A welding process control system using artificial intelligence is being developed by Dmitry Bublik, a postgraduate student at INRTU

Dmitry Bublik is a first-year postgraduate student in the field of informatics and computer science, and he teaches students about neural networks and machine-oriented programming languages. Andrei Trufanov of the Centre for Interdisciplinary Research and Projects is the polytechnic student’s supervisor.

The talented student, together with Roman Kononenko, head of INRTU’s computer hardware and software laboratory, has been developing robotic complexes for five years.

«Artificial Intelligence technology is a hardware and software system comprising a video camera, a microphone, a signal matching device for the robotic welding machine and a computer. The welding process can be monitored in real time and optimum welding parameters can be set to achieve a perfect welded joint thanks to these devices. The hardware and software system allows welding in all spatial positions — vertical, overhead, and underneath," said Dmitry Bublik.

The introduction of such systems at enterprises will increase economic efficiency, production safety, increase the speed of welding and improve the quality of welds, he said.

«One of the main advantages of our machine is its versatility. We are not tied to any manufacturer, you can install the system on any machine," the graduate student said of the features of the development.

Dmitry Bublik also cited monitoring data fr om domestic enterprises, including Gazprom, Transneft and Transtroy, wh ere welding is actively used. There are about 50 thousand people involved in welding production. More than 500 employees work in robotic welding, which is a low number.

During experiments the polytechnic student has already received input data, identified regularities of welding parameters influencing formation of joints, among them speed of wire feeding, movement, movement of the torch. The graduate student is now working on the correlation between the parameters and their influence on the weld.

Dmitry Bublik plans to create a full-fledged autopilot for welding robotic systems that can be put into mass production.