Opportunities and challenges of applying artificial intelligence to improve production processes in the context of welding manufacturing

Abstract

As manufacturing is increasingly transforming into Industry 4.0, artificial intelligence is playing an important role in the development of production processes in manufacturing. Despite growing research on AI applications in manufacturing, there have been limitations in specific industrial contexts, such as welding production. This thesis investigates the opportunities and challenges of applying AI to improve production processes in welding manufacturing, focusing on three key areas: production scheduling and workload balancing, predictive maintenance, and quality control and inspection. The research uses a qualitative case study methodology, combining theoretical background with empirical insights from semi-structured interviews conducted with four personnel from a Finnish welding manufacturing company. The theoretical background examines AI technologies, including machine learning and deep learning, and their applications in welding contexts. The findings reveal that while AI has lots of potential for enhancing efficiency, product quality, and cost reduction, there are still challenges to overcome in its adoption. Key challenges include data quality issues, integration complexity, implementation costs, and organizational readiness. Interviewees identified an AI-based production scheduling system as the most promising area currently, while quality control and predictive maintenance were seen as concepts with potential in the future. The study concludes with practical implications for the case company and for welding manufacturers considering AI integration.