Predictive maintenance through data-driven decision making

Abstract

This thesis investigates the deployment of Predictive Maintenance (PdM) systems and solutions at Parker Hannifin Manufacturing, a producer of control valves. PdM uses advanced analytics, machine learning algorithms, and real-time sensor data to predict maintenance needs before machine breakdowns occur. This approach contrasts with traditional preventive maintenance (PM), which follows predetermined schedules. The research aims to identify critical parameters for equipment monitoring, the necessary data infrastructure, and the potential cost benefits of PdM implementation. The study involves a case study of spool cell machines at Parker Hannifin Manufacturing, addressing challenges in data collection, algorithm development, and cultural shifts within the organization. The expected outcome is strategic roadmaps for transitioning from PM to PdM. The study concludes that applying PdM can reduce maintenance costs and enhance operational efficiency