Case Study: Uddeholm reduces steel cracks and waste with CGI machine learning

Uddeholm predicts steel cracks with CGI at over 70% accuracy

Uddeholm, a Swedish multinational producer of high alloyed tool steel, faced a significant challenge with cracks forming in its finished products, leading to costly waste and energy usage as damaged steel had to be re-melted. They were unable to effectively use their abundant data to identify the causes and improve their process. To address this, CGI developed a machine learning-based solution to help predict and reduce these production errors.

CGI implemented a high-powered machine learning model that analyzed Uddeholm's data using big data and an IoT platform. The solution could predict cracks with over 70% accuracy, enabling Uddeholm to pinpoint the causes of quality issues and adjust its manufacturing process. This resulted in significantly less waste and improved profitability, representing an important first step in the company's digitization journey. CGI also helped implement satellite technology for tracking materials.

Uddeholm