Taiwan Association of Engineering and Technology Innovation: E-Journals
Not a member yet
887 research outputs found
Sort by
Dimensional Assessment and Error Analysis of 3D CAD Models Manufactured with Bound Metal Deposition
Full text linkThis study evaluates dimensional deviations of 316L stainless steel components produced using the Bound Metal Deposition (BMD) method. The evaluation covers four production stages: CAD modeling, slicing, printing, and sintering. The approach involves experimental measurements and finite element simulations to assess accuracy. Results indicate an average dimensional deviation of 6%, with height-related features showing the most significant errors due to anisotropic shrinkage and inconsistencies in layer deposition. The widely used 1.16 scale factor is inadequate for precise dimensional recovery. Instead, a revised scale factor of 1.13 is developed through empirical analysis and validated using ANSYS simulation. This new factor shows better agreement with the original CAD design. Furthermore, the achieved dimensional accuracy falls within clinically accepted tolerances for dental implants (0.19-0.36 mm), making it suitable for biomedical applications. Overall, these findings provide a validated framework for dimensional calibration in BMD, improving the accuracy of patient-specific medical devices
Evolutionary Tuner and Selective Kernel Attention for Improving YOLOv11 in Underwater Fish Detection and Recognition
Full text linkThis study aims to enhance the accuracy of underwater fish detection by proposing a dual-enhanced YOLOv11 model. The approach leverages two key improvements: First, a selective kernel attention (SKA) mechanism is incorporated into the YOLOv11 architecture to enable dynamic selection of multi-scale convolution kernels, improving adaptability to various target sizes. Second, an evolutionary tuner (ET) is employed for hyperparameter optimization to refine model performance further. The proposed model achieves significant gains over the baseline, with improvements of 2.06% in mean average precision (mAP)@0.5 and 6.30% in [email protected]:0.95, attaining final scores of 98.629% and 86.933%, respectively. The dual-enhanced model demonstrates superior accuracy and robustness in complex underwater environments, ultimately achieving a precision of 99.069% and a recall of 95.968%
Paraffin/Bamboo Carbon Composites for Electrical Vehicle Battery Thermal Management
Full text linkThis study aims to develop a paraffin-based phase change material (PCM) modified with bamboo waste carbon (BWC). The modification is intended to enhance the performance of a passive battery thermal management system for electric vehicles. PCM composites containing 0, 5, 10, and 15 wt.% BWC are prepared and characterized using simultaneous thermal analysis, fourier transform infrared spectroscopy, and scanning electron microscopy. The thermal performance of the composites is evaluated through battery module cooling simulations under a constant discharge load. The simulations assess the ability of the PCM composites to maintain battery operating temperatures within a safe range (293.15–313.15 K). The results indicate that the composite containing 10 wt.% BWC achieves optimal performance, with a 33.9% increase in thermal energy absorption. The peak battery temperature is reduced to 309.03 K. These findings demonstrate BWC’s potential as an effective and sustainable thermal additive for paraffin-based PCMs in passive BTMS applications for electric vehicles
Assessment of Hot-Water-Alkali Treated Bagasse Fiber in Metakaolin-Based Geopolymer Using Machine Learning
Full text linkThis study aims to develop metakaolin-based geopolymers reinforced with sugarcane bagasse fiber (BF) and to evaluate the effect of BF treatment on the composite's performance. The BF is pretreated with hot water and sodium hydroxide before being incorporated into the geopolymer matrix. Metakaolin-based geopolymer specimens containing 0%, 3%, 4%, and 5% BF by weight are prepared, and their mechanical properties and water absorption are analyzed. Scanning electron microscopy and Fourier transform infrared spectroscopy analyses reveal that the combined hot-water-alkali treatment significantly modifies the fiber surface. The treatment removes impurities and increases surface roughness, thereby enhancing fiber–matrix bonding. As a result, this treatment improves compressive and splitting tensile strength (STS) while reducing water absorption compared to untreated BF. Furthermore, machine learning algorithms, including random forest, AdaBoost, and XGBoost, are applied to predict STS. Among the three models, XGBoost demonstrates the highest predictive accuracy (R^² = 0.95, MAE = 0.28), indicating reliable predictions of mechanical strength
Development of an Electromagnetic Pollution Rating Index for Buildings
Full text linkRapid technological development and the intensive use of electrical energy in buildings create electromagnetic pollution. This pollution is further intensified by external radiation from base stations and transmission lines. The World Health Organization (WHO) classifies electromagnetic field (EMF) exposure as a Class 2B carcinogen. To address this, this study proposes a novel electromagnetic pollution rating index (EPRI). This index integrates into green building certification systems, such as leadership in energy and environmental design (LEED) and building research establishment environmental assessment method (BREEAM). The EPRI methodology applies risk reduction factors (10–1000×) to international commission on non-ionizing radiation protection (ICNIRP) reference limits. A case study conducted in a 7,600 m² municipal building involves 760 measurements at extremely low frequency (ELF) and radio frequency (RF) ranges. Results reveal maximum values of 4.9 V/m and values of 0.43 µT. These qualify for the highest A+++ rating, demonstrating the practical applicability of EPRI in certifying electromagnetically clean indoor environments
Classification of Leftover Shrimp Feed Based on Lift Net Design Utilizing the k-Nearest Neighbors Algorithm
Full text linkAn effective Feeding Management System (FMS) is crucial in shrimp farming, as both overfeeding and underfeeding can adversely affect shrimp growth. To ensure optimal nutrition, an accurate FMS must account for factors such as shrimp size, weight, age, and leftover feed. This study presents a method for detecting leftover shrimp feed using custom-designed lift nets equipped with paired ultrasonic sensors. Two critical aspects are examined: the optimal timing for measurement and the ideal placement of the transmitter. Results show that measurements should be taken within 10 minutes of feed immersion to avoid feed disintegration. Additionally, placing the transmitter on the outer side of the lift net improves measurement accuracy. Ultrasonic echoes are analyzed to classify leftover feed using the k-Nearest Neighbors algorithm. Root Mean Square voltage-based classification effectively groups leftover feed into five classes, highlighting its potential to improve aquaculture feed management
Smart IoT Irrigation System Using PID-PSO Optimization Method
Full text linkAn irrigation system stands as one of the most efficient real systems of sustainable water and farming fields. This work aims to enhances the proportional-integral-derivative (PID) controller by using particle swarm optimisation (PSO), which relies on the internet of things (IoT) for a smart irrigation system. The system employs a PID controller to dynamically adjust water flow by integrating environmental data, including humidity, soil moisture, and temperature, gathered by IoT sensors. PSO is utilised to optimize the PID parameters and overcome the limitations of traditional PID tuning. Furthermore, the proposed work improved stability, reduced overshoot, and provided faster response times. The experimental results indicated significant gains in crop health and water use efficiency. The moisture stabilizes and maintains the target of 60% with the optimized PID parameters in the case study. The smart system assisted in managing water resources sustainably by providing a scalable and energy-efficient precision agriculture solution
Real-Time Table Availability Detection in Dynamic Dining Environments Using YOLOv8 and Geometric Overlap Analysis
Full text linkThis study addresses the challenge of determining table availability in dynamic restaurant environments. Customer mobility, visual obstruction, and irregular table configurations make direct visual classification ineffective. A real-time table availability detection system utilizing YOLOv8 and simple online and real-time tracking (SORT) is proposed to address these difficulties. The primary innovation is a geometric overlap-based inference technique for table status assessment. The method examines the spatial link between customer centroids and table polygons. Centroid area expansion is applied to mitigate bounding box noise. The system is evaluated using an annotated dataset and compared with a direct detection baseline. Experimental findings indicate that the method attains an accuracy of 91.08%, markedly surpassing the baseline accuracy of 35.75%. Real-time performance assessment indicates a processing speed of 14.58 FPS with a latency of 68.57 ms during CPU-only execution. This performance satisfies real-time criteria. The study demonstrates that the method offers a dependable and efficient alternative for automated table availability monitoring
Intelligent TNVR Ear-Tag Recognition and Monitoring System for Stray Animal Management
Full text linkThis study aims to enhance stray animal management by improving the efficiency and sustainability of the trap-neuter-vaccinate-return (TNVR) process. An intelligent monitoring system integrating image recognition and radar sensing is proposed for real-time detection and identification. The system utilizes a domain-specific ear-tag recognition model based on OpenCV preprocessing and YOLOv8, achieving an accuracy of 91% under various environmental conditions. Captured data are automatically uploaded via 4G to a centralized server, supporting continuous monitoring and instant alerts. Designed for high-density urban settings, the system mitigates manual workload and enhances decision-making efficiency, contributing to sustainable and humane stray animal control. Although the proposed system demonstrates high detection accuracy and robust performance under real-world conditions, the current evaluation is conducted on a moderate-scale dataset; future work will focus on large-scale deployment and cross-context validation to further examine system generalizability
Remaining Useful Life Prediction of Milling Tool Based on Improved PSO-MultiAM-BiLSTM
Full text linkTo improve the accuracy of remaining useful life (RUL) prediction for milling tools, this study proposes an enhanced PSO-MultiAM-BiLSTM model integrating particle swarm optimization (PSO), multi-head attention mechanism (MultiAM), and bidirectional long short-term memory (BiLSTM). The model captures key information in input sequences, alleviating early feature attenuation in BiLSTM from “chain propagation.” A logarithmic decreasing strategy adjusts PSO inertia weights, balancing global and local searches while optimizing BiLSTM parameters. Validated on the PHM2010 dataset, the model attains an average coefficient of determination of 0.97, with average root-mean-square error and mean absolute error of 0.062 and 0.045, improving prediction accuracy by 9.64% and 4.06% over MultiAM-BiLSTM and PSO-AM-BiLSTM, respectively. Such a result attests to the effective extraction of degradation features of tools and provides a valuable reference for predicting the RUL of milling tools