Machine-learning (Sentinel)

Energy data arrives like many pieces of a large puzzle – time-series data, meter readings, weather changes, and fluctuating use patterns all play a role. Machine learning is a type of artificial intelligence that evaluates data to find patterns, then uses those patterns to predict future behaviors. Sentinel, EnergyCAP’s smart software engine, predicts what your energy use should be at any given time, for every circuit communicating data into the platform. 

Sentinel is constantly looking at the past and present of your energy data, in order to predict the future. Sentinel then continuously compares your real energy use against the predicted amount to discover unusual energy performance. When a piece of equipment fails to turn on when it should, or runs long after it shouldn’t, Sentinel alerts notify you of unusual usage, bubbling issues in your building portfolio to the top of your list, saving you time and money, and acting like an on-call equipment specialist 24/7. 

Sentinel provides you with a deep understanding of where your energy profile is, and where it should be, so you can focus on energy saving projects and corrective actions. You can do things like:

• Address problems in real time. Set up Sentinel alerts to inform you via email, text, or on your SmartAnalytics dashboard when abnormal demand is detected.
• Call for action. Enlist the help of your staff to switch off equipment when you’re alerted to unnecessary use over breaks or offtime. 
• Save for the future. Sentinel powers the Operations Analyzer module, which lets you simulate energy saving actions and provides specific predictions for cost, energy, and GHG savings.
• Make informed decisions. Use Sentinel predictions to choose the energy tariff structure that will save you the most money.

Sentinel is constantly learning. The adaptive engine uses ever-amassing historical data to make better predictions about future energy consumption patterns. From these patterns, the system is able to give you a range of the expected load (shown in green). The black line represents the actual load. 

In any of these graphs, you can hover over a particular point to see the real values of consumption. You can switch between kilowatt hours and kilowatts to see trends for active power or active energy of a particular monitored point or circuit. Sentinel alerts can be automatically created, or set up by you, to notify you when actual load falls outside of the expected range. This means your team can save cost and usage by jumping into action to address potential problems before they get worse, or simply by managing equipment, HVAC, lighting, and other circuitry settings. 

Sentinel works in tandem with other capabilities, such as data normalization, to evaluate energy usage regardless of outside factors like the weather. Sentinel knows that your air conditioning units should be running in the summer, but will flag the issue if they turn on in December. Not only do you have the actionable data you need to keep building conditions ideal year round, but letting Sentinel monitor and alert you to unexpected changes in your energy profile means you have the time and resources to focus on energy efficiency projects and actions.