The building industry is on the cusp of a major paradigm shift to connected, smart buildings using Internet of Things (IoT) and artificial intelligence (AI). Every facilities and energy manager’s inbox is full of blogs, newsletters, and products proclaiming this impending revolution. Yet, the asterisk accompanying each proclamation lists a caveat: “only those with good data may apply.” Time and again this vision crumples and fails when faced with the reality of the limited and poor-quality data in existing buildings. This dearth of data is sometimes a technical issue, and other times a budget issue. Therefore, the question every facilities executive faces is how to best use limited funds to work toward a smart building vision, and what to do in the meantime?
To address this issue, we re-thought the traditional approach to building optimization and created a flexible, outcome-focused data platform that can not only accommodate a wider range of data quality and availability, but also help building executives understand the most critical area to invest. Instead of using inflexible rules-based approaches that must use a specific data point to function, we use flexible machine learning data analytics that look for patterns in the data. Instead of attempting to create a rule for every possible fault, our tools look at the influence of each point and piece of equipment on whole building power and identify when they are suddenly using more power than anticipated. This tool can work with 100 or 10,000 data points and identify energy impacts for whatever data is available. In fact, this approach is so data agnostic, it can be used on any time series data, from industrial process to substation data.
But AI and data analytics are only part of the optimization equation. Technology-enabled services include a human layer —round-the-clock monitoring for all systems and faults so our building engineers can triage every anomaly. These are people that understand the dynamics and history of each building and can fill in the data gaps with contextual knowledge. This approach focuses on changes that result in measured, not calculated, increased energy consumption, removing the noise of every day faults that do not truly impact building performance.
The combined use of AI and building experts means that we do not necessarily need all the data to reach actionable recommendations. Often, the action of a missing upstream point can be inferred by watching the response of downstream systems. Facilities managers intuitively fill in these gaps everyday by watching a room’s temperature response to identify a VAV valve failure, or an AHU discharge air temperature to estimate a reheat coil’s performance. These intuitive jumps require a deep knowledge of system interactions, often across multiple subsystems. To date, this has been hard to teach a computer, but with a new analytics tool, we can now track and quantify the impact of each point on every other point. The tool creates a map of all upstream and downstream points to uncover the systems driving or being influenced by each point.
This serves as both a powerful tool for tracking the propagation of a fault through a system during root cause analysis, and a means of identifying downstream system responses to critical equipment despite data gaps. These analytics build on this intuitive process of inferring cause and effect by fully mapping and quantifying these interactions. A team can then identify faults or tune system response using the behavior of the downstream point’s response as a proxy for the missing upstream data.
Often times, a single room or zone can drive an entire system’s performance due to uneven loading or equipment issues. In parallel with these relationships, teams can use the measured impact on whole building power for these critical zones to verify priorities and estimate potential savings. Exposing these interdependences and energy impacts can allow FM teams to prioritize retrofits or IoT sensor installations in these critical areas first. Our tools help teams identify which systems, zones, or data gaps have the largest influence on whole building energy consumption, have the highest risk for failures, and have the greatest impact if left untended.
Helping understand the upstream and downstream interactions of systems means facilities management teams work smarter, more efficiently, and at lower cost; yet with greater positive impact —a key part of building optimization. The added data flexibility provided by data agnostic statistical approaches means that facilities teams and building owners have a tool set that adapts to their current building data capabilities, and critically, lets them focus their upgrade budgets where they can achieve the most impact.
Chris McClurg is Sr. Product Manager, Software & Services at BuildingIQ, and a mechanical engineer focused on energy efficiency in large portfolios and net zero developments. Chris has worked on deep retrofits, integrated design, integrated project delivery, and buildings as a grid asset. She is a PE, CEM and LEED AP certified.