Peter Sondergaard, VP Gartner Research
By using our client’s quality management expertise and data as well as craftworks’ predictive analytics know-how, we can predict and prevent quality issues in production processes by identifying the root causes. The goal thereby is not only to improve the quality of the products, but also to decrease potential scrap and rework costs as well as warranty claims and to decrease quality related machine downtimes.
The production process of brick production companies starts with the preparation steps such as grinding and milling before the actual bricks are being shaped and dried.
Together with an interdisciplinary team from both the client's side and craftworks, the concrete questions that should be answered were evaluated. Using the gathered data from different production steps, we developed a machine learning algorithm detecting quality problems and its root causes. The results are visualized in order for the client’s operators to easily read and understand the system and its recommendations. After a successful PoC, the algorithm can be transformed to other production machines step by step as well. Moreover, instead of acting upon recommendations of the system, the settings of parameters can be automated allowing a highly efficient process.