BuildingIQ Epiphany Reveals Interdependent Data Relationships for AI-Driven System Optimization

Powerful inference engine creates virtualized network of disparate data points within systems

SAN MATEO, Calif. – August 6, 2019 – BuildingIQ (ASX: BIQ) has accelerated its leadership in artificial intelligence (AI) solutions with Epiphany – an AI-driven engine that uncovers how data is interconnected, dependent and influenced within a system. Epiphany is used to create foundational knowledge about a system that can enable insights like root cause analysis of anomalies; predictive failure and maintenance of equipment; better capital planning; and prioritization of service based on the real, hidden impact of issues versus simplistic cost-to-replace estimates.

Inferences derived from Epiphany are today powering BuildingIQ’s Outcome-based Fault Detection (OFD) service, bringing deeper anomaly detection and more valuable insights to fault detection within a building. With the introduction of the Epiphany full suite of tools, BuildingIQ can create a virtualized network of a system that brings an even more holistic view of system interactions among myriad data points and assets, regardless of how they are wired or logically connected.

The four components of Epiphany as initially applied to building systems are:
·   Influence Engine -  Reveals the impact of all points relative to whole building power, inferring what is driving a building’s energy consumption.

·   Influence Delta - Examines how an influence changes over time, looking at associated power impact from equipment due to that change. 

·   Delta T -  Correlates real energy impacts with comfort impacts, allowing facilities teams to prioritize work better than with simple rule violation faults.

·   Correlation Map - Uses machine learning to create a map of the relationships between points, quickly inferring root cause through powerful filtering capabilities.

For more information about Epiphany, watch the video at: https://youtu.be/ktXans2qHj4

[Home Page] [The Automator] [About] [ Subscribe ] [Contact Us]