Assessment of Critical Functions in Battery Production Lines for Predictive Maintenance

Sammanfattning

In an increasingly automated and data-driven manufacturing environment, the utilization of real-time data for predictive maintenance plays a crucial role in minimizing downtime and improving reliability. This thesis focuses on identifying and assessing critical functions in the battery module production lines at Scania’s battery plant in Södertälje, Sweden. The goal is to support smart maintenance by evaluating where the pain points exist in the battery production lines and ensuring that these areas are effectively monitored. A condition-based risk ranking model was developed using three key parameters: failure visibility, frequency, and downtime. Data for this analysis was extracted from Scania’s existing process monitoring system (PROMO) and the Computerized Maintenance Management System (CMMS) - IBM Maximo. To structure and understand the production process flow as functions, the Function Analysis System Technique (FAST) was used at both the machine and overview levels. This helped create a logical mapping of how, why, and when each function occurs. Based on the functional risk assessment, this work also evaluates the sufficiency of the current sensor and vision system coverage. For functions with inadequate monitoring, sensor or vision system upgrades are proposed, supported by both functional risk justification and historical qualitative work orders from Maximo. These findings will be used to support dysfunctional analysis for future research. Additionally, a conceptual framework for integrating Machine Learning (ML) is proposed to support maintenance teams by providing data-driven recommendations for incoming work orders. The outcome of this thesis provides a pathway toward smarter, more proactive maintenance planning that is grounded in operational data and human reasoning, a practical step toward more resilient and efficient EV battery production lines.