Journal of Engineering and Industrial Technology, Kalasin University

Improvement of a Preventive Maintenance Plan for Machinery in the Canned Beverage Production Line

Rungprai Sarakum — 2026-08-25

This research aims to enhance the reliability and maintenance efficiency of a Tray Packer machine in a canned beverage production line, which was identified as the bottleneck machine with the highest downtime rate through the Weight Score Method. Conventional maintenance approaches that rely solely on manufacturer manuals or operator experience are insufficient for accurately identifying true machine degradation behavior. This study therefore proposes a novel framework by applying reliability engineering principles through Weibull Distribution Analysis, yielding a shape parameter (β) of 2.79, to verify wear-out failure behavior using actual time-to-failure field data. The findings were then integrated with the Total Productive Maintenance (TPM) concept. Unlike conventional TPM applications that depend primarily on operator judgment, this study uniquely employs statistical evidence derived from Weibull analysis as the basis for defining inspection items and maintenance intervals, resulting in a data-driven Preventive Maintenance plan that reflects actual engineering operating conditions. A comparative performance evaluation before and after implementation revealed a statistically significant reduction in total Breakdown Time by 44.03% and Failure Frequency by 35.20%. Furthermore, the Mean Time to Repair (MTTR) decreased by 13.63%, while the Mean Time Between Failures (MTBF) increased by 55.70%, at a statistical significance level of 0.05 (p < 0.05). The findings demonstrate that a data-driven preventive maintenance plan developed through systematic analysis can effectively reduce production losses, enhance machine stability, and serve as a replicable model applicable to other machinery with similar operating characteristics in the production line.

Design and Evaluation of Machine Safety Control Systems Based on ISO 13849-1 Using 3D Model-Based Design : A Case Study of an Automated Sorting Conveyor System

Sompoch Veangkum — 2026-08-25

The objective of this research is to propose a design framework for safety control systems in automated machinery by integrating the international standards ISO 12100 for risk assessment and ISO 13849-1 for safety design. A three-dimensional (3D) model of the machinery was developed to guide the definition of safety functions from the initial design phase. An automated conveyor system equipped with a pneumatic sorting mechanism was selected as a case study, representing a type of machinery that involves close human-machine interaction. The study focuses on the design and selection of input, logic, and output devices in accordance with ISO 13849-1 principles. This process relies on the technical data of the selected components, including MTTFd, DCavg, and the evaluation of Common Cause Failures (CCF). The research results indicate that the Required Performance Level (PLr) was determined to be at level c, derived from the risk reduction measures of ISO 12100 specifically related to the Safety-Related Parts of Control Systems (SRP/CS). This study identifies two distinct safety functions: SF1 and SF2. The designed SRP/CS successfully achieved a Performance Level of level e, which exceeds the required threshold (PL: e > PLr: c). These findings demonstrate a systematic design approach that can serve as a comprehensive guideline for developing safety control systems in automated machinery, in strict compliance with international standard requirements.

Small Object Detection for Red Ant Egg Classification and Detection

Thummarat Boonrod — 2026-08-25

This research aims to apply small object detection techniques to classify and detect red ant eggs from images captured in artificial nests, with the objective of improving detection accuracy for estimating egg quantity. The study employs a deep learning model based on YOLO, specifically the YOLOv11n version, which is an effective approach for small object detection tasks. The research methodology is divided into four main stages: data collection, image preprocessing, model training and testing, and performance evaluation. The dataset consists of 47 images of red ant eggs, divided into 32 training images, 9 validation images, and 6 testing images, corresponding to a ratio of 70:20:10. Data augmentation techniques, including image scaling, rotation, brightness and color adjustment, were applied to increase data diversity and reduce overfitting. The experimental results found that the model can detect red ant eggs with a precision was 0.66 and a recall was 0.79. The mean Average Precision (mAP) at an Intersection over Union (IoU) threshold of 0.5 (mAP@0.5) was 0.76, while the mAP@0.5:0.95 was 0.49. The results also show that the model performs better in detecting larger objects compared to smaller ones. This research shows the potential of applying deep learning-based object detection techniques in smart agriculture systems to support the detect and estimation of red ant egg production.

Effects of Rotational Speed and Feed Rate on the Performance of a Dual-Blade Herbal Chopper for Cassumunar Ginger and Turmeric

Thitinun Pongnam — 2026-08-25

This study aimed to investigate the effects of cutting blade speed and feed rate on the performance of a dual blade herbal chopping machine optimized for Cassumunar ginger and Turmeric. The machine's performance was evaluated based on throughput capacity, chopping efficiency and chopping loss. The experiments were conducted across three levels of blade rotational speed as 536.25 715.00 and 893.75 rpm. For Cassumunar ginger, the feed rates were varied at three levels as 359.64 617.14 and 872.37 kg/h, while a Turmeric, feed rates were 471.00, 621.76, and 916.81 kg/h. The results showed that, the feed rate for Cassumunar ginger and Turmeric had no statistically significant effect on the chopping efficiency. Conversely, the blade speed was a critical factor that significantly affected both throughput capacity and chopping efficiency at a 95% confidence level, particularly during Cassumunar ginger processing, for turmeric, the blade speed exhibited a similar trend influencing operational performance, with its p-value approaching the threshold of statistical significance at the 95% confidence level. To optimize operational parameters, the recommended settings for during Cassumunar ginger and Turmeric processing are feed rates of 872.37 kg/h and 916.81 kg/h coupled with blade rotational speeds of 715.00 rpm and 536.25 rpm, respectively. These optimal configurations yielded average chopping efficiencies of 97% and 98%, with corresponding chopping losses minimized to 3% and 2%, respectively.

Dynamic Response Analysis of a Strain Gauge Load Cell under Impact Loading for Accuracy Improvement in an Automatic Bottle Sorting System

Sawanee Jansawang — 2026-08-25

This study aims to investigate the dynamic response behavior of a strain gauge load cell under impact conditions and to apply the findings to improve the accuracy of an automatic sorting system. The experimental setup employed a load cell integrated with an HX711 signal amplification module and an Arduino Uno microcontroller for real-time data acquisition. Experiments were conducted by dropping objects with masses of 30, 50, 120, and 250 grams under controlled position and height conditions. The results showed that the output signal exhibited damped oscillation behavior during the transient period, and the settling time (T_s) increased significantly with the mass of the object. The average settling times for 30, 50, 120, and 250 grams were 50.22, 65.14, 100.02, and 145.03 milliseconds, respectively. To prevent signal overlapping in continuous measurements, the minimum release time was defined as approximately 2T_s, which improved system stability and reduced measurement errors. Furthermore, a linear relationship between object mass and release time was observed, leading to the development of an empirical calibration equation T_release = 0.832M+86.46 with a coefficient of determination R² = 0.982, indicating a high level of model accuracy. The findings demonstrate that the dynamic response characteristics of the load cell can be effectively utilized to determine optimal timing in automatic sorting systems, thereby reducing noise effects and signal overlap while enhancing overall system accuracy and stability.

Design and Development of a Robotic Glove Prototype for Hand Rehabilitation Using a Remote Control System

Theerathawan Panklang — 2026-08-25

This research aimed to design, develop, and evaluate the engineering performance of a robotic glove prototype for hand rehabilitation. The prototype consisted of a robotic glove, a pneumatic actuation system integrated with servo motor-controlled pneumatic valves, an ESP8266 microcontroller, and a remote control system via the Blynk platform. The system was capable of controlling hand grasping and finger extension movements in both collective and independent finger control modes. The system performance was evaluated at the engineering prototype level by assessing flex sensor readings, control system response, and wireless communication stability. The results showed that the system was able to accurately detect and transmit finger movement data, with an average response time of approximately 59 milliseconds. The system also demonstrated continuous and stable operation. In addition, user satisfaction was evaluated with 30 participants selected through purposive sampling, representing general users rather than target patients. The evaluation results indicated that the overall user satisfaction was at a high level (Mean = 4.42), covering usability, accuracy, safety, and system stability. However, this study was limited to engineering prototype testing and did not include clinical effectiveness evaluation with actual patients. The findings suggest that the developed robotic rehabilitation glove has potential for application as a hand rehabilitation assistive device and could be further developed for future medical and rehabilitation technology applications.

Development of a Stroke Tele-Rehabilitation System Using Markerless Augmented Reality and a Hybrid AI-Based Clinical Decision Support Dashboard

Sarawin Rachanikorn — 2026-08-25

This study aimed to develop a stroke tele-rehabilitation prototype integrating markerless augmented reality with a hybrid AI-based clinical decision support dashboard, in order to reduce the limitations of wearable-device-dependent rehabilitation and to support continuous home-based monitoring. The system employed MediaPipe for real-time tracking of 21 hand landmarks from a standard webcam and transmitted the extracted features to a Hybrid AI architecture that combines a rule-based engine with a large language model to generate feedback and clinical summaries under the Serious Game Design Assessment framework and the concept of neuroplasticity. The methodology consisted of engineering evaluation, simulated-data testing using 100 clinical cases, and expert validation by nine specialists. Evaluation in a simulated environment showed a frame drop rate of 5.36%, a spatial error of 10.51 pixels, a frame rate of 49.6–58.4 FPS, and a latency of 45.2–68.7 ms. For AI performance, the dashboard achieved 92.00% accuracy and 89.36% recall on 100 simulated clinical cases, while expert evaluation rated the system architecture and functional appropriateness at the highest level (mean = 4.59–4.75). The key contribution of this work lies in the integration of Markerless AR, Hybrid AI, and a clinical dashboard into a single prototype. However, the findings are limited to prototype evaluation in a simulated environment and expert-based validation; therefore, further clinical studies are required before real-world deployment.