Smart Grid

August 15, 2013

NREL's Open-Source 'DataBus' Tracks Power Usage at Any Size Facility

All aboard! The U.S. National Renewable Energy Laboratory (NREL) has launched an open-source software solution called the Energy DataBus. It will enable facility managers to collect, integrate and analyze data on power usage on a large scale from a wide variety of sources.

NREL currently is using the Energy DataBus to track and study energy usage on its own campus in Golden, Colorado. The system is applicable and available to other facilities of practically any scale, from a single building to a large military base or college campus or for other energy data management needs.

"It's similar to the one that Facebook and Twitter (News - Alert) use to collect all this data and information on people. We're using the same giant, scalable capability to collect energy data," explained Aaron Beach, Energy Informatics (News - Alert) researcher.

Managing and minimizing energy consumption on a large campus is usually a difficult task for facility managers.There may be hundreds of energy meters spread across a campus, and the meter data are often recorded by hand. Even when data are captured electronically, there may be measurement issues or time periods that may not coincide. Making sense of this limited and often confusing data can be a challenge that makes the assessment of building performance a struggle for many facility managers.

Four Trillion (News - Alert) Data Points Annually

The Energy DataBus software was developed by NREL to address these issues on its own campus, but with an eye toward offering its software solutions to other facilities. Key features include the software's ability to store large amounts of data collected at high frequencies. NREL collects some of its energy data every second. As a result, nearly half a million data points stream into the database over the course of an hour. That adds up to more than 4 trillion data points per year —and rich functionality is required to integrate this wide variety of information into a single database.

"The data gathered by NREL comes in different formats, at different rates, from a wide variety of sensors, meters, and control networks," says Keith Searight, development manager of the Energy DataBus. "The Energy DataBus software collects all this data and aligns it within one scalable database."

Existing energy data systems tend to be designed for particular applications such as reporting or control and do not provide general data-analysis interfaces flexible enough for NREL’s requirements. But rather than creating a new solution from the ground up, NREL built the Energy DataBus on existing open-source software products that are used to manage complex data problems in other industries. The Energy DataBus supports the popular Cassandra database and PlayORM, which facilitates the easy integration of widely varying data. PlayORM provides the Energy DataBus with the capabilities it needs to interact with many types of data, including time-series, textual, or numerical data.

NREL has developed tools that allow the Energy DataBus to collect data:

  • From control systems using the BACNet protocol,
  • From electrical meters via the ModBus protocol, and
  • From weather sensors via Web requests to an on-site weather station.

High-Quality Data

The flexibility of the Energy DataBus Web API has made all of these systems easy to integrate. Additionally, the Energy DataBus software incorporates tools that help assure the highest quality of data in the database.

"No matter what you do, when you start collecting data off meters, you get bad data, you get data dropouts, you get all sorts of issues," Searight explained. "So we created some tools to clean and pull out relevant information from the datasets—including such things as filling in missing data and identifying bad data in the system. Being able to mine the data and perform some analytics on it helped us pinpoint where potential issues were with some of the control strategies in our buildings."

The Energy DataBus can operate at widely varying scales and can be integrated into small desktop applications and run on a laptop or virtual machine with only a few gigabytes of RAM (News - Alert).

On the other hand, Energy DataBus was designed to support cloud architecture and can be scaled to hundreds—or possibly thousands—of nodes across the globe. For instance, NREL currently runs its production version of Energy DataBus with 12 database nodes on high-performance servers and four webserver nodes on virtual machines. However, for development, the team uses a version that runs in memory on a laptop.

Looking to the Future

"The plan is to roll this platform out to the public in open source format so that university campuses and other institutions can tap into these tools and build off of them," stated Beach.

To employ the Energy DataBus, other facilities would need to connect their existing data collection systems with it and then configure it to meet their particular needs. But NREL has found that it's worth the effort. Already, a team at NREL has developed an application that uses the Energy DataBus to provide data to a set of "energy dashboards" —enabling anyone on site to monitor the energy performance of the NREL campus. Ergo, the task of minimizing energy use falls on each person, instead of just the facility managers.

Another team at NREL has developed an app that allows building occupants to report their comfort levels. More apps are in the works.

For software developers, the Energy DataBus software is now available for free download. For more information on how to obtain, test, and implement the Energy DataBus software, contact Keith Searight and the Energy DataBus team at


Edited by Rich Steeves