Journal of Engineering and Digital Technology (JEDT)

Biochar and Wood Vinegar Production from Pineapple Leaves Using Low-emission 200-liter Kiln and the Feasibility Study of Production

2023-09-12T08:34:13+07:00

This research aimed to develop a 200-liter iron kiln to produce biochar and wood vinegar from pineapple leaves provided from Khao Khan Song sub-district, Sriracha district, Chon Buri province. The preliminary experiments were conducted in the laboratory scale to determine the optimum operating temperature and time for biochar production. The experimental results showed that the highest solid yield was 33.02 wt.%, obtained at 400 °C and for 1 h, with the highest heating value as 23.25 MJ/kg. The 200-liter iron kiln was then developed consisting of 3 parts: biochar production kiln, condensation apparatus for storing wood vinegar and combustion kiln for heat production. It was found that the developed 200-liter iron kiln could produce biochar and wood vinegar from pineapple leaves. The 20 kg of pineapple leaves can produce approximately 2 kg of biochar and 10 kg of wood vinegar. The analysis of biochar revealed that biochar had 70.49 wt.% of Carbon content with the highest heating value of 25.97 MJ/kg. The generated wood vinegar is mainly composed of Phenols and Phenol derivatives, including Guaiacol which has antifungal and antibacterial properties. The wood vinegar could be used in agricultural areas to eliminate root rot disease and plant fungi. It could also be sprayed in order to inhibit germs growth and some pests as well. The analysis of costs, the possibility of product creation and implementation presented that the number of payback operations was 16 and resulted in a wide variety of value-added products.

Classification of Rice Varieties from Milled Rice Grain Images by Object Detection Method

2023-05-28T21:46:37+07:00

This research has applied the YOLOv5 object detection model to help classify rice varieties from images of milled rice grains from the following varieties: Karacadag, Jasmine, Ipsala, Basmati, and Arborio. The research was divided into three main parts: data engineering, which involved developing a Python program to prepare data for artificial intelligence learning; data science operations using Python programming in conjunction with Google Colaboratory for milled rice grain detection; and the development of model accuracy evaluation method. In the data preparation phase, single-grain JPEG images were obtained from https://www.muratkoklu.com/datasets/, and noise reduction, background removal, and conversion to PNG format were performed. These images were then placed into 800 x 800 pixels images, each containing 20–64 grains, with varying degrees of overlapping: no overlap, and 5, 10, 15, 20, and 25 percent overlap. The non-overlapping images were used to train the YOLOv5 model, which was then used to classify rice varieties and identify the locations of various milled rice grains in the images. The research results showed that the YOLOv5 model could effectively classify all five rice varieties. Evaluating the model's accuracy at a threshold of 0.6, it was found that the model could correctly classify rice varieties in images with 0, 5, 10, 15, 20, and 25 percent grain overlap, with accuracy rates of 99.13, 99.00, 98.62, 98.19, 97.56, and 96.89 percent, respectively.

Developing a Credit Approval Determination Model Using Principal Component Analysis with Machine Learning Techniques

2023-05-29T09:41:06+07:00

This research presents a model and tests the performance of a credit approval determination model using principal component analysis with machine learning techniques. Four algorithms were tested: Naïve Bayes, Decision Tree, Random Forest, and Support Vector Machine. 441,335 examples of lending data. There are 4 model performance test results: Accuracy, Precision, Recall, and F-Measure.
The result of Accuracy from the experiment found that Random Forest model The best efficiency is 92.9%, followed by Support Vector Machine and Decision Tree was 87.0% and Naïve Bayes was 83.4%, respectively. It was found that reducing the data dimensions resulted in improved model performance by eliminating insignificant features and solves the problem of the traditional credit approval model that considers the old attribute variables. The model has improved performance by considering new attribute variables.
The results of the completeness Recall and the overall performance F-Measure from the experiment showed that Random Forest model provides the best performance as well as Accuracy was 99.64% and 99.35%, respectively, and the highest Precision was 99.07% for Random Forest Model and Decision Tree.

Economic Rebar Order Quantity Model for Construction Project

2023-04-09T17:14:21+07:00

Planning and ordering construction materials is a highly complex problem in the construction industry as it has to order the right type of material in the right quantity and at the right time. Otherwise it will cause losses and consequences. This study is to develop an Economical Order Quantity Model for Construction Projects. By reducing the limitation of the current method to calculate by ordering the same amount every time and have the same frequency this may result in higher expenses, for example the cost of storing the remaining rebar after ordering in excess of daily demand. Therefore, the researcher has studied and developed a Modified Economic Order Quantity Model in order to calculate the timing of ordering rebar in accordance with the project schedule to reduce the ordering cost, holding cost, delivery cost, handling cost, opportunity cost, penalty cost, indirect cost and idling cost and reduce delays in construction activities by using a Genetic Algorithm to help find the best answer for the time and quantity to order construction materials. This method has the advantage of helping to find the closest answer to the best answer can find the trend of improved answers and help create a plan for ordering steel bars in many forms the above methods have been used to experiment with case study projects, both of which consider the fastest time and the case that takes into account the cheapest cost. It was found that the modified economic order quantity model method compared to EOQ with Planned Shortage and EOQ with quantity discounts found that it can help reduce the time and cost to be less.

Foot Traffic Pattern Analysis of Vadhana District in the Spreading of COVID-19 Periods

2023-06-28T17:13:18+07:00

Between 2020 and 2022, several businesses in Thailand had deficits and wound up due to the government declarations for controlling the epidemic situation with large number of COVID-19 casualties. In this way, it is essential to analyze the impact on such businesses and prepare for the outbreak that may occur in the future. This research aimed to study the effects of businesses in Vadhana district of Bangkok, Thailand, using foot traffic data obtained by anonymized cell phone GPS location. The goals of this research are to study and differentiate foot traffic patterns of business types by percent difference condition between average foot traffic ratio of the group and the place. Moreover, foot traffic can evaluate the impact of government declaration and business recovery during the sample periods using statistical methods, Kolmogorov-Smirnov and Wilcoxon Signed-Rank, which performed to analyze foot traffic amounts. The study indicates that foot traffic pattern of each business type related to the number of foot traffic average per day. The results of the number of sample places which passed condition/total sample places are bars and night club 130/204, medical center and hospital 143/204, office building 199/282, restaurant 304/480, and retail shop and department store 145/180. Specifically, the government declarations, emergency decree, curfew, and preventing foreign tourists from entering the country, caused foot traffic reduction at the lowest level of enforcement periods. The second measure of closing bars and entertainment venues resulted in foot traffic slowly decreased to low level for two months and peaked in some period of this reduction that went against declaration’s objectives. Finally, foot traffic statistical analyses in October 2022 showed that all business types had not yet recovered from pre-pandemic period.

Improvement of Harmonic Detection Using SDF for Shunt Active Power Filter in Single-phase Power Systems

2023-09-27T10:01:16+07:00

This paper presents the performance improvement of harmonic detection using the Synchronous Detection with Fourier analysis (SDF) method to calculate the reference current of a shunt active power filter (SAPF) for single-phase power systems. The Positive Sequence Voltage Detector (PSVD) is applied to the SDF method to improve the accuracy of the reference current calculation in the case of a distorted voltage source. Where the objective is to enhance the effectiveness of harmonic elimination for single-phase power systems. For the harmonic detection testing, the hardware-in-the-loop simulation technique of the Simulink/MATLAB program and the TMS320C2000TM Experimenter Kit DSP board are used to simulate the harmonic elimination system. The simulation is divided into two cases of voltage sources: a pure sinusoidal waveform and a distorted waveform caused by harmonics. The simulation results show that harmonic detection using SDF in cooperation with the PSVD method (SDF+PSVD) can reduce the percentage of the total harmonic distortion (%THD) value of the source current to 3.07% for the case of a pure sinusoidal voltage waveform and to 2.00% for the case of a distorted voltage waveform. These %THD values are lower when compared with the SDF method and synchronous detection (SD). From the result, it confirms that the SDF cooperated with the PSVD method can provide better performance for harmonic detection. Therefore, the shunt active power filter can effectively eliminate current harmonics in the single-phase power system. Moreover, %THD of the source current after compensation is also satisfied under the IEEE std. 519-2022.

Loss and Efficiency Improvement of EV Traction Drive System by Using SiC in Inverter

2023-10-09T01:15:06+07:00

This paper investigates the effectiveness of employing SiC MOSFETs compared to Si-IGBTs in the two-level inverter for the traction motor drives in electric vehicles. The study focuses on the differences in structure and material between SiC MOSFETs and Si-IGBTs by which the conduction and switching losses can be reduced by over 40% and 60%, respectively, when a SiC MOSFET inverter is employed. This advantage enables the SiC MOSFET inverter to drive the traction motor at a higher switching frequency, surpassing traditional Si-IGBT inverters' capabilities. To assess the performance of the systems, simulations are conducted using the PLECS simulation platform. The comparison includes evaluating the traction drive system's efficiency and losses with SiC MOSFET and Si-IGBT inverters on the IPMSM traction motor torque-speed curve; the result shows that the efficiency is improved by over 1% on the entire torque-speed curve. Moreover, the paper also explores the trade-off between the switching frequency and motor harmonic core loss of the motor. This evaluation offers a comprehensive understanding of the interplay between these factors and aids in optimizing the performance of the traction motor system.

Physical and Biological Properties of Silk Suture Soaked in Levofloxacin

2023-10-04T08:16:12+07:00

Silk suture was a type of suture material commonly used by dentists to close wounds in the oral cavity. It provides strong tensile strength and was consistent with wound healing by maintaining suture integrity to promote wound healing for at least 3-5 days. However, non-resorbable sutures made of silk were known to accumulate food debris and increase the risk of infection, particularly in patients with underlying conditions such as diabetes or immune deficiencies. There has been recent development in coating silk sutures with antimicrobial agents to reduce the risk of infection, but no studies had been conducted using silk sutures coated with such agents. This study aimed to evaluate the physical and biological properties of non-absorbable and multifilament braid silk sutures that have been coated with levofloxacin. The physical properties of strength, amount of pore with SEM and BET analysis, and the presence of levofloxacin solution were assessed, as were the biological properties of drug release profile. All groups were compared between dry and wet conditions of the coated silk sutures and non-coated silk sutures. The results showed that the levofloxacin-coated silk sutures had no significant difference tensile strength between non-coated silk sutures, and the release of levofloxacin was sustained over a period of 7 days. BET analysis showed that the wet group had less porosity and wider pore diameter than the other groups. Therefore, the use of levofloxacin-coated silk sutures was provided comparable suture strength to non-coated silk sutures while reducing the risk of infection.

Sentiment Classification Based on Term Weighting with Class-mutual Information

2023-08-04T22:26:24+07:00

Online platforms and information technology are developing quickly, which boosts the popularity of online e-commerce. Nowadays, posting content from client sentiments is vital for product makers to improve product quality as much as possible to exceed customers' expectations. Sentiment analysis is a process of natural language processing that finds out the sentiment and attitudes of users towards a product, whether positive or negative. Most sentiment and text classification research uses term weighting with inverse document frequency (idf). However, assigning term weights using the idf method alone may not be effective enough to classify sentiment because this weight does not consider vital information classification views. This paper presents a supervised term weighting using the class mutual information calculated with the term frequency and the inverse document frequency. Experimental results show that the proposed method performs more effectively than term distribution and the term weighting that uses only the inverse document frequency when considering by the performance indicator value: Accuracy, Precision, Recall and F1-Measure.

The Prediction Model of PM2.5 in Phayao Province

2023-04-25T22:49:08+07:00

The dust in the atmosphere (Particulate Matter with a diameter less than or equal to 2.5 micrometers; PM_2.5) is the most significant air pollution in the northern region of Thailand, directly and indirectly affecting the health of the people. This pollution in Phayao province has consistently ranked in the top 3 of Thailand’s northern regions. Forecasting the concentration of PM_2.5 in the atmosphere using a mathematical model is an alternative that can be employed as a guideline for planning solutions or preventing air pollution problems. Logistic regression analysis and data from the Thai Meteorological Department and Pollution Control Department were applied to this prediction model. The 5 years of data between 2015–2021 were utilized for this model and created using Python programming. The results demonstrate that the independent variables of this model are PM₁₀, NO₂, O₃ and CO. The sensitivity and specificity are 0.96 and 0.72, respectively. The area under the curve (AUC) is 0.98.