Taiwan Association of Engineering and Technology Innovation: E-Journals
Not a member yet
887 research outputs found
Sort by
Swarm Intelligence Algorithm Based on Plant Root System in 1D Biomedical Signal Feature Engineering to Improve Classification Accuracy
Get PDFThe classification accuracy of one-dimensional (1D) biomedical signals is limited due to the lack of independence of the extracted features. To address this shortcoming, the study applies a swarm intelligence algorithm based on plant root systems (PRSs) to feature engineering. Some basic features of 1D biomedical signals are integrated into a digitized soil, and a root matrix is generated from this digitized soil and the PRS algorithm. The PRS features are extracted from the root matrix and used to classify the basic features. Following classification with the same biomedical signals and classifier, the accuracy of the added PRS set is generally higher than that of the base set. The result shows that the proposed algorithm can expand the application of 1D biomedical signals to include more biomedical signals in classification tasks for clinical diagnosis
An Improved MobileNet for Disease Detection on Tomato Leaves
Get PDFTomatoes are widely grown vegetables, and farmers face challenges in caring for them, particularly regarding plant diseases. The MobileNet architecture is renowned for its simplicity and compatibility with mobile devices. This study introduces MobileNet as a deep learning model to enhance disease detection efficiency in tomato plants. The model is evaluated on a dataset of 2,064 tomato leaf images, encompassing early blight, leaf spot, yellow curl, and healthy leaves. Results demonstrate promising accuracy, exceeding 0.980 for disease classification and 0.975 for distinguishing between diseases and healthy cases. Moreover, the proposed model outperforms existing approaches in terms of accuracy and training time for plant leaf disease detection
Flexural Behavior of Built-Up Cold-formed Steel Channel Section Strengthened with Oriented Strand Board
Get PDFThe objective of the study is to determine the flexural behavior of the built-up cold-formed steel (CFS) channel section strengthened with an oriented strand board (OSB) in the three-point bending experiment. CFS with a variety of shapes and grades is classified as a steel-based material and exposed to buckling failure when subjected to compression or flexural load. Thus, the CFS channel section with 100 mm of the web, 50 mm of the flange, 12 mm of the lip, and 1.55 mm of thickness has been selected. Then, the built-up CFS channel section is designed by filling with an OSB between the gap of each section. Channel, face-to-face built-up, and back-to-back built-up CFS sections are three types of tested specimens. From the result and discussion, the specimen with back-to-back built-up CFS section is recognized to sustain the ultimate load with the highest value when compared with other specimens
Flexural Strength and Porosity of NaOH-Treated Maize Stalk Cellulose-Fibers-Reinforced Geopolymer Composites
Get PDFThis study characterizes the flexural strength and porosity of NaOH-treated maize stalk cellulose-fibers-reinforced geopolymer composites. Flexural strength tests are conducted, and the fracture surfaces of the composite and geopolymer powder are observed using a scanning electron microscope (SEM). Moreover, porosity analysis is also performed using Image J software from SEM images. The formation of geopolymer is confirmed using X-ray diffraction (XRD) and Fourier transform infrared (FTIR) analysis. The addition of 1.5 wt% of NaOH-treated maize stalk cellulose fibers improves flexural strength by 2.4 times. The results show that the main failure mechanisms, namely fiber breakage, fiber pullout, and debonding of the fiber and matrix, can increase flexural strength and reduce failures during service life. During the analysis for fiber and particle pullout, SEM images under 25^2 pixels of pore areas are not considered, and an average porosity of 36.7% is achieved
DV-EXCCCII Based Resistor-Less Current-Mode Universal Biquadratic Filter
Get PDFThis study aims to present a new resistor-less current-mode multi-input single-output universal filter. The current-mode’s design approach is used to obtain the proposed circuit. This circuit employs a single differential voltage extra-X current controlled current conveyor (DV-EXCCCII) and two grounded capacitors. This multifunction filter circuit offers low-pass, high-pass, all-pass, band-pass, and band-reject filters at a single output terminal without passive component matching constraints. The same circuit topology can obtain all second-order filter functions with different input conditions. The proposed circuit design is electronically adjustable with the bias current of DV-EXCCCII. Because of its high output impedance, this arrangement is suitable for cascading other current-mode circuits. The proposed circuit is simulated by Cadence Spectre with 0.18 µm UMC CMOS technology process parameters at ± 0.9 V supply voltages. The simulation results agree well with the theoretical concept of the proposed circuit
An Image-Based Rice Weighing Estimation Approach on Clock Type Weighing Scale Using Deep Learning and Geometric Transformations
Get PDFAI impacts surrounding human life, such as the economy, health, education, and agricultural production; however, the crop prices in the harvest season are still on manual calculation, which causes doubts about accuracy. In this study, an image-based approach is proposed to help farmers calculate rice prices more accurately. YOLOv5 is used to detect and extract the scales in the images taken from the harvesting of rice crops. Then, various image processing techniques, such as brightness balance, background removal, etc., are compiled to determine the needle position and number on the extracted scale. Lastly, geometric transformations are proposed to calculate the weight. A real dataset of 709 images is used for the experiment. The proposed method achieves good results in terms of [email protected] at 0.995, mAP@[0.5:0.95] at 0.830 for scale detection, and MAE at 3.7 for weight calculation
The Performance of Machine Learning for Chronic Kidney Disease Diagnosis
Get PDFThis paper aims to review the performance of different machine learning (ML) models and develop models for the automated diagnosis of chronic kidney disease. To detect chronic kidney disease with better precision, selecting the right and better-performing ML model is significant as it improves the precision and accuracy of the chronic kidney disease diagnosis. The study uses the Joana Briggs Institute (JBI) scoping review methodology, which involves different steps such as searching relevant literature, conducting the review, and reporting the review result. In the search, the year of publication and the indexing of journals where the studies are published is used as inclusion and exclusion criteria. The review result shows that the current chronic kidney disease detection has focused on the development of ensemble-based and deep-learning methods. The deep learning method can achieve a higher accuracy of 99.75%
Using Feedback Control to Control Rotor Flux and Torque of the DFIG-Based Wind Power System
Get PDFDirect torque control (DТС) is a method of controlling electrical machines that are widely used, and this is due to its simplicity and ease of use. However, this method has several issues, such as torque, rotor flux, and current fluctuations. To overcome these shortcomings and improve the characteristics and robustness of the DTC strategy of the doubly-fed induction generator (DFIG), a new DTC scheme based on the feedback control method (FCM) and space vector modulation (SVM) is proposed. In the proposed DTC technique, a proportional-integral controller based on feedback control theory is used to control and regulate the torque and rotor flux of the DFIG. On the other hand, the SVM technique is used to control the rotor side converter (RSC) to obtain a high-quality current. The simulation result shows that the proposed DTC technique has the advantages of faster dynamics and reduced harmonic distortion of current compared to the conventional technique.
A Hybrid Metaheuristic Algorithm for Stop Point Selection in Wireless Rechargeable Sensor Network
Get PDFA wireless rechargeable sensor network (WRSN) enables charging of rechargeable sensor nodes (RSN) wirelessly through a mobile charging vehicle (MCV). Most existing works choose the MCV’s stop point (SP) at random, the cluster’s center, or the cluster head position, all without exploring the demand from RSNs. It results in a long charging delay, a low charging throughput, frequent MCV trips, and more dead nodes. To overcome these issues, this paper proposes a hybrid metaheuristic algorithm for stop point selection (HMA-SPS) that combines the techniques of the dragonfly algorithm (DA), firefly algorithm (FA), and gray wolf optimization (GWO) algorithms. Using FA and GWO techniques, DA predicts an ideal SP using the run-time metrics of RSNs, such as energy, delay, distance, and trust factors. The simulated results demonstrate faster convergence with low delay and highlight that more RSNs can be recharged with fewer MCV visits, further enhancing energy utilization, throughput, network lifetime, and trust factor
Green Building Materials for Circular Economy - Geopolymer Foams
Get PDFThis study aims to design and investigate foamed geopolymers as a green material dedicated to the circular economy. For synthesis as raw material, the main waste materials of two Polish coal mines, Wieczorek and Staszic, are applied. Additionally, various foaming methods are employed to utilize the by-product of energy production, especially the fly ash generated by the Skawina power plant. In this study, the main issues addressed are related to the selection of the most appropriate foaming agent and the optimization of the process parameters, including temperature, time, and mixture components. Hydrogen peroxide, aluminum powder, and a commercial foaming agent are selected as foaming agents in this research. During the process of sample preparation, stabilizers are applied in the form of polyglycol and cellulose. Through the conducted test, the results show that hydrogen peroxide and aluminum powder emerged as the two most optimal foaming agents