Cloud Computing
Cloud
computing has been
prevalent in consumer products for nearly a decade. Perhaps the most
common example we see in our day to day lives is in intelligent
personal assistants on our phones, such as Siri. The technology to run
this natural language user interface does not exist in each device,
rather it exists in a cluster of computers in a data center. Each time
an individual makes a request of Siri, it is sent to a data center,
analyzed, and a solution is returned to the individual device. This
happens so quickly that it may seem that each device is equipped with
its own personal assistant.For the purposes of building automation systems, we are already starting to see many front-end user interfaces being moved to the cloud. Rather than having a physical head-end computer that is used to access the BAS in each facility, there are now hubs that receive and transmit data to the cloud. This allows the facility operators the ability to access their system on any internet connected device.
Moreover, by moving this data to the cloud, we are able to analyze how buildings are operating with more capable machines, giving facility managers insights into how to operate their HVAC and lighting systems more efficiently. Accumulating this “big data” and its subsequent insights for hundreds or thousands of buildings then enables BAS companies to utilize machine learning algorithms for research into the patterns and commonalities of well-run buildings.
Machine Learning + Artificial Intelligence
Yes, this is where Skynet
begins, but rather than using science fiction movies as examples of
machine learning, let’s try a harmless, real life music app that most
people have come across at some time or another.Pandora is a free music streaming application that makes automated recommendations based on the music listeners have previously “liked”. First, users choose a genre, artist, or song they like and a “station” is created. Based on particular features of that genre, artist, or song, Pandora plays what it perceives to be similar tracks. Users then indicate, in the app, whether or not they like this song. As listeners feed Pandora more information about their preferences, the app “learns” what kind songs they will and will not like.
Unlike cloud computing, machine learning and artificial intelligence are not nearly as common in today’s building automation systems. Some systems have features built into their software that represent the beginnings of this technology, but none have been able to capture the true benefits we foresee for large environments.
Much like Pandora, your BAS needs some initial input to effectively start learning. This initial input will likely come in the form of amps, voltage, temperature, etc. from your various field devices. One of the greatest advantages we gain from machine learning is the ability to recognize patterns and anomalies. It’s in this advantage where we find our first application for machine learning in building automation.
As your BAS is able to start distinguishing what is “normal” for your system through pattern recognition, it can also start detecting abnormalities. For example, your system might be able to notice if an air handling unit has a progressive decrease in amperage over a long period of time and notify you to make a preventative maintenance call on the fan belt. This is only possible because your system was previously given data that indicated what the normal amperage was.
The other, perhaps even more exciting, use case for machine learning in building automation is in predictive analysis. By combining historical weather and BAS data with future weather forecasts, your building automation system of the future will be able to predict how our facility is going to run the next day. Facility engineers will be able to fast forward through the following day and see how each piece of equipment behaves and how much energy is used at various times of the day. They can then make adjustments to optimize the functions of their mechanical systems.
Connectivity + Integration
In recent years, home automation
has taken off. Between smart thermostats, light bulbs, plugs, speakers,
locks, and security systems, the home automation market has progressed
exponentially faster than that of larger commercial facilities. The
Internet of Things is truly becoming a reality in the homes of those
who can afford it.Furthermore, many of these devices can be interconnected and automated via mobile applications. Based on an individual’s location and personal preferences, these apps can lock and unlock doors, change the setpoint, set the alarm, and even turn on/off lights. This is where the IoT rubber meets the road.
On a commercial level, there are some (but not many) buildings that are implementing similar concepts. The Edge, in Amsterdam, is perhaps the best example of this. The Edge is a net zero facility that utilizes a smartphone app to customize each tenant’s experience. Among its list of impressive features, The Edge’s app assigns tenants a desk each day based on what is on their calendars. Their assigned desk will then be set with their preferred lighting settings.
[an error occurred while processing this directive]The Reality
Often times these technologies are discussed as if they already exist and can easily be implemented, but in reality, neither is typically true. Many of them, such as machine learning, have been conceptualized for decades, but we are only now getting to a point where the technology is catching up enough to put the concepts into action. It’s an exciting time in the building automation industry as we look forward to seeing more automation implemented to make our lives easier and make facilities more efficient.