SAU JOURNAL OF SCIENCE & TECHNOLOGY

Hydrodynamic Performance Analysis of Hull Propellers for Enhancing Flow Efficiency in Water Propulsion Systems

Fri, 26 Jun 2026 00:00:00 +0700

This research presents a case study on the design of a Surface Piercing Propeller (SPP) with an optimized geometry for water propulsion systems. The study focuses on a 3-blade propeller featuring a 12-inch diameter and a 22-degree rake angle, specifically engineered for integration with a 16-horsepower engine. Performance evaluations were conducted in an open-water flume to simulate real-world operational conditions. The experimental results demonstrated that the newly designed SPP achieved a maximum propulsion speed of 6 knots (11.112 km/h) when positioned at the water surface. In contrast, at a submersion depth of 30 cm and an approximate shaft angle of 36 degrees, the propeller maintained an average speed of 5.6 knots (10.371 km/h). Furthermore, the optimized propeller design contributed to a 16% reduction in fuel consumption compared to conventional standard propellers available in the market. These findings indicate that the enhancement of the propeller’s rake angle plays a critical role in improving thrust efficiency and energy conservation for flood mitigation missions.

A Conceptual Framework for Applying Artificial Intelligence in Shoplifting Prevention within Department Stores

Tippapa Meesin, Chutima Pisarn — Fri, 26 Jun 2026 00:00:00 +0700

This study aims to examine the application of Artificial Intelligence (AI) technology for preventing merchandise theft in department stores. The research focuses on exploring AI technologies that can be applied to theft prevention, developing a conceptual framework for applying such technologies, and evaluating the suitability of the proposed framework through expert assessment. This research adopts a qualitative research approach using documentary research, emphasizing the analysis of relevant documents to develop a conceptual framework for this study in preventing theft in department stores. The proposed conceptual framework consists of theft risk management, the identification of risk-prone areas based on retail business formats to determine appropriate technological applications, budget and consideration of legal frameworks related to theft prevention technologies in department stores. The results from expert evaluation indicate that, overall, the proposed framework is appropriate for preventing merchandise theft in retail businesses. It can enhance the effectiveness of security systems and aligns with modern retail risk management practices, particularly through the integration of AI technologies with (Closed Circuit Television: CCTV) systems and behavioral analytics. Organizations can practically apply this conceptual framework for shoplifting prevention under a clear risk management approach by starting with high-risk areas and selecting technologies that align with the business model, budget, and existing organizational systems. In addition, legal requirements and cost-effectiveness should be carefully considered before implementation.

Design and Development of a Miniature Automated Storage and Retrieval System Using a Gantry Mechanism for Automation Education

Shanaphat Dechnaraphat — Fri, 26 Jun 2026 00:00:00 +0700

This research presents the design and development of a gantry-based Mini Automated Storage and Retrieval System (Mini-AS/RS) as a prototype for automation systems. The system employs an X–Y–Z motion structure controlled by a PLC via push buttons and a touch screen interface to enhance positioning accuracy and reduce reliance on manual labor in storage operations. Quantitative performance evaluation shows that the system achieves an average operating time of 23.1–30.2 seconds per cycle, with a standard deviation of less than 0.23 seconds. The average positional error is 0.2 mm, and the positioning success rate is 100%, indicating high accuracy and repeatability. Furthermore, a pre-test and post-test assessment conducted with 40 students revealed a 30.22% increase in the average score, with statistically significant improvement. These findings demonstrate that the developed system not only achieves reliable engineering performance but also effectively enhances learning outcomes.

Surrogate Model Optimization Using Gaussian Process Regression and Particle Swarm Optimization for Engineering Design Problems

Fri, 26 Jun 2026 00:00:00 +0700

This research proposes surrogate model optimization (SMO) for engineering design by the use of machine learning regression algorithms and an optimization algorithm. The SMO proposed in this research features an optimization process designed to use only a small number of generated solutions. Additionally, variable scaling is applied prior to constructing the surrogate model to ensure that each variable has a similar impact on finding the optimal solution. The machine learning regression algorithms to be evaluated are gaussian process regression (GPR), support vector regression (SVR), and artificial neural network (ANN), while the optimization algorithm used is particle swarm optimization (PSO). After testing with benchmark function optimization problems, GPR with PSO has the highest performance and outperforms the design of experiment (DOE) technique, Taguchi method. For three standard benchmark functions, the proposed SMO using GPR and PSO yielded average errors of 0.0001, 0.0000, and 0.0856, whereas the Taguchi method resulted in average errors of 0.0000, 0.1249, and 1.0000, respectively. Thereafter, SMO using GPR and PSO can effectively solve engineering design benchmark problems and engineering design problems with the numerical calculation by Solidworks software. It produced design objective values with deviations of only 0.63% and 0.58% from the exact solutions for two standard engineering design problems where the exact solutions are known. Therefore, SMO using GPR combined with PSO is considerably suitable for surrogate model optimization in engineering design problems.

Multi-Objective Optimization of a Multi-Link Slider–Crank Mechanism for Vertical Motion Generation in Legged Robots

Fri, 26 Jun 2026 00:00:00 +0700

This study presents the design of a multi-link slider–crank mechanism for generating vertical motion in legged robots. The design problem is formulated as a problem of multi-objective optimization (MOO) with two primary objectives: (1) maximizing the vertical motion range to cover a specified target interval, and (2) minimizing horizontal deviation of the center of mass during motion. To prevent bias toward oversized mechanisms, the total link length is incorporated into the formulation. The MOO employs the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to obtain Pareto optimal solutions. NSGA-II yields 508 Pareto-optimal solutions, notably, 40.35% of the solutions achieve horizontal deviation below 1 mm. Compared to single-objective optimization (SOO), a specific solution obtained from the MOO solution set achieved a reduction in total link length of over 30% while maintaining high vertical motion efficiency and low horizontal deviation. Consequently, the solutions obtained from MOO offer more compact mechanisms than those from SOO. These findings highlight the effectiveness of the MOO framework in providing design flexibility and practical suitability for implementation in mobile robotic systems.