Automatic Diagnostics is part of Aquicore’s new Load Analytics, a machine-learning capability designed to detect usage anomalies from your building’s electricity curve.
What is it?
We know your day is busy, and it can be difficult to dedicate time to reviewing Load Analytics and creating relevant Alerts for your building. Automatic Diagnostics is Aquicore’s way of providing you with actionable insights into possible problems.
Automatic Diagnostics ultimately save you and your team(s) valuable time by reviewing electricity curves and identifying (possible) issues that would likely have gone unnoticed. Most importantly, Automatic Diagnostics highlight the estimated Potential Savings of reducing the identified consumption.
How does it work?
Compared to other Artificial Intelligence solutions, Aquicore's Automatic Diagnostics...
is based on a robust machine learning model
relies on a feedback loop and improves with engagement
is building specific
ingests billing data, building schedules, and weather to inform its analysis
is built from a "crowdsourced" knowledge base of Energy Notes captured within the platform
We use all available data for your building(s) as part of our algorithm, so it is important to ensure your building schedule(s) and utility bills are recorded and kept up to date within the Aquicore platform.
Automatic Diagnostics handles robust energy analysis for you behind the scenes to identify possible issues and their associated cost, enabling your teams to do their best work.
Who is it for?
Automatic Diagnostics deliver insights that benefit Building Engineers by identifying possible issues that could impact building operations. Property Managers benefit from Automatic Diagnostics by having visibility into the associated financial impact.
Where do I find Automatic Diagnostics?
Automatic Diagnostics will appear in your building's Activity Feed Widget on your Custom Dashboard.
Within the Activity Feed Widget, Automatic Diagnostics will appear under the Load Analytics section and are organized by event type (e.g. Holiday, Weekend, or Nighttime Run).
Clicking an individual Automatic Diagnostic opens a pop-up that provides more details into the anomaly. Users are also asked to confirm the cause of the anomaly by clicking "Yes" or "No." Clicking "No" prompts you to reclassify the anomaly.
Classifying the anomaly is an imperative part of our feedback loop, and effectively trains our machine learning model. In other words, the more you engage with Automatic Diagnostics, the smarter they become.
Reclassifying Anomalies
In the event, that an Automatic Diagnostic is not accurate, and the user reclassifies the anomaly…
The potential cause listed in the Automatic Diagnostic modal will be updated
That day’s data is excluded from later baseline calculations
The algorithm will exclude similar events moving forward
The new anomaly cause is fed back to our algorithm’s training model to improve future detection
When do I see Auto Diagnostics?
Once a day, our platform runs an analysis of the last 36 hours of building data to detect any diagnostics. For buildings using interval utility data, we run the analysis as we receive new building data.