ML Based Data Driven Energy Centered Predictive Maintenance

Energy consumption is a critical issue, with $500 billion spent annually in US alone on energy. Energy-centered maintenance (ECM) is a continuous improvement maintenance regime designed to reduce energy use by identifying equipment or items that consume excessive energy while still performing their function. The primary purpose of ECM is to prevent equipment from becoming energy hogs, which supports the energy reduction program adopted within the facility and helps achieve the projected energy savings. Machine Learning based Data-driven energy-centered maintenance (DDECM) is the key focus of this paper, which involves using modern AI technologies to automate the majority of the task, especially in identifying maintenance requirements of a chiller system. This approach uses software and hardware technologies for real-time monitoring of the equipment's performance by collecting energy consumption data, identifying anomalies, and classifying them into typical failure conditions by comparing them to the historical patterns that define a baseline of its ideal performance. Artificial intelligence and machine learning take the maintenance intelligence to the next level by providing early alarms related to abnormal equipment performance that were previously not possible to trigger. Implementing this model supports organizations in their digital transformation strategy by providing a business case for implementing cost-efficient maintenance tasks based on real-time data and digital analytics.