Out of the living room, into the boiler room

AI is being used across industries to derive insight and drive value. Can it help eliminate energy waste in multi-residential buildings?

Artificial Intelligence (AI) is playing an increasingly important role in our everyday lives. Whether it’s in suggesting a reply in your email inbox, finding a rideshare in minutes, picking a song on a playlist or checking credit worthiness in seconds, it’s hard not to recognize the growing impact of AI and its applications. But if AI is so powerful and pervasive, what role can it play in helping re-imagine efficient energy use in the buildings we live in? We’ve embraced AI in our living rooms, but perhaps it’s time to introduce it into our boiler rooms as well.

Multi-residential buildings face an ongoing and significant challenge: energy waste. Typically, the energy (electricity and gas) that buildings manage and control is primarily consumed by HVAC equipment (provided each unit in the building does not have its own heat pump or equivalent system). This can account for 35 to 50 per cent of the utility spend for any particular building.

Innovation has been slow to come to HVAC management. For the most part, today’s multi-residential buildings manage their HVAC energy the same way they did 20 years ago. The process is not data driven, dynamic, or measured against historical patterns and it certainly is not forward-looking. If anything, it continues to reinforce an unacceptable pattern of waste.

How does it work? First, let’s talk about data. AI depends on large sums of data. The most successful AI implementations in industry pair data-processing AI engines with human expertise. An example from healthcare is the potential for AI to assist in identifying different skin cancers through images; a study out of Stanford demonstrated that a trained system can meet the success rates of fully-trained dermatologists.

Thus, the first substantial barrier for AI use in HVAC energy management is obtaining these large sets of information. With sets of labeled data, AI systems can learn what normal operations look like, how to identify states of failure, and where efficiencies can be found. Thinking about HVAC in particular, we need data sets that capture the interplay of sub systems both across a range of multi-residential buildings (with differing demographics and usage patterns) as well as across year-long timelines. This is important because a building with an older, largely retired demographic will present a different data pattern than a much younger working demographic at 9 a.m., for example. AI can help us segment the exact needs and usage patterns of varying buildings and see in usage patterns how best to provide comfort to residents.

Multi-residential boiler rooms like this will need to be connected and controlled to leverage AI.

In addition to data on the primary equipment, we will also need data on affected building areas such as hallway temperature. Further, any fulsome analysis requires additional input including external weather and third-party data where available (ie. utility pricing rates).

The more data the better

In order to capture this data, building owners and property managers need to acquire capabilities to collect and store it. In the past, this requirement may have been enough to dissuade property teams from considering adopting new processes. New technology, however, can help facilitate powerful information gathering. Improvements such as lower-cost sensors, and better connectivity to pass data to the cloud, can help alleviate the burden on property managers and building owners.

With hardware in place to collect data in real time, the industry will need to think about new levels of control that can leverage AI data processing. If existing equipment and systems can be controlled with new levels of accuracy, there is a greater opportunity for uptake; retrofitting existing equipment with new levels of control that are powered by cloud-based AI engines will likely be the most cost-effective way to leverage AI.

Unsurprisingly, a Berkeley Lab study found that, “the higher the initial cost and the longer the wait for returns, the less likely a project is to be approved.” And thus, the challenge when capital-intensive equipment upgrades are proposed is they are riskier with longer payback times. Smart solutions don’t require new investments in big equipment; they can connect to legacy systems and use AI to make them efficient.

With lightweight controls in place and data pipelines that can fuel AI engines, the real work will begin. Energy Management Platforms will need to be developed that can both make smart decisions and offer insight and control to domain experts such as mechanical technicians. The net result should be systems that can provide the right level of detail and information to the right parties whether that’s the condo unit owner, condo board member, property manager, mechanical partner, energy provider or research partner.

The potential benefits of this kind of system are many including energy savings, environmental impact and greater operational insight.

Condo managers and boards are under increased pressure from their residents to control escalating fees. This is no easy task. If buildings can reliably lower their largest operating expense — utilities — they can help to secure the financial viability of their building and better secure the investment of all unit owners. In my experience, when AI is leveraged appropriately, buildings can realize 20 to 30 per cent savings on utilities, without any upfront capital investment. This is prudent financial stewardship.

With lightweight retrofits we can help to eliminate energy waste in condo buildings – AI can help.

A second benefit of reduced energy waste is the positive environmental impact. According to a UN report, buildings and construction account for 39 per cent of global CO2 emissions. As energy consumers, it behooves all of us to think about how we utilize resources and where we can be more energy efficient. As society transitions into denser and denser urban environments, our buildings’ environmental impact will only become a bigger issue. In Ontario alone there are more than 1.3 million condo residents; across Canada, it is estimated that one in eight households live in condo buildings.

A third benefit of AI in buildings is the opportunity for more advanced alerting and notification when equipment is operating incorrectly. Real time data processing can help the equipment maintenance staff and give property managers better peace of mind that issues will be caught immediately on behalf of residents.

A final benefit for AI in buildings connects to resident health. If we can better instrument our buildings and capture data related to the quality of our buildings’ environment (air and water in particular) we can better inform residents on the health of their environment and better help safeguard it. Increasingly, we are able to monitor all aspects of health with the help of AI and data (ie. the heart monitor on your smartwatch). The same should be true of the environment where we spend so much of our time.


John Macdonald is Chair of the Board of Parity, a Canadian proptech company. Previously, Macdonald was the first employee and CEO of Enercare, which has grown to be North America’s largest HVAC company. www.paritygo.com

 

You might also like

Comments are closed.

single-podcasts