Engineering Transactions: A Research Publication of Mahanakorn University of Technology

Time Series Forecasting using Diffusion Model in Case Water Quality of Chao Phraya River

2025-08-04T11:11:32+07:00

Time series forecasting is a fundamental task for predicting future values in order to understand underlying behavior. Numerous forecasting models rely on learning patterns from historical observations. However, these models face limitations when time series data exhibit inherent uncertainty, multi-modality (i.e., multiple plausible outcomes), abrupt regime changes, or complex noise. Recently, an alternative solution has emerged in which models aim to approximate the distribution of possible future outcomes. One such solution is the use of Denoising Diffusion Probabilistic Models (Diffusion Models), which offer a promising framework by learning a generative process that models complex data distributions in a probabilistic and iterative manner. This capability enables the model to effectively capture uncertainty and multi-modality in time series forecasting. In this study, we implement a Diffusion Model for time series forecasting. We conduct an evaluation on a real-world case study involving water quality in the Chao Phraya River. The results demonstrate that the Diffusion Model outperforms classical techniques such as Long Short-Term Memory (LSTM) in time series forecasting.

Design of equipment to reduce defect in anchor plate production

2025-09-05T18:00:48+07:00

This research aimed to reduce defect in the manufacturing process: a case study of the production of anchor plate parts. The study of the anchor plate manufacturing process revealed a defect in the second step, drilling small holes, which accounted for 83.49%. Waste characteristics were then analyzed using the Pareto principle. The first defect was misaligned holes, was found to be 91.84% and the second was misaligned plates, was 5.81%. Remediation efforts were undertaken to address both problems. ECRS principles were used to identify solutions and design equipment using tool engineering principles to assist with hole positioning. After the design of the positioning device, defect was reduced from 6.22% to 0.29%.

Rice husk ash-derived catalyst for diethyl ether production through ethanol dehydration in a large-scale reactor

2025-09-10T15:02:59+07:00

Ethanol dehydration to diethyl ether (DEE) was carried out in large-scale reactors using rice husk ash (RHA), a low-cost biomass catalyst derived from agricultural waste. This paper investigates the impact of temperature and liquid hourly space velocity (LHSV) to determine the optimal conditions and stability for ethanol dehydration to DEE over an RHA catalyst in a large-scale reactor (2.1 cm i.d., 30 cm length). The RHA catalyst was synthesized using a reflux process with sulfuric acid, followed by calcination at 600 °C, and characterized through X-ray Diffraction (XRD), Fourier-Transform Infrared Spectroscopy (FT-IR), Field Emission Scanning Electron Microscopy (FE-SEM), X-ray Fluorescence (XRF), Surface area and pore size analysis, NH₃ temperature-programmed desorption, and thermogravimetric analysis (TGA). The optimal conditions were determined to be 360 °C and an LHSV of 1.3 h⁻¹, which resulted in the highest DEE yield production of 99.21% and ethanol conversion exceeding 99%. The results of the stability tests, conducted over a 24-hour period, demonstrated that RHA is resilient to deactivation and suitable for long-term catalytic applications. These findings indicate the potential of RHA as a promising and sustainable catalyst for the large-scale production of DEE.

Complex Numbers and Functions and Harmonic Analysis

2025-07-04T11:56:27+07:00

The purposes of the present academic paper are to review and summarize the basic of complex numbers and functions including the harmonic analysis as necessary that involved to analyze the vibration of mechanical and structural systems together with demonstrating the system modelling.

Buckling Criteria of an Euler Column with Both Ends Elastically Restrained: Analytical Solution and Verification

2025-07-04T11:55:18+07:00

This study is concerned with the buckling problem of an Euler column with the condition of both ends elastically restrained under an axial compressive force. The objectives are, therefore, to derive a general buckling criteria equation based on an analytical method and to present verification for the obtained results. The highlight of the study is that the problem governing differential equation which is formulated and solved in the form of non-dimension. Significantly, the considering column can be re-conditioned to other cases of column having various end restraints.

Design and Implementation of an Unmanned Fire Detection and Extinguishing System for Greenhouse

2025-09-12T20:04:30+07:00

Greenhouse fires pose severe risks due to enclosed structures, dense equipment, and rapid flame propagation. To address the limitations of manual inspection and fixed monitoring devices, this paper presents the design and implementation of an unmanned fire detection and extinguishing system for greenhouses, based on an STM32 microcontroller with Raspberry Pi co-processing. The system integrates flame sensors, temperature–humidity sensors, LiDAR, and a high-definition camera, mounted on a wheeled platform with a rotatable extinguishing module. It supports autonomous patrol, manual control, and point-to-point extinguishing modes. Wireless video and data transmission enable real-time fire monitoring and remote command. Experimental results demonstrate flame recognition accuracy exceeding 95% within 10 m, positioning error below ±0.5 m, and fire response time under 8 seconds. Compared with conventional methods, the proposed system achieves faster response, broader coverage, and stronger adaptability, showing great potential for intelligent greenhouse fire prevention.

Workpiece Inspection and Sorting System using Image Processing on CiRA CORE Integrated with PLC and IoT

2025-11-05T15:56:15+07:00

This paper presents a workpiece inspection and sorting system using image processing on the CiRA CORE platform integrated with a PLC and IoT technology to enhance accuracy and reduce human errors in workpiece size inspection. The system is designed to use 2 cameras to capture images of the workpieces from the front and top views, with the acquired data transmitted to the CiRA CORE platform for classifying the workpieces according to their diameters and heights. The inspection results are then transmitted to the PLC via the Modbus protocol for automatic sorting control. In addition, the system is connected to the Node-RED program to display and record real-time operational data through IoT technology. The experimental results were performed using cylindrical workpieces with diameters of 5 and 6 centimeters and heights of 4 and 5 centimeters, with each size inspected 20 times. The experimental results showed that the system could accurately detect and measure the dimensions of the workpieces, achieving an average accuracy rate of more than 98 percent. Furthermore, the overall machine performance was evaluated through the Node-RED program.

A Back-to-Back Matching Network Implementation Technique for Accurate Characterization of Source/Load Impedance Seen by Power Transistor in RF Power Amplifier Design

2025-11-18T15:44:55+07:00

This paper presents a simple and more accurate measurement of matching network characteristics for an RF power amplifier, specifically its driving point impedances as seen by the power transistor, which are critical to its operational performances. The proposed technique relies on an implementation using the same matching network connected in a back-to-back configuration to avoid direct measurement on the matching network port that is connected to the power transistor, which has a relatively large pad size so as to connect with the gate and drain terminals of the power transistor package. This serves to mitigate an interconnect discontinuity between the matching layout and the 50-ohm transmission line layout to a network analyzer. The measured responses based on the back-to-back implementation is then employed to optimize the component values of the matching network schematic so as to fit the measurement with simulation, thereby enabling the implementation of the schematic to more accurately determine the driving point impedance. Validation of the technique is given via practical design and implementation of a 3.1-GHz GAN power amplifier.

The Development of an Online Blood Donation Queue Management System with Advance Notifications to Enhance Operational Efficiency

2025-12-03T18:26:08+07:00

This study aims to develop and evaluate the performance of online blood-donation queue reservation system equipped with advance appointment notifications, designed to address difficulties in booking, communication constraints, and inadequate donor preparedness factors that critically affect donation regularity and the adequacy of blood reserves in healthcare facilities. The proposed system was designed through a systematic user requirements analysis combined with structural design techniques, including the Context Diagram, Data Flow Diagram, Data Dictionary, and Entity Relationship Diagram, to ensure architectural and functional coherence. The system evaluation indicates that all modules operate correctly according to the specified requirements, with data management conforming to relational database constraints. The booking, notification, user-management, and information presentation functions demonstrate stability and effective responsiveness for both administrators and general users. The developed system therefore exhibits strong potential for real-world implementation to support blood supply management, reduce operational errors, and enhance donor readiness at the operational level.