Data-Driven Health Status Monitoring on Electrical Pumps

This study presents a data-driven approach for detecting failures and anomalies in electrical pumps as well as identifying pumps’ health state, particularly in industrial paint shops. By analysing pump pressure, electrical current, and operational frequency, we employed algorithms to detect failures and anomalies on pumps and to monitor the health status of pumps. This approach surpassed traditional diagnostic methods by enabling proactive maintenance, thus enhancing operational efficiency, and reducing down-times. Results indicate that the algorithm to monitor pumps’ health successfully identified potential issues before they developed into critical failures, and model to detect sudden failures on pumps successfully identified failure in its early stages. The precision and reliability of algorithms was further validated through comparison with historical maintenance data, demonstrating a significant improvement in the early detection of pump malfunctions and overall state of electrical pumps systems. This research showcases the potential of data-driven strategies in revolutionizing industrial maintenance, underscoring the benefits of predictive analytics in achieving more reliable, efficient, and cost-effective operational practices.