Journal of Engineering and Digital Technology (JEDT)

An Application of Text Mining and Association Rule Mining to Job and Skill Recommendations for IT Jobs

2023-10-22T23:53:24+07:00

This research implemented a web application that employed text mining to extract skill requirements from online IT job announcements, and association rule mining to discover the co-occurrence of hard skills (technical abilities) and soft skills (personal competencies) specified by the jobs. The matching score of each job was calculated by comparing hard skills extracted from the job announcement with a user’s current hard skills. Jobs were recommended to the user based on their matching scores. In addition, the discovered association rules were used to recommend new skills as follows: (1) based on the user’s current hard skills as antecedents, new hard skills as consequences would be recommended; and (2) based on the user’s current hard skills or soft skills as antecedents, new soft skills as consequences would be recommended. Online training courses to obtain such new skills were also recommended. The application was evaluated by 40 users, and received high satisfaction scores on both job recommendation and skill recommendation.

Improvement of Inbound Logistics Process in Coconut Manufacturing Using FlexSim Simulation: A Case Study

2023-12-15T14:02:43+07:00

This research focuses on improving the inbound logistics process in coconut manufacturing by employing simulation techniques using the FlexSim software tool. Data of the processes were collected and validated to generate the simulation model. Multiple scenarios were proposed, analyzed, and assessed against the key performance metrics, including throughput, work in process (WIP), stay time, and resource utilization. All scenarios were generated depending on the number of work centers for each process. The findings indicated that Scenario 3, Scenario 4, and Scenario 7 offered higher average throughput with 384.73, 384.60, and 384.73 units, respectively, and Scenario 7 provided the most efficient average WIP with 53.01 units and average stay time with 1263.7 seconds. Resource utilization was moderate across all scenarios. It is recommended to adopt strategies similar to Scenario 7 for optimizing inbound logistic case study. Nevertheless, real-world validation is necessary to ensure the implementation. The study contributes valuable insights to decision-makers in the coconut manufacturers. It serves as a basis for future research in logistics optimization, utilizing simulation techniques to enhance manufacturing performance and competitiveness.

Increasing of Stretch Formability on Hole-flanging Parts of Aluminum Alloys by Smooth Sheared Surface

2023-10-24T21:39:56+07:00

Aluminum alloy is a lightweight material that fractures often occur during stretch forming, especially in the hole-flanging parts. However, the stretch formability of aluminum alloy increased by improving a smooth-sheared surface. The resistance of materials to edge fracture in complex shape forming is typically measured using the hole expansion ratio (HER). This paper presents the concave piercing punch design to increase smooth sheared surface which induces the increasing stretch formability of aluminum alloy. The effect of a smooth-sheared surface on stretch formability was examined via a hole expansion test (HET). The aluminum alloy grade AA1100, and AA5052 of 1.2 mm of thickness with three levels of clearance 5%t, 10%t, and 15%t were used in the experiments. The HER of aluminum alloy is influenced by a smooth sheared surface from conventional and concave piercing punch with varied clearances. In the experimental aspect, the clearance level and shape of the piercing punch, which are factors for the stretch formability of aluminum alloy, were observed by hole HER. The concave piercing punch design improves the hole expansion ratio more than a conventional piercing because a smooth sheared surface around the hole edge increases the hole expansion ratio.

Low Cost and Simple Soil Moisture Measurement Using Multi-Level Capacitive Technique

2023-11-03T11:09:36+07:00

Global warming and climate change can cause water scarcity and drought for agricultural areas. Automatic irrigation can be one of the possible solutions for optimum water usage but has to cooperate with soil moisture measurement. However, the devices for soil moisture measurement at present are relatively expansive and require high technical setup and test skills; especially, for multi-depth soil moisture measurement. This paper proposes an alternatively low cost, simple soil moisture profile measurement using the multi-level capacitive technique. The proposed measurement technique was developed and tested by observing the moisture and water absorption capacity of sand, loam and clay soils at a depth up to 30 cm from the ground surface. It is found that the proposed measuring prototype could clearly classify levels of water infiltration, distribution and storage for particular different levels of the soil samples (The uncertainty values: RMSE from soils sandy, loam and clay by less than 8.50, 10.72, and 16.19 VMC%). The results also showed feasibility of the technique that could be used to study behavior of plants and crops in order to achieve the optimum water and moisture supply profile for different types of their roots in particular different soil depths for the best growth rate or quality.

Method Enhancement of Quality Control in Brake Pads Manufacturing

2024-02-25T15:44:06+07:00

The disc brake production is coupled with a defect detection process to control quality. The current manufacturing quality is high. However, detection relies on visuals by the operator, posing challenges in terms of accuracy and time. In a high-quality production, using human labor for detection is time-consuming and labor-intensive. This research is aiming at exploring the possibility of applying object recognition using deep learning approach for the in-line defects detection on disc brake pads. Faster R-CNN, Scaled YOLOv4, and YOLOv5s were compared for the detection of two major commonly defective brake pads of brake pad model A. The main criteria for the detection are (a) the detection time must be shorter than residence time of brake pads on the detection station on the conveyor, at 498 milliseconds, and (b) the precision must be higher than 70%. The detection time and precision of YOLOv5s, Scaled YOLOv4, and Faster R-CNN are at 13.9 milliseconds and 83%, 20.0 milliseconds and 83%, and 20.2 milliseconds and 92%, respectively. The detection time of all algorithms investigated in this study is far shorter than the residence time at the checking station with the precision exceeding the criteria. The training time for Faster R-CNN, 220 minutes, is five times longer than that of YOLOv5s (49 min) and Scaled YOLOv4 (41 min). All three algorithms are capable of real-time detection and yield a consistent result on both splits and poorly consolidated friction material workpieces. Faster R-CNN is chosen because it has the highest precision.

Monitoring Blood Coagulation Behavior Using Quartz Crystal Microbalance with a Mason Equivalent Circuit Model

2024-02-26T19:28:01+07:00

Blood clotting ability is a vital function of life forms, aiding survival when the body sustains cuts or wounds. Patients lacking this ability, or those on blood-thinning medications, can suffer from excessive bleeding, posing challenges for medical treatment. This is particularly relevant for dentists, as excessive bleeding may occur during tooth extractions or other dental procedures. Hence, there is a need for real-time blood coagulation detection. We propose using the quartz crystal microbalance (QCM) as a sensor. Its ability to detect mass loading and viscosity makes it feasible for coagulation testing in dental clinics. However, conventional equations like Sauerbrey’s and Kanazawa’s do not account for the dissipation factor, which plays a crucial role in determining liquid viscosity. We suggest using the Mason equivalent circuit model for interpreting blood viscosity and its clotting on the QCM surface. Observations show that film thickness increases with viscosity. Different samples exhibited varying viscosity changes, while the increase in film thickness was relatively comparable. This demonstrates the potential of the Mason equivalent Circuit model to differentiate results and estimate the underlying physical properties of blood coagulation.

Optimized PIDA Controller Design with Bat-inspired Algorithm for Temperature Control of Electric Furnace System

2023-11-03T08:17:23+07:00

The electric furnace temperature control system is one of the real-world second-order systems plus time delay (SOSPD) widely used as the process control in industries. It is normally operated under the PID feedback control loop. The PIDA controller, however, performed better response than the PID controller for higher order plant. In this paper, an optimal PIDA controller design for the electric furnace temperature control system is presented. Regarding to modern optimization context, the Bat-inspired Algorithm (BA) which is one of the most efficient population-based metaheuristic optimization techniques is applied to search for the appropriate PIDA’s parameters. The proposed BA-based PIDA design framework is considered as the constrained optimization problem. System responses obtained by the PIDA controller designed by the BA will be compared with those obtained by the PIDA controller also designed by the BA. As results, it was found that the PIDA can provides the very satisfactory tracking and regulating responses of the electric furnace temperature control system superior to the PIDA, significantly.

Performance Analysis of a Low-cost Automatic Single-Dispenser Drinking Water Filling Machine Controlled by a Programmable Logic Controller

2023-12-20T08:29:28+07:00

This research presents the application of a Programmable logic controller to control the operation of an automatic drinking water filling machine. The PLC is used to control the operation of the DC motor and control the stepper motor of the automatic drinking water filling machine. The working principle is as follows: 1) When a water bottle is placed on the conveyor belt, a conveyor belt delivers the bottles to the water-filling point. 2) When the water is filled to the right volume, the water bottle will be moved from the water-filling point to the capping point using the stepper motor rotation. 3) The process of capping the bottle cap tightening by direct current electric motor and air cylinder with a sensor to order the air cylinder to push down to close the lid, and the DC electric motor rotates to close the lid to the screw, and 4) When the lid is closed, the stepping motor rotates to bring the closed water bottle to the leg conveyor. The operation of the automatic drinking water filling machine was tested using statistical analysis standards with a total of 30 tests consisting of 3 values: finding the average value of the bottled drinking water volume at 597.27 milliliters and the standard deviation is 2.87 and the root mean square error (RMSE) is 1.95, meaning the water filling machine has an error of ± 1.95. The test results to find the optimum production cycle time of 12 minutes and the water filling rate found that it could fill an average of 8 bottles of water per minute. The energy consumption was found to use a maximum power of 32.1 W with an energy of 0.6 Wh per bottle. The automatic drinking water filling machine that has been created has an Availability% is at 99%, a Performance% is at 100%, a quality rate is at 88.89% and an overall equipment effectiveness is at 88.00%. Compared to industrial machines, automatic drinking water filling machines are generally considered to have good overall performance at a lower cost than the market price.

Single Order Inventory Replenishment Policy for the System Having Multiple Substitutable Inventory Items

2024-02-02T01:11:09+07:00

The research discusses the single order inventory replenishment policy for a system consisting of three inventory items with substitutability. The leftovers of lower-cost inventory items can be either used to satisfy those demands occurred from the shortage of higher-cost inventory items with a specified ratio or sold at their salvage value. In order to determine the proper order quantity for each inventory item, the method based on the concept of mathematical solving on a Monte Carlo simulation model is proposed. This can be done by modeling the system on the Microsoft Excel platform and, with the objective function of maximizing the expected system profit, applying the solver function to determine a good solution. For the performance evaluation, 120 experimental problems are randomly generated considering two main factors, salvage value (2 levels) and substitution ratio (2 levels), with 30 replications for each treatment combination. The solution obtained from the proposed method is compared with the solution yielded from the traditional benefit analysis method, which is to independently determine the single order quantity of each inventory item without the consideration of substitution opportunity. According to the study result, in comparison to the benefit analysis method, the proposed method provides a larger expected profit for all 120 experimental problems. Moreover, the main influences of salvage value and substitution affect the percentage difference in profit between the proposed method and the benefit analysis approach. Here, the percentage difference in profit increases at the lower levels of both factors.

The Automatic System for Perpetual Futures Contracts on Cryptocurrency Trading Using an Algorithm of Multiple Indicators to Predict the Trading Positions

2023-10-15T12:21:47+07:00

At present, perpetual futures contracts on cryptocurrency trading still face the risk of ruin. Therefore, this research was conducted with the objectives of creating an algorithm of multiple indicators to predict the trading positions, developing the computer program for generating trading strategies, creating the automatic system for cryptocurrency perpetual futures contract trading, and measuring the financial performance of the automatic trading system. The financial performance measurement was divided into two parts: a backward test, which involved evaluating historical data over a 3-month period, and a forward test, which involved testing system using current real trading data during a 3-month period. The total testing duration was 6 months. The results revealed that both the backward and the forward tests achieved a 3-month profitability efficiency of 130% and 140%, respectively. The overall accuracy rates were 72.00% and 80.39%, respectively. It has been shown that investing with the automatic system for perpetual futures contracts on cryptocurrency trading using the algorithm of multiple indicators to predict the trading positions has the potential to generate greater returns than the general traders.