Robotisation of welding process using information from 3D model

Abstract

The fourth industrial revolution, characterised by increased automation and robotics, is reshaping industries worldwide. In response to the growing demand for cost-efficient and rapidly evolving construction methods, with environmental considerations, construction firms are increasingly motivated to refine their operational processes. This drive is aimed at enhancing efficiency and achieving significant time and cost savings. Aibel, in particular, is actively pursuing these advancements to maintain its competitive edge in the market. Therefore, this thesis aims to develop an automated proof-of-concept process for welding specific structures—pipe support structures, that are currently welded manually. To achieve this, establishing an information flow between software tools is imperative. More specifically from a 3D model to simulation software, effectively integrate this into a simulation of an automated welding process, that would be a proof-of-concept simulation demonstrating the possibilities.

First, the capacity of the 3D model to store all essential information must be evaluated. In this project, the 3D model is managed through Tekla Structures software and its Tekla Open API functionalities. Tekla has proven capable of containing all necessary data, including material, weight, and size attributes of the profiles. This data is seamlessly collected using the Tekla Open API, which has been integrated with a custom-built application designed for information retrieval, storage, and conversion to a neutral file format. This neutral format allows the data to be further utilised in KUKA.Sim, ensuring the implementation of accurate welding procedures. The combination of these tools has established a proof-of-concept simulation for an automated welding process, enabling effective collaboration between robots. The simulation tool KUKA.Sim, known for its complexity, has facilitated a deep learning session to refine the simulation creation.

By effectively finding a common ground, such as a neutral file format that both Tekla and KUKA.Sim can use, the 3D model information can be efficiently utilised to select welding procedures in KUKA.Sim. The simulation has been successfully implemented using a robotic system comprising two robots. The first robot is responsible for picking and manoeuvring support structures while the second, an arc robot, carries out the welding operations. This strategy has effectively demonstrated the feasibility of integrating 3D model information into a simulation process.