Background
In a densely populated metro like the City of Washington DC, heating and cooling buildings consumes the most energy and makes the largest contribution to greenhouse gas (GHG ) emissions. DC’s built environment contributes 75% of all its GHG emissions, which is almost double the national average of 38%. In 2013, the District’s Department of General Services launched a new energy-efficiency retrofit program for owned properties and engaged an energy consultancy to lead and coordinate the efforts. By summer that year, the BuildSmart DC platform was up, with a page dedicated to almost every District property. This website discloses metered energy data to building operators and the public from local Potomac area utility, Pepco.

While energy data is valuable, it is 100 times more valuable when you add building operational data. The District’s building energy managers wanted to take their efforts to this next level by bringing the data streaming from Building Automation Systems (BAS) to an open source platform where it can be used to enable ongoing commissioning with automated Fault Detection and Diagnostics software and possibly other dashboard visualization programs in the future. For this expertise, they hired Intellimation LLC.

Pulls and Tugs on BAS Data
The concept of a major metropolitan district extracting and using BAS data to drive energy efficiency decisions at portfolio-wide scale was pioneering in 2013, but indicative of the industry as a whole. The BAS industry has entered a time of rapid change, and business-as-usual is being challenged and disrupted from multiple directions:

• Widespread adoption of BACnet/IP is providing a building-level standard protocol for connectivity. The proprietary ‘walled gardens’ protecting legacy building equipment vendors are crumbling as more building owners insist on open BAS architectures and BACnet compatibility.
  
• Widely-available, inexpensive cellular data connectivity means that moving data has become easier and less expensive. Building owners understand that there is need to get data into the right hands for decision-making. More and more often, this often means an energy management expert monitoring and controlling from afar.
• Falling price of data storage and new technologies for aggregating data is undermining the common practice of pricing the trending of BAS points at a premium. In the past, automation contractors worked with customers to keep monitored points below a threshold determined by the capacity of BAS internal storage and processing resources. And they programmed the system to delete trend logs after a week or so to conserve storage. It was typical for the estimated cost of adding even a few more points to the trend logs of some BAS-managed piece of equipment to be in the thousands of dollars. None of this makes sense today. After years of fierce competition, the IT industry has brought down the cost of storage to a small fraction of what it was a decade ago and has introduced ‘big data’ methods of managing vast data stores for fast and easy access.
  
• Finally, Automated Fault Detection & Diagnostics (AFDD) for commercial buildings is emerging as an attractive market for software developers with expertise in rule-based analytics programs. New entrants and familiar BAS-industry open-source software innovators are evolving this category at a fast pace. However, legacy BAS systems are not designed to trend most their data points, so their integrated analytics and visualization programs don’t often have all the data they need to deliver insightful results. This introduces need for an open architecture platform for trending and archiving BAS data that is extendable to integrating data sources external to the BAS, such as data from mobile messaging and other apps, wireless thermostats, occupancy sensors, etc.

Criteria for a Trending and Archiving Platform
When Intellimation arrived on the scene in the summer of 2013, we found that among the approximate 100 buildings in the District’s portfolio that had BAS systems, several different branded control systems were in place and not all of these were BACnet/IP enabled. Certain BAS systems were simply run on local PCs with no outside connectivity. Our first challenge was to get all the buildings with BAS systems connected and to create an architecture that was conducive to utilizing the growing number of AFDD building analytics solutions.

We piloted various tool sets and approaches as we started to get the feeds as trend log objects.  At first, we couldn’t succeed with fewer than 5 to 15 clicks per trend point. This was unacceptable because, with thousands of points to setup, the time investment was too great and would run up costs.  At this point, it occurred to us that a better solution would be a universal trending and archiving platform that would bypass the trending tools built-into the various BAS systems altogether. So we began a search for such a solution with the following selection criteria in mind:

• Auto discovery of points from any BACnet/IP controller
• Acquire data without setting up trend logs
• Data historian that could be located within customers firewall
• The platform had to be fast, easy and vendor-neutral 
  
• Have a powerful API, making it easy to onboard software vendors
  
• Data needed to remain customer-owned

There is an array of AFDD software solutions on the market, and new entrants keep arriving. The right solution for a given building can depend upon its particular age, purpose (healthcare, education, civic business, etc.) size and mix of building equipment. DC’s internal energy managers wanted a trending and archiving platform that did not narrow their future AFDD options. They wanted to ensure the path would remain open to selecting the best toolset— based on analytics rule libraries and price point— for any given building, years into the future. They didn’t want the trending and archiving platform that they standardized on today to lead to paying high switching costs for BAS equipment, middleware or AFDD software later.  Finally, the District of DC wanted to own all data acquisition and archiving infrastructure. They did not want to end up in a situation whereby the data they needed to address some new line of questioning existed only on someone else’s remote data stores.

Engine for High-Performance, Time-Series Data Collection Emerges
After some research, Intellimation learned of an innovative trending and archiving device built upon an open-source platform called sMAP, for Simple Measurement and Actuation Profile.  Born out of DOE-funded research at the University of California, Berkeley, sMAP was developed by the two founding members of Oakland, California, -based Building Robotics. The Building Robotics sMAP product is called Trendr™ and also includes a library of drivers for collecting building data. By the time Intellimation heard about the sMAP platform, the standard was already gaining support in other university research settings and by a loyal community of private software developers interested in the building industry, making it a truly open platform. So we joined their beta program. Building Robotics offers the option of either hosting Trendr in the cloud or, as was required by the City of DC, installing it on a server located within the perimeter of their firewall. Intellimation contracted with Building Robotics for a trial device and began using Trendr for the City of DC’s retro-commissioning  project.

[an error occurred while processing this directive] Using Trendr gets us past the first major issue with the trending and archiving tools built into legacy BAS boxes—their capacity constraints. We can trend most, if not all, available BAS data points—hundreds of thousands of points across the portfolio, without worrying about capacity or the affordability of storage. These comprehensive ‘data lakes’ are what is required to effectively deploy AFDD analytics. Next, Trendr got us past the tedious, cumbersome process of trending points manually.  Without Trendr, it would take us 6 to 12 mouse clicks to create a trend log in a typical BAS. That was if we could get access to the BAS data at all. There are incumbent BAS vendors that block access to feeds via special keys and proprietary formatting tricks that only their partnering automation contractors know how to disentangle. Now we can place a Trendr anywhere on the BAS network and, in one click, trend all points in the system. The system auto-polls the BACnet points without having to configure the local BMS, saving us from having to know all the addresses.  So there is no need to involve the BAS vendor at all.

With Trendr, we can be pushing data within four to eight hours of walking on site when the site is native BACnet. What can take hours or even days to create in a BAS — 1 to 2 thousand trend logs — takes 30 minutes with Trendr.  Also, with Trendr all the data stays within the firewall.  And, Trendr imposes no constraints concerning either BAS or AFDD vendor selections. It sits in the middle and is neutral. The data originates in the BAS system and flows through sMAP and then on to an AFDD vendor via a standard API. Any BAS or FDD vendor could be replaced without replacing Trendr.

Implications for the Future of BAS Data
To date, Intellimation has deployed 20 Trendr boxes in DC buildings and we’re planning to get to 50 in six months. The underlying sMAP platform is a powerful tool for openness in the building industry. It’s an ideal platform for building a low-cost data acquisition infrastructure, particularly for the large number of building owners who don’t currently systematically manage their building data. Analysts and vendors can build upon the open sMAP APIs, so the growth path for this approach is secure. 

[an error occurred while processing this directive]
[Click Banner To Learn More]

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