A few weeks back, I was chatting with a CEO in the building technology industry whose company is developing a digital twin for buildings. I asked a simple question and assumed I’d get a simple answer.
Can your technology can serve as a platform for data analytics?
Spoiler alert: I was wrong.
Fault Detection and Diagnostics (FDD) is a specialized field with plenty of actors while not many active users, so we’re waiting to see when it becomes a normal practice. We have enough new frontiers to break in addition to FDD.
His answer made me take a step back and ask myself how well this industry actually understands an essential tool.
Many experts have already responded to this on my LinkedIn post. Below I’ve added my thoughts and sprinkled in my favorite comments from the experts.
I have three main issues with the CEO’s answer:
- Analytics does not equal FDD. FDD is an analytics capability. One among many.
- Some level of analytics is a no-brainer for all building owners. Questioning the need for analytics is like saying:
- “my building is running perfectly”, or
- “my building is not running perfectly, and I don’t want to know about it or how to fix it”, or
- “my building is not running perfectly, but I’d like to analyze it by walking around or crunching some numbers in excel or glancing at the BAS every so often”
- Every new technology starts out as “specialized” with “not many active users”… They’re called early adopters.
Let’s focus on the second and third bullets. Analytics are inevitable because every building owner needs them and the obstacles preventing scalability will soon be overcome.
Every building owner needs analytics
The need for analytics in buildings is well-illustrated in Derek Cowburn‘s story about his parents’ net-zero energy home:
My parents’ home is Net Zero. Their monthly power bill is $10 (grid connection minimum charge). Last winter they had a $400 bill in one month. Turns out a board cracked in their groundwater heat pump furnace and it wasn’t running the water pumps so the emergency resistance heaters were on. Their honeywell wall thermostats did show EMHEAT in a little LCD icon but the system kept temperature pretty well. Not as good as when the mechanical heating was functioning properly.
My parents are in their 80s but they both have smartphones and email (even wechat). Point is, there should have been a big buzzer or light or some internal diagnostic indicator on the $20,000 waterfurnace that said “hey, I think I’m broken!”. That month and a half undetected problem cost more than the 6 years prior total net operating cost.
In summary, we can design perfect buildings, but the ongoing performance is always going to degrade. It’s a proven fact.
With ever more stringent energy performance laws—like New York City’s Local Law 97, which will spread across the country and perhaps the world—and better building performance demanded by occupants, building owners need improvement, not degradation.
Building analytics software provides an unparalleled suite of tools required for improvement. The situational awareness gained by bringing diverse sets of data together into one single platform provides insights and prioritizes the best actions to optimize performance.
Facility managers are feeling the pain of the well-documented and growing shortfall in building operating labor, knowledge, and talent. Analytics not only helps fill that gap but can also be used as a recruiting tool to attract new talent who want to work with innovative technology.
Think about every bit of your own years of learned experience and thought process that drove your troubleshooting to eventually find the root cause. Imagine a platform that had access to every variable that went into that process, that could not only pinpoint the issue the absolute second it happened and alert you (text, email, Slack, etc), but also predict its inevitable next failure by regressing against the same system of variables to thousands of other homes. That’s where machine learning is taking us.
While many commercial building owners have invested a lot of money in building automation systems, they’re built on technology that’s 15 (or more?) years behind the smartphones in our pockets. General technology has left building technology in the dust. Analytics provide a control system overlay that’s cheaper than replacing old technology from the ground up.
The nature of control system efficiency is best described by the Grumman X-29 jet. Crazy high efficiency thanks to the forward swept wing design, BUT, if the control system fails it falls from the sky.
And even if a building owner wanted to replace everything with the latest tech, there’s no iPhone-equivalent for buildings. Analytic platforms are the closest thing—the best bet for making up for the failure that is modern controls.
With supplemental supervisory controls, analytics platforms augment and increase the capabilities of underlying control systems, enabling automated optimization and smarter responses to ever-changing conditions.
Analytics are (almost) ready to scale
The best analytics platforms are not specialized in any way—they’re built to scale across all building types and applications.
FDD software doesn’t need to be terribly specialized, only the application of it (service)… provided the software has a flexible, configurable framework/interface. I think about this as a data continuum that has these basic pieces:
[data acquisition > FDD software > application]
I think that the two ends of this continuum are where most of the specialization come into play. Data acquisition (in the broad sense that encompasses all the variations previously discussed in a Nexus newsletter post) has a huge amount of specialization in technology and methodology.
Application has a huge amount of specialization based on what you want to monitor and alert on, and for what purpose(s) you want to monitor/alert. In this continuum context, I believe the FDD software can be (and should be) the least specialized piece.– Joel Urban
And some of those vendors are seeing the technology beginning to scale up.
I agree that there are many frontiers still to break through (in our industry), many of which are challenged primarily by scaling technologies that we know can have significant impacts. I do see Fault Detection and Diagnostics as a very foundational technology which is now scaling fast with an organizational user base that is increasing rapidly.
– Nick Gayewski, CEO, KGS Buildings
Meaning this is a problem of overcoming final obstacles to scale, not a question of whether it will scale.
I don’t see the field as that different to other innovative and disruptive technologies – many actors, ways of approaching the problem, business models etc. which will eventually shake out.
It’s inevitable that FDD will become a standard operating technology as the installed base of equipment turns over, specialist industry labor continues to decline and wireless communications technologies become ubiquitous, reducing the cost of implementation.
And we’re making great progress shepherding analytics through the early adopter phase.
FDD veteran James Lee said it at Haystack Connect and I agree – FDD/Analytics is through the early adopter phase and entering the mass adoption phase.
In conclusion, companies developing platforms for buildings should recognize the need for analytics, see where the industry is going, and build a product accordingly.
To me, waiting and seeing ignores the inevitability of analytics. When platforms are built without analytics in mind, it complicates the user experience for building owners and adds more delays to digital transformation and improved performance.
If these building analytics are your thing, I invite you to subscribe to my weekly(ish) newsletter on smart buildings and analytics. It will keep you up to date on the industry and any new article I write.