September 20, 2020

HVAC optimization: an interview with Fulvio Bartolucci, Tera GM

Let’s start with something really basic: what is HVAC and how does it connect to Tera’s work?

HVAC is heating, ventilation and air conditioning – the air that circulates through a building to make it comfortable. For Tera, HVAC optimization has always been a core solution. In fact, automating and optimizing HVAC was the first thing we focused on when we created our AI energy platform. Today, it’s still a specialty of ours, especially for air conditioning.

Why air conditioning especially?

Part of it is simply geography – we’re based in the APAC region where it’s hot and humid for much of the year. But there are other reasons the demand for cooling has grown so much. Today, Asia’s workforce is more white-collar and service-based than ever. With that comes more office space, virtually all air-conditioned.

The projections are that by 2030, 95% of non-residential property in China will be cooled. So, if you want to be serious about cutting C02 emissions, balancing the need for comfortable, cooled workspaces with HVAC efficiency is incredibly important.

What’s the simplest pitch you can make for HVAC optimization?

There are many ways to answer that, but I usually start by talking about speed. This is a solution where we can quickly make your facility much more energy-efficient without causing any major disruptions to normal business or requiring any system overhauls. So the flow of work at your site continues as normal, and within a month, you’re spending less on electricity than ever before; contributing far less to carbon emissions than you were before.

Of course, everyone is looking for value and savings – this is a quick way to cut your energy bills, and if you’re being taxed on carbon use, you’ll see compounded savings from a reduction in emissions as well.

What is the biggest cause of HVAC inefficiency?

There are two major causes of HVAC inefficiency; one is technical, the other is human.

On the technical side, there is a sense in which HVAC system controls are “blind”, and this causes delays. By this, I mean that HVAC systems usually operate without directly tracking variable factors like how many people are in the space, or external conditions like weather. That lack of data input leads to a delay in reaction, and from there to overproducing chilled air, wasted energy and higher bills.

And the human side?

Lack of understanding. Somewhere around 80% of properties aren’t doing anything to optimize HVAC usage. Most property owners can only say “I pay this much, this is all I know.” They aren’t collecting the right data, and so any attempt at optimizing their HVAC system will essentially be guesswork.

What does Tera do to solve this?

What you could say we do is “hack” the existing system. We start by installing sensors around the property, then take over the control logic and give our software – our AI – control over the HVAC system.

The goal is to reduce the time lag between the data input and the central chiller control. The “hacked” system will have a more sophisticated range of inputs, such as room use and external temperature, to forecast the cooling load. This data, along with more accurate internal temperature data, is sent in real-time, for a highly efficient HVAC system that uses machine learning to automate use and save energy.

What kind of properties can you optimize?

We mainly work with properties that have one central AC system – industrial and commercial properties. HVAC optimization is a holistic project that takes on usage across an entire building.

What about properties without centralized HVAC?

If we are talking about properties that have split two or split one AC— the type you would turn on room by room— then you need to find a different solution. That can’t be controlled centrally, but what we could do is basically remove the possibility of manually turning the AC on and off. We replace that with an IoT device to control it.

So by taking away manual control, you cut take away the human error?

Yes, through automation. When you just forget to turn the AC off; when there’s a space that people aren’t spending a lot of time in, and so on, it’s very easy for the air conditioning to stay on at 18 degrees all day, wasting energy and creating unnecessary emissions.

But, it is not only about error. It is also about finding a compromise between individual preferences and collective impact.

Could you explain that a bit more?

Here’s a scenario that is easy to imagine. In “Room A” there is a person who wants things to be very cold and he always turns up the AC so it is like a refrigerator. Next door, in “Room B” the employees want it to be rather warm. If you leave things completely to individual preferences, you may end up with a hot area next door to a cold one. That creates a situation where both rooms become inefficient because both are spending more to maintain their preferred temperature.

With a centralized system (wired or virtual) we manage the temperature in the whole area so that it’s more or less homogeneous, but still gives enough individual control that people are satisfied. They can still modify temperatures and get their cooler or warmer rooms, but within a controlled range.

Let’s come back to your idea of “hacking” HVAC systems. What happens once you’ve done this?

The “system hack” is Phase 1. By setting up the optimized machine learning system, we will unlock access to extremely precise data about how heating and cooling are used at a property from day to day, month to month and across seasons—what we would call the load curve. That’s how we generate savings, and that’s where AI and machine learning really come into play.

Phase 2 is a further step that some clients opt for. That is when we use the data generated in Phase 1 to recommend upgrades to equipment, such as chillers, fans or heat storage, or beyond, to safely integrate low carbon technologies, like geothermal storage for reduced emissions.

How does the HVAC system data enhance the asset upgrade?

Once we have the precise load curve from Phase 1, we use data modeling to tell you exactly what the savings would look like if we replace your chiller. We remove the assumption factor. We can look at the data to plan the right moment to go in and replace equipment. That’s possible because now you have a model and can adapt it to the situation. You can say “okay, what if we replace it now or then—how will that impact the rest of the system, how will that impact operation?”

We have a very science-based, mathematical method to tell you precisely how an equipment upgrade will look and what the savings will be. We’re removing the guesswork, taking the assumptions out of the project.

Let’s talk numbers. How big are the savings generated by HVAC optimization?

All our clients can expect an average of 17% in energy savings related to HVAC usage. And that is just from Phase 1.

If a client decides to move on to Phase 2 and begin replacing equipment, then they will begin to see compounded numbers and easily reach over 40% of reduction in energy use from the baseline. If they choose to go the extra step of heading to low carbon technologies like geothermal storage as well, then you’re looking at reducing your carbon footprint by 80%.

In just a few words, what is Tera’s mission in offering HVAC optimization?

Tera’s mission, through all of the solutions we offer, is to make the future of industry smarter, more sustainable and more efficient. HVAC optimization is one of the simplest and most convenient ways for a business to do this.