foaming properties of the material. The results demonstrate that the pore size of the CM/EVA foam material is smaller than that of pure EVA foam. Additionally, the cell density of the CM/EVA foam material can reach 3.23 × 1011 cells/cm3
Self-Sensing Potential of Metashale Geopolymer Mortars with Carbon Fiber/Graphite Powder Admixtures
Get PDFMultifunctional building materials with self-sensing capability have great potential for civil engineering applications. The self-sensing capability of typically calcium aluminosilicate matrices of cementitious or geopolymer materials is adopted by admixing electrically conductive admixtures in an amount that ensures optimal electrical properties and their proportionality to mechanical loading. The paper aims to evaluate the self-sensing capability of 4 metashale geopolymer mortars with graphite powder (GP) and carbon fibers (CF) in different ratios, including MGF 5/0, MGF 4.5/0.5, MGF 4/1, and MGF 3/0. The 4-probe measurements at 21 V DC input voltage on (100 × 100 × 100) mm3 samples with embedded copper-grid electrodes evaluate the gauge factor, which corresponds to the monitored changes in electrical resistivity. Despite the limitations of DC measurements, the self-sensing capability is observed for all the mixtures. The most promising response to dynamic loading with an FCR of 0.018%, is observed for the MGF 4.5/0.5 sample
Prediction of Distribution Network Line Loss Rate Based on Ensemble Learning
Get PDFThe distribution network line loss rate is a crucial factor in improving the economic efficiency of power grids. However, the traditional prediction model has low accuracy. This study proposes a predictive method based on data preprocessing and model integration to improve accuracy. Data preprocessing employs dynamic cleaning technology with machine learning to enhance data quality. Model integration combines long short-term memory (LSTM), linear regression, and extreme gradient boosting (XGBoost) models to achieve multi-angle modeling. This study employs regression evaluation metrics to assess the difference between predicted and actual results for model evaluation. Experimental results show that this method leads to improvements over other models. For example, compared to LSTM, root mean square error (RMSE) was reduced by 44.0% and mean absolute error (MAE) by 23.8%. The method provides technical solutions for building accurate line loss monitoring systems and enhances power grid operations
Recognition of Ginger Seed Growth Stages Using a Two-Stage Deep Learning Approach
Get PDFMonitoring the growth of ginger seed relies on human experts due to the lack of salient features for effective recognition. In this study, a region-based convolutional neural network (R-CNN) hybrid detector-classifier model is developed to address the natural variations in ginger sprouts, enabling automatic recognition into three growth stages. Out of 1,746 images containing 2,277 sprout instances, the model predictions revealed significant confusion between growth stages, aligning with the human perception in data annotation, as indicated by Cohen’s Kappa scores. The developed hybrid detector-classifier model achieved an 85.50% mean average precision (mAP) at 0.5 intersections over union (IoU), tested with 402 images containing 561 sprout instances, with an inference time of 0.383 seconds per image. The results confirm the potential of the hybrid model as an alternative to current manual operations. This study serves as a practical case, for extensions to other applications within plant phenotyping communities
A Novel Hybrid Approach for Feature Selection in Cardiovascular Risk Assessment
Get PDFEarly detection of cardiac risk is crucial for accurate diagnosis and treatment of fatal cardiovascular diseases. Selecting relevant features is essential for machine learning in building an effective decision support system of cardiovascular risk assessment, ensuring accuracy of high-dimensional data. This study aims to propose a novel hybrid feature selection approach, termed ant colony optimization with hill climbing (ACOHC), integrating ant colony optimization (ACO) and hill climbing (HC) algorithms. The accuracy metric and various classifiers are deployed to evaluate the effectiveness. Additionally, comparisons are made with nine alternative feature selection techniques. The feature subset identified through the ACOHC attains a classification accuracy of 95.1% with the support vector machine classifier
Performance Evaluation of Neural Network Models for Autism Detection Using EEG Data
Get PDFThis study aims to leverage a promising avenue for the precise and early detection of Autism. Autism is a multifaceted neurodevelopmental condition marked by challenges in social interaction, communication, and repetitive behaviors. Traditional diagnosis relies on time-consuming behavioral assessments, necessitating reliable and non-intrusive biomarkers for early and accurate detection. This paper analyzes eleven linear and non-linear features across time and frequency domains from an EEG dataset. Four neural network models, such as convolutional neural network (CNN), deep neural network (DNN), long short-term memory (LSTM), and a custom neural network are employed for classification. The CNN achieves the lowest accuracy at 89.02%, while the custom neural network reaches the highest accuracy at 94.02%, and the DNN and LSTM achieve 91.98% and 93.83% accuracy, respectively. Other metrics such as precision, recall, specificity, and F1-score, are also evaluated. This research underscores the efficacy of neural network in detecting Autism, advancing diagnostic tools
Possibility of Using a Geopolymer Containing Phase Change Materials as a Sprayed Insulating Coating - Preliminary Results
Get PDFGeopolymers have been known for decades and classified as inorganic polymers, characterized by high resistance to high temperatures. They can be successfully used for the thermal insulation of buildings, especially in the foamed form. The addition of phase change materials (PCMs) in such materials may also increase the heat capacity of the materials, therefore, using them for building cladding can increase the thermal comfort of the building and prevent it from overheating. This study tests the addition of PCMs to geopolymers by spraying and presents the results. Additionally, the study includes preliminary experience concerning the technology of applying these materials, along with selected test results that assess the properties of the produced coatings. The results indicate that the addition of PCMs in the amount of 15% can increase the heat capacity of geopolymer materials by about 150-180%, and the foamed geopolymer coatings produced have a thermal conductivity in the range of 0.07-0.09 W/mK
Environmental Odor Analysis in West and East Java’s Ambient Air and Odor Reduction Using Biofilter Model
Get PDFThe odor affects both one’s health and quality of life. This study measures and analyzes odor concentration and odor sources in ambient air, the correlation between odor gas concentration and the hedonic scale, and the design of an odor-reduction instrument. The research commenced from February to May 2022 in the small industrial area (SIA) of Magetan Regency and compost bins of Bogor City. Data was collected through chemical analysis, and the hedonic scale was measured at four points divided into radii one and two. The concentration of odor parameters in Magetan and Bogor City is below the quality standard, while the correlation between ammonia gas concentration and the hedonic scale is low. Regarding the biofilter, its odor reduction efficiency is 35% for rotten fish, 70% for goat manure, 82% for compost waste, and 47% for chicken carcasses
Multifeatured Electronic Helmet to Enhance Road Safety and Rider’s Comfort
Get PDFThis paper presents a multi-featured electronic helmet designed to tackle critical road safety issues, including accidents, drunk driving, and over-speeding. The design incorporates a global system for mobile communications (GSM) and global positioning system (GPS) modules to accurately capture driver’s current location and send messages to predefined contacts. When encountering an accident, the helmet promptly notifies designated contacts and authorities, ensuring swift assistance. By integrating features for detecting over-speeding, accidents, and drunk driving, the helmet renders real-time alerts to both the driver’s family and traffic police, enhancing accountability and safety measures. Additionally, the helmet includes solar charging functionality for mobile devices receive 100% charged in 3.37 hours, thereby optimizing usability and emergency communication. A rain detection system protects mobile devices, while an internal warming mechanism caters to adverse weather conditions, and proffers enhanced comfort by maintaining internal form temperature between 26-27 ℃ and safety, especially in cold regions or for military personnel
Generalized and Improved Human Activity Recognition for Real-Time Wellness Monitoring
Get PDFHuman activity categorization using smartphone data can be useful for physicians in real-time data monitoring in sports or lifestyle monitoring. The goal of this research is to develop a methodology that can identify strong machine-learning classifiers applied to various human activity datasets. The first step is pre-processing the data, followed by feature extraction, selection, and classification. Relying on a single dataset does not yield high confidence in the findings. Instead, examining multiple datasets is crucial for a comprehensive understanding, as it avoids the pitfalls of basing conclusions on one dataset alone. Multiple datasets and classifiers are applied in different experiments to achieve improved and generalized human activity recognition performance. Experimental results of the support vector machine (SVM) with its generalized performance of 99% encourage us to use the trained SVM-based model to monitor normal human activities inside the home, in the park, in the gym, etc. enhancing wellness monitoring