Energy Management Platform for Centralised Chiller Plant using AI




The centralised chiller plant of a building accounts for roughly half of all energy consumed by a building. Therefore, operating the chiller plant in the most energy-efficient manner help to optimise building energy utilisation considerably.  This centralised chiller plant energy management platform identify energy-saving opportunities and addresses maintenance issues, improves equipment lifecycles maintenance and productivity, and improves energy-savings. with the proposed technology relates to advanced machine-learning algorithms that identify inefficiencies in the chiller plant and recommend better operating set-points resulting in 5% to 40% reduction in energy consumption. The platform further detects anomalies, provides actionable root-cause diagnosis, and condition-based monitoring of chiller plant equipment, including chillers, cooling towers, and pumps.  This energy management platform is designed to provide energy-saving optimisation of a chiller plant, simplify its maintenance and maintain the energy-savings achieved. Equipment being optimised in a centralised chiller plant includes the chillers, cooling tower, and pumps.  The technology provider is seeking partners to collaborate through various modes including joint-research, sub-licensing, Platform-as-a-Service and/or other commercial arrangements.


A set of AI algorithm identifies optimal operating conditions of the chiller plant taking into consideration the dynamic building cooling load, utilisation cycle, external weather, plant equipment condition and long-term trends of equipment efficiency. The energy optimization algorithm further takes into account degradation in equipment efficiency, and improvement of equipment efficiency due to recent maintenance to optimise operating set point.  Machine-learning algorithms further detect degradation in energy efficiency with: 1. A machine-learning algorithm that detects anomalies in plant efficiency. The use of sensor fusion techniques and multivariate analysis helps to identify deviation in chiller plant efficiency. Identified anomalies are presented to the domain expert for further analysis and labelling. 2. Beside anomaly detection, alarm detection with false alarm suppression capability help to monitor the permissible operating range of critical parameters. The alarms come with advanced false-alarm suppression techniques allowing the user to provide feedback and moderate permissible conditions according to plant condition. This reduces false alarm rates and this reduces alarm fatigue. 3. With user labelling, identified anomaly and alarms are then translated into Fault Detection and Diagnosis (FDD) algorithms to detect future occurrences of energy inefficiency incidents. This includes instances where chiller programming resulted in non-optimal part-load utilisation, operating of unnecessary pumps or unnecessarily high flow resulting in inefficient operation, and many other incidents resulting in loss of energy efficiency.  4. An equipment lifecycle algorithm scores the efficiency of the equipment by monitoring a combination of degradation in efficiency and occurrence of incidence where equipment operates in high wear conditions.


The energy management platform is designed for facility managers seeking to improve chiller plant performance, reducing energy and improving productivity in the maintenance of the plant.  The energy management platform can also be integrated into a more comprehensive smart building technology suite helping buildings manager manage energy utilisation of the plant. More algorithm from OEMs are welcome and can be added to the platform improving energy efficiency and simplifying maintenance, improving productivity.


The energy management platform is designed for facility managers seeking to improve chiller plant performance, reducing energy and improving productivity in the maintenance of the plant.  The set of algorithm identifies energy-saving opportunities in a centralised chiller plant, detects and identifies deviation of energy efficiency achieved, reducing the energy consumption of a building. This helps to reduce cost and meets heighten energy efficiency targets and sustainability goals of smart buildings. More advanced algorithms from equipment OEM can also be integrated onto the platform, providing between equipment insights.