ario@eng.ui.ac.id
1998 – Mechanical Engineering, Universitas Indonesia
2004 – Mechanical Engineering, Universitas Indonesia
2006 – Mechanical Engineering, Keio University, Japan
2009 – Mechanical Engineering, Keio University, Japan
We want to contribute to the manufacturing world with automatic welding technology that the industrial world and for current human civilization benefits for.
Welding is a manufacturing process carried out to make a metal joint. Welding results must have the same quality control and standards even though they are carried out by different operators. Armed with these conditions, we made innovations to facilitate welding activity
https://www.youtube.com/watch?v=MffNGmD9D24
Technological and industrial advancements demand excellent products and quality support, especially in the rapidly-growing manufacturing sector. One of which is achieved from welding results that meet the required standards, are consistent, and uniform in order to be able to function properly. However, some obstacles generally arising during welding activities include a difficult welding result achievement. Accordingly, any welding innovations designed through automatic movements are required to facilitate the operator‘s work and to ensure that the results are consistent and meet an established industrial quality standard
In this 4.0 industrial era, the production process in the industrial world has taken advantage of artificial intelligent technology establishing a piece of equipment work automatically that the industry needs since time efficiency is always calculated to achieve massive and quality production targets. Welding works require accuracy and inspection processes to keep results accurate and precise. Therefore, to support automatic inspection processes, we develop a simulation of welding and visual inspection systems to obtain an image that is automatically processed with a machine vision.
In this study, we also developed a Neural Networking system to automatically regulate and control the welding speed, which could be achieved by carrying out the training process of Artificial Neural Networks using previous welding data meant to estimate and control the welding penetration. Welding penetration is estimated using a hybrid estimation method by combining welding simulation and sensor-based visual observation using a Charge-Coupled Device (CCD) camera. The experimental results revealed that the control system was effective enough to detect a Molten Pool in real-time and produce suitable welding.