Energy Management Systems (EMS) use sensors and meters to capture energy consumption, occupancy, and temperature and provide operational and analytical insights to optimize energy consumption. For small offices and homes, managing data can be fairly simple, but with a wide network of buildings spread across geographies, with 1000s of meters capturing data every few minutes, it can quickly become very complex. Here are some of our learnings to help you navigate around potholes as you embark on your journey of implementing an Energy Management System.
Our experience of implementing Energy Management Systems for multiple enterprises shows that every such implementation would necessitate addressing various challenges related to ingestion, storage, and processing of data. Additional capabilities of real-time monitoring, visualization, analytics, and prediction make this exercise further demanding.
Using a proven, robust and scalable IoT platform goes a long way in addressing many of these challenges. Among other things, such a platform helps in:
- Easy scalability of storage and processing capacity with growing needs
- Connecting with and ingesting data from diverse data sources
- Extensibility and flexibility of handling new requirements with minimal code change
- Security of the solution
Like most real-life IoT implementations, an enterprise-scale EMS would need to address situations of missing and spiky data resulting from infrastructure failure, network delays, and other reasons. A robust and resilient EMS would need to have the necessary intelligence to distinguish between good data and bad, i.e. ‘Is it a bad data spike that needs to be ignored (or otherwise handled) OR a genuine, good data spike that occurred because someone just turned on a chiller?’ It is crucial for an EMS to model the data in a manner that facilitates such intelligence based on self-learning from past data, and its understanding of what constitutes normal behavior. Right data models coupled with advanced machine learning capabilities can enable an EMS to continuously improve, leading to better predictive capabilities and an increased accuracy of anomaly detection.
The most important aspect of an EMS is deriving actionable insights. To prevent drowning in the data deluge of numerous alerts, create a profile of the buildings based on the distribution of energy consumption across different equipment/appliances (such as air conditioning), reduce your scope, and target those areas which provide maximum savings benefit.
Another key element for effective comparison across a portfolio of buildings is normalization of their performance based on factors like local weather, working hours, building usage and other business specific drivers. Such benchmarking enables identification of real outliers among them and throws up various additional possibilities of energy savings.
Successful deployment of the system critically hinges on what the operators of the EMS feel about it. Real-time applications often require information regarding availability and quality of data, to help trace faults and raise alarm when required. Exhaustive reports that help view and monitor the flow of data across systems go a long way in ensuring the success of the EMS. Finally, involving the operators early in your journey of EMS development and deployment helps – the experience and practical insights that they contribute can save you a lot of trouble in the long run.
Energy savings is largely dependent on the behavioral patterns of occupants. It is essential to generate awareness among occupants through public dashboards and infographics based on the data analysis and provide feedback options. Gamification and crowd-sourcing ideas from occupants can also generate great savings opportunities.
These are just some of the lessons learned through our experiences of developing and deploying EMSs. Every journey is unique and will surely have its own set of successes and learnings.
Here is another blog you might be interested in reading on Energy Management based on IoT & predictive analytics