AutomatedBuildings.com Article - Easy Wins in Building Data Collection

Have you spent time recently begging a facility manager to send you that one CSV data dump you need to close out your project? Asked to trend a few more points and discovered that the system is already at capacity, and adding more points will cost tens of thousands of dollars? Discovered trend logs only hold a week’s worth of data when you need an entire season’s-worth of data? Only a few organizations systematically collect, store, and exploit the potential of trend data. There are lots of reasons for this -- primarily, limitations of legacy products, and a perception that the cost of collecting and storing the data is very high due to this sort of factors.

Several existing technology trends should combine to significantly reduce the cost and complexity of data collection. In particular, the wide adoption of BACnet and availability of gateways from all major vendors means that the multi-protocol integration challenges of yesteryear are, while not gone, less common than they used to be. Widely-available, inexpensive cellular data connectivity means that moving the data off site no longer requires dedicated phone lines, USB stick drops, or manual emails with CSV attachments on a monthly basis, except in the most remote sites. And storage costs continue to plummet to the point where the hassle of deciding which data to keep is more expensive than simply keeping all of it.

Despite these underlying trends, building professionals still face limited choices and high costs when designing a data collection infrastructure. Data collection is often packaged as part of a larger solution for campus energy management, fault detection, controls, or commissioning, and despite the technology advances, vendors are still tied to legacy pricing models, and are in the habit of charging high per-point prices that encourage only collecting a subset of the available data. These stovepipe solutions increase costs for the customer because they lock the customer into a single vendor who can then bundle more services than needed to increase the size of a sale.

Over the past several years, we have been working on breaking past these barriers, through designing, implementing, deploying, and operating tools based on the Simple Measurement and Actuation Profile (“sMAP”), first through research at U.C. Berkeley, and now at Building Robotics.

If the data are not collected, they will be lost, and so the first critical step is for the data to be extracted from the control systems and meters and archived for future use and analysis. Taking this step is not necessarily expensive -- a few hundred dollars of hardware and tens of dollars per month for maintenance are sufficient in many cases to collect data every five minutes from essentially all of the points in a large building. Critically, building owners or operators should own this infrastructure and data. Even without sophisticated analysis tools, this infrastructure practically always pays back essentially immediately -- it takes only one avoided truck roll or chiller placed back on a schedule to recoup the modest investment in data acquisition.

Once the data are collected, though, is when true utilization of it can begin. Because the data are exposed and made available, the value isn’t limited to a single application, and no application needs to foot the bill for installing and managing the data. Building data finds many different customers within the enterprise.

Furthermore, by owning the infrastructure and data itself instead of tying those to a particular application, the enterprise can even use different, competing solutions where they each make the most sense, and enable them to shop their projects to multiple vendors. Independently, vendors can avoid building the in-house technical expertise in data acquisition that is currently required, and focus on their true competitive strengths: the services and value delivered to the customer.

At Building Robotics, we believe that the data should be owned by the customer, and that the basic systems and protocols used to gather that data should be open, with open-source implementations existing alongside proprietary ones. That’s why we’re going to keep supporting the open source sMAP project, which allows anyone to use a high-performance time-series engine, as well as a library of drivers for collecting building data. sMAP is a powerful tool for openness in the building industry, since using it, anyone can building their own low-cost data acquisition infrastructure, and then ask analysts and vendors to play on top. We believe there are easy wins on the table for a large number of building owners who don’t currently systematically manage their building data, and that a modest investment will pay back immediately, with returns compounded down the road, as the market evolves to offer better products that are easier to install and use, and lower cost.

Steve Dawson-Haggerty, PhD, is the CTO at Building Robotics and has extensive experience in building wide-scale data collection infrastructures. Steve leads the team's efforts in new approaches to optimizing building controls through elegant software, machine learning and carefully tailored occupant-facing controls. In 2015, Building Robotics released its latest product, Trendr which leverages the sMAP open source project to change how people access building data, making it quick, easy and intuitive.

Events	Want Ads
Our Sponsors	Resources