Project KISSS

In shipbuilding, reliable and effective welding processes are hugely important, which is why many shipyards rely on automated processes for this work. However, such processes reach their limits when it comes to producing individual components, taking manufacturing deviations into account, or functioning in environments that are difficult to access. The German research project KISSS is developing an innovative welding process to overcome these limitations and aims to apply it in these use cases, which are rapidly growing in number. To do so, four partners from industry and research are counting on artificial intelligence (AI) methods. Over the coming three years, they hope to considerably increase quality and productivity for the industry and thus secure their competitive edge on international markets.

What ist the approach within the project?

In many industrial sectors, welding is one of the basic processes for joining components together. The welding methods currently used in shipbuilding are particularly efficient when it comes to repeatedly prefabricating similar and flat components. But in the field of component manufacturing, more and more new, complex and even three-dimensional tasks are emerging – intensified by the effects of the coronavirus pandemic. Today’s welding robots are not designed for these. To avoid major time losses and a not insignificant increase in the cost of manufacturing a ship, the partners in the KISSS research project (the acronym stands for “AI-based manufacturing technology for autonomous laser welding of complex ship structures”) are working on a new and robust laser welding process. Their goal is for it to form the basis for significant productivity and quality improvements in component manufacturing for shipbuilding. The project will employ AI methods in different phases of the welding manufacturing process in order to make the high complexity of the application manageable.

AI is still mostly associated with – preferably human-like – robots. However, the rapid development of this field in recent years has resulted in a host of new areas of application, such as the condition monitoring of machines and systems or the optimization of quality control for almost any process. As the most cutting-edge form of data analysis, AI opens up unimagined possibilities with regard to the evaluation of the ever-increasing volume of data generated in today’s industrial processes. Where humans can no longer gain an overview of highly complex processes and the associated measurement and control variables, AI is the tool of choice.

By employing suitable AI methods, the project partners want to draw up approaches that can help improve the entire planning of a sequence of welds. The partners will also apply these methods to find the optimal design for each individual weld seam in a new type of automated laser welding process that uses targeted heat input.