Data Wrangling for Breakthrough Energy Performance Analytics can easily shave 20% off your energy costs

Jack Mc Gowan, CEM Principal The Mc Gowan Group Contributing Editor published May BUILDINGS

Buildings are complex organisms with systems and occupant patterns that are in continuous flux. Thanks to ENERGY STAR and other sources, benchmarking data exists to compare the energy performance of similar buildings. However, such data shows wide and unexplained differences in performance among similar buildings. What are the roots of superior performance? And for buildings already performing at high levels, is it possible to break through to higher performance?

For answers, we must drill deeper than benchmarking data allows and tap the power of big data energy analytics. Even skilled managers confront a bewildering host of known and unknown factors that affect performance in every building. Analytics technology can unlock dramatic savings, easily 20% of energy costs, while ensuring ideal occupant environments.

Making Analytics a Part of the Routine

Building managers tend to focus on maintenance and system upgrades. Analytics – or “data wrangling” as I like to call it – elevates such routine activity to continuous building improvement by adapting to equipment operations and occupancy patterns. Through combinations of commissioning, preventive maintenance and equipment upgrades, it can facilitate performance breakthroughs that exceed building benchmarks.

Because the diversity of systems in buildings adds complexity, the first step is understanding applications that should be optimized, such as HVAC systems and the facility’s Energy Utilization Index (EUI). Once applications have been identified, engineers and other “data wranglers” are deployed to create rules or parameters that define ideal operation for these applications.

With rules defined, analysts can identify what data is necessary to test whether a building is operating within parameters. That data is then evaluated against the rules to determine performance. Analytics software engines crunch the data to identify patterns, issues, faults and trends. In this sense, the big data approach is different than, say, buying a new high-efficiency LED lamp, which is a product-based solution.

Rules in analytics are best illustrated by referencing a single piece of equipment. For example, a rule might define ideal operating parameters for a particular type of VAV box, including a set of values for each operating condition (discharge and return air temperature, zone temperature compared to setpoint, static pressure, etc.). Analytics tools access the data from the equipment and compare it to the rule to determine if a unit is operating properly.

Getting started with analytics is easy. If you have access to operating data (energy consumption, equipment trends, etc.), you can start the journey to data-driven performance. Because you can’t manage what you can’t measure, a single building that is utility metered or a metered process is a good starting point. If your building has a building automation system (BAS), it should be a valuable resource of data.  Typically there is some level of integration needed to get data from the BAS to analytics tools, but once data is flowing, these tools will begin producing ideas for improvement. As financial results accrue, go further. 

The Potential of “Self-Describing” Data

Facilitating the flow of data from the BAS to the analytics application has great potential for improving performance results. The data should follow standards that make it able to define itself or “self-describe.” Project Haystack (project-haystack.org) and ASHRAE (BACnet.org) have assembled industry experts to work on consensus-approved approaches to describing data so it can easily be consumed by cloud-based analytics tools. Algorithms evaluate whether conditions meet the rules or require adjustment.

This software analysis converts volumes of unconnected data into knowledge. Analytics tools don’t just trigger alarms, they provide actionable knowledge in a format that can be quickly understood by operations personnel. Managers know whether buildings are performing to expectations, and if not, why not. Tools provide lists of recommendations summarizing what’s wrong and what’s necessary to meet or exceed performance expectations. They tell us how long the problem has existed, what it will cost to fix it and what skills are needed. This information is conveyed through executive dashboards that provide the knowledge necessary to truly break through previous barriers to high performance buildings.

About the Author

John J. “Jack” McGowan is Principal with The McGowan Group and former CEO of Energy Control Inc. (ECI), an OpTerra Energy company. He is Chairman Emeritus of the U.S. Department of Energy GridWise Architecture Council. In 2003 the Association of Energy Engineers (AEE) admitted him to the International Energy Managers Hall of Fame. His book, Energy and Analytics: Big Data and Building Technology Integration was recently published by Fairmont Press (www.fairmontpress.com) and he instructs an AEE seminar on analytics (aeeprograms.com/Realtime/BigData/).