ASHRAE Journal presents.
John Falcioni:
Hi everyone. And welcome back. Today on ASHRAE Journal podcast, number 15: Industry 4.0, the beckoning of the Fourth Industrial Revolution. I'm today's host, John Falcioni, the editor of ASHRAE Journal.
Our conversation today revolves around the impact of industry 4.0 on HVAC&R technology usage, adoption, and the impact on the end user. Consider a physical world transformed into a digital one, where everything is connected. An explosion of smart devices and technologies allowing us to be in constant communication, 24/7, from anywhere in the world. An Internet of things where physical objects with sensors, processing ability, software and other technologies are connected and exchange data with other devices and systems over the internet and through other networks. A sub-segment of the Internet of Things called the Industrial Internet of Things, or Industry 4.0, is where we focus our attention today.
We're on the cusp of the fourth Industrial Revolution, a revolution that's characterized by a fusion of technologies that is blurring the lines between the physical, digital and biological spheres. The speed of current breakthroughs has no historical president compared with previous industrial revolutions. Through its velocity, scope and systems impact, it's disrupting almost every industry in every country in the world. It is heralding the transformation of entire systems of production, management and governance. And these will be multiplied by further breakthroughs in things like artificial intelligence and robotics, autonomous vehicles, 3D printing, nanotechnology and biotechnology, material science, and energy storage, and quantum computing.
To talk about how this revolution is impacting the HVAC&R industry, we're joined today by Michael Cooper, an ASHRAE member and president of the sustainability division at Bernhard, a full-service design and energy engineering company.
Michael Cooper:
Thank you, John.
John Falcioni:
We're also joined by Stephen Heard, who is vice president of logic at Bernhard.
Stephen Heard:
Thanks, glad to be here.
John Falcioni:
And by our third guest, Bob Tonner, the CEO and founder of Aviexx, a company with a technology that links different parts of the HVAC industry into a single ecosystem.
Bob Tonner:
It's my honor, John.
John Falcioni:
To view the full bios of our speakers, please visit us online at ashrae.org. Michael, we'll start with you. So, in simple terms, what does IIoT mean to the HVAC&R industry?
Michael Cooper:
In simple terms, it really means taking the data that we have within our design elements, through the design process, from fabrication of pipe and equipment, and having that all brought into the building and into the central plant and so forth, and actually being able to utilize that data to the next degree. You think about some of the data collection that's brought forth and being used maybe for advertising or trying to understand what people's needs are in social media. It's really taking that same data and analytics that we have within a built environment, and being able to apply that to both energy efficiency, as well as thermal performance of our buildings to mature, to evolve, to impact the people inside of it. Not just based upon a built-in schedule or something that might be in the building, but actually based upon how the end user inside of that building operates. So, it's really data collection with the existing functions that we have within our HVAC systems, and then applying it to the next level. That's the biggest piece.
John Falcioni:
Stephen, you're out in front when it comes to designing and implementing IIoT-based solutions to your customers. What are some of the new HVAC&R technologies that have emerged because of these advanced processes that Michael's talking about?
Stephen Heard:
Well, I think it's interesting to think about new HVAC technologies versus new technologies, because I think a lot of what's happening is that we're actually reaching out into other industries that have been doing this sort of thing for a while and bringing a lot of these technologies home. So, while they are new to us, it's really great that we can stand on the shoulders of some of the work that has been done elsewhere.
I think some of the biggest technology emergence has really been around the concepts of edge computing and the ability to send computational loads and computational tasks from the cloud to the edge and do things that are closer to the place that's actually going to benefit from the data. So, traditionally with a lot of systems, we're on the edge and this is in the buildings with the traditional hardware and that's sending information up to the cloud, and then there's some sort of arbitrary calculation or analysis is happening, the information is sent back to the edge for use.
But one of the big things about IoT and the concept of edge computing is that you can actually have those calculations happen on the edge in a way that's centrally managed and performed the same as if it was done in the cloud. But instead of having the delays in the latency of transmission from the site to the cloud and back, it can all happen on the edge, close to where the data's going to get used. So, when we're talking about control of systems, we're talking about real-time analysis, we're talking about really low-level control that's required to really get to the next level of control of these systems. This transfer of computational location from the cloud to the edge is going to be really critical for that. So, I think that's one of the big emergences, one of the big technology shifts that's happening. That of course is requiring a different kind of hardware than we've traditionally been installing in the systems, which is where the IoT hardware itself comes in.
Bob Tonner:
The whole idea of edge computing, sometimes people in the industry look at us and they don't understand the word edge. And so we have to explain, it's a local, low-cost, let's call it a medium- to low-power computer on the location of the equipment that can do the critical task offload that duty from the cloud, get rid of the latency, like you said, and make local decisions quickly, and then report the results back up. It's going to be a bit of a challenge for us in the industry, in the IoT industry for HVAC, to get that message across what edge computing is. But I agree with Stephen, that's definitely one of the new things that's happening.
Michael Cooper:
And I think just to add on to Bob's comment, the transition is to take this technology and this knowledge and how do you get it to the end user in a facility? So, at the end of the day, they're functionally taking action. We're being proactive about a hot or cold call because there's certainly these algorithms in the background that are seeing different dynamics in a building, but that information gets brought down to the operator in the building, to take action to deal with maybe a mechanical issue or something in space, so it can be resolved proactively as opposed to some type of inefficiency or function in the building that's not positive.
So, that to me is the biggest piece in the reality, is taking this edge computing and the dynamics and the data that's in the background and bringing all of that to the boots on the ground that can impact change so it provides the right outcome. I think that's the beauty of it all, is bringing that data full circle to provide a significant improvement to the next level of improvement, to energy efficiency, as well as thermal comfort, is making those dots come back together.
John Falcioni:
Yeah, that's great.
Stephen Heard:
Yeah. I couldn't agree with Michael more. I think that one of the stark realities of our industry is that we're not technology limited. We're human limited at the moment. We're limited by our ability to take even the stuff we can find now and push it out. So, the user experience side of it, the technology side, the display side of it, is going to be critical. And even being able to adopt the technologies that are already commercially available outside of our industry.
Michael Cooper:
And you think too, the energy code is always evolving. Just use that as one example, certainly there's interesting technologies that are always going to increase energy efficiency of maybe a coil or some of those functions but there's limiting returns. At the end of it, if we can take that data and impact the human performance, that's the next level of efficiency and that's really something that I think ASHRAE, and not just ASHRAE, but the code standards, the technologies are really going to push in how we can evolve as an industry and really push that to the next level because there's limiting returns on thermodynamics and mechanical efficiencies, but we're really just starting to touch how we can use that information on a day-to-day basis.
Bob Tonner:
The thing that we have been seeing a lot in our work is that getting the data delivered in a way that's actionable. It's very easy to deliver data that's aimed at a high-level engineer or a scientist, but when we're talking about the boots on the ground, they don't have time to sit there and try to figure out this geeky stuff. We have to get it delivered to them in a very actionable way, in a familiar way, that allows them to dispatch the people and the equipment and the parts to get the problem rectified ASAP.
And that's a big challenge for us. It's one thing to get all this wonderful, wonderful data and throw tons of sensors at everything. But until we can get that information interpreted and put in the hands of the people that need it in a way that, once again is actionable, we're going to be, it's going to be a struggle to get it accepted.
Michael Cooper:
Bob, I tell you know, it reminds me of one of my mentors. He told me years ago, "Never design a control system that's more complicated than the end user can actually operate." I could use some personal examples with my kids in a seven day programmable thermostat that I won't get into. But at the end of the day, it has to be used by an operator or an HVAC technician, not a professionally licensed engineer. It has to be intuitive to that person to be able to take that action.
I think sometimes as engineers, we lose sight of our audience and our audience has to be the boots on the ground at the end of the day. And that is, the communication function is probably almost as important as collecting the data and coming up with the analysis. We've got to communicate it in a way to get it to someone who is taking that action and providing change.
Bob Tonner:
When we come from an engineering perspective, long, thoughtful analysis is something we tend to enjoy and we can ponder these things, and we get a lot of self-satisfaction out of that. But the reality is in the HVAC industry, stuff has to be done quickly and efficiently to keep costs under control. And so, that behooves us to make sure that the data is in a way, like you say, that's consumable by those people and not too much information and just the right amount of information. That's been what we've found is a big challenge, is what is the right amount of information to give people and allow them to drill down deeper if they want to, but don't flood them with information that's unnecessary for the task at hand.
John Falcioni:
So Bob, let me follow up along the same lines. You work mostly with residential settings, while Michael and Stephen focus on large commercial type buildings, what are some of the HVAC technologies and innovations that have emerged in your space and what is their potential to change how things are done today?
Bob Tonner:
Well, there's quite a bit, there's some great companies that have come up with some great technologies out in British Columbia. They put an enormous amount of effort into indoor air quality and so monitoring indoor air quality, and then tying that to the equipment so that they can automatically control everything from particulate matter to VOCs, to CO2 those types of things in the air, and really improve occupant comfort with that.
The smart thermostats, I'll be honest, maybe this is going to alienate some of the customer base. I'm not a huge fan of smart thermostats. I don't think they're necessarily smart for the system. I'd like to see that intelligence moved closer to the actual equipment, but there has been some huge advances in the smart, Wi-Fi-based thermostats. So, we're seeing tons of stuff and lots of sensor development too. Now, a lot of sensors that are developed for heavy industry have got high price tags, things like pressure sensors and stuff like that for refrigerant lines. And now we're starting to see companies develop lower cost things that can be rapidly and cost effectively deployed in the residential space, so that we can start picking up data that up until recently has just been cost prohibitive. So, a lot of developments.
Michael Cooper:
I guess Bob though, I want to just tack on one item you mentioned. So certainly, you mentioned getting that information to the technicians. They can do the calculations and so forth, which no doubt, I think the other challenge is bandwidth and the quantity of points. So, I know projects that Stephen and I work on, if you have a hospital that has 200 air handlers and couple thousand VAV boxes and pick your quantity of ancillary pieces of equipment, you have to have the Internet of Things and the analytics, because you just don't have enough time to go and flip through the graphics to see every issue or alarm and really, it's a function of bandwidth we believe as well because there's only so much time in the day to do that diagnosis and having it in the background has been powerful for certainly, our teams that we work with that are boots on the ground. I know Stephen is on the analytics side, but that's the biggest thing too, is just bandwidth.
Bob Tonner:
Yeah and on the residential side of things, it gets really hairy when the seasons change. So when heating season starts, that's when equipment breaks down and a lot of no heat calls. I don't know about your end of the spectrum, you're in the big institutional applications, but in residential applications, every customer considers themselves the single most important customer on the planet, especially when you've got a mom with small children at home and the furnace isn't working, she expects that service guy there instantly to fix it.
Well, she doesn't realize that there's 15 other people that have also called in that day. And so, the time to fuss around with looking at data is not a luxury they have. Sometimes these guys and gals are pulling 18-hour days on when those polar vortexes come into New York in mid-February, or you get a heat spell in Texas in mid-July. They're doing 18-hour days trying to keep people comfortable and the urgency is so strong that if they have to spend too much time pulling out calculators or looking up in manuals to try to figure things out, that's time lost and customer satisfaction plummets. So, presenting data, getting the sensors, getting the information to them in a very quick and actionable way, is going to be a game changer for the industry.
Stephen Heard:
Talk about bandwidth as well. I mean, bandwidth isn't just with the engineering bandwidth. I mean, there's, there's the needle in the haystack problem of being able to find the problem like Michael was saying, and once our engineers find the problem, they know what the problem is, however if you have to monitor thousands of pieces of equipment simultaneously, finding the problem is harder than solving the problem, but then there's also literal network bandwidth issues when it comes to dealing with these amounts of data on a big network, like at a hospital or a higher ed institution, where you might want to be able to get every single point you might have access and hard IOed every single point on a piece of equipment, but trying to get it out of that network, these networks are already overloaded. A lot of times, their infrastructure's ancient. A lot of times their infrastructure's not set up correctly.
So, we talked earlier about one of the key abilities of IoT being able to move calculations from the cloud to the edge. Being able to do that sort of calculation on the edge, solves the literal bandwidth issue of not having to transmit all of this data, not having to wait for data transmission due to latency, being able to make real-time decisions and being able to impact it in a way that might be able to be in a smaller than 15 minute window. You might be talking about a five minute, one minute, 30 second window of being able to enact changes for energy efficiency and automatic response to some of these issues that we're talking about.
John Falcioni:
So, Michael, what is it about the IoT in our industry that excites you the most?
Michael Cooper:
There's two things. One, I'm curious to see how the industry continues to evolve. I'll give some just simple examples. Hypothetically, there's access control systems in a number of—and I'll use a hospital, or higher ed, or maybe a large corporate office building. There's going to be an access control system to get into the building or into the space. They're going to know when you go into the garage, they're going to know when that person, because it should be specific to the individual, comes in and out. And then how do you take that and then tie it into the building automation system, so it knows that this individual goes into the garage at seven o'clock in the morning, they like their thermostat at 74 degrees, they leave at 5:10 every day and then we can offset it to some other temperature. And that's for not just one occupant, but certainly all of the occupants within that building based upon the usage and the integration of multiple systems.
I think that to me, is the most powerful opportunity. We operate and maintain facilities throughout the United States and one of the things that we're curious how that evolves when we're talking to end users, is how do those technologies merge to really provide that next layer of opportunity?
I think the other piece that we see the Internet of Things really coming together is the integration of design to implementation to the end user. I mentioned earlier, so during design, we receive level 500 BIM files, that's brought to a fabrication that then how do we integrate that into the CMMS system? And then how do we integrate activities from the remote monitoring or fault detection services? And then how does that integrate to now create a work order that creates action by that end user?
So to me, the integration of that data, that information, all the way from design, to manufacturing, installation, and now the end user gets to use it over the term and it's tied into all those different systems, that opportunity is huge. I see fundamentally, the challenge is the data is siloed, because we have some sites that we're beginning this integration, but what you find is from an end user perspective of, well, this manufacturer is using BACnet® and then they only want to share certain things and then it's not clean and how that comes together. And that's the big challenge, is getting the information and data to come together so the owner can really realize the full potential.
John Falcioni:
So, let's look at the other side of the equation. Stephen, what worries you about the advent of this technology and then if you would also comment on market acceptance for the technology?
Stephen Heard:
When it comes to worries about advent of the technology I think we are our own worst enemy in a lot of ways when it comes to adopting this cutting edge technology. I think that our industry is used to moving at a different pace than technology in a lot of ways. In our industry, we look at technologies that are five or 10 years old and we think, "Well, we're still using this. It's still nearly mainstream." Whereas you go out to California and something is three years old isn't alive and is being replaced.
So, just the fact that we're a little slower and a little more hesitant, I really feel like the boat's leaving us at any given moment. And the fundamental of it is just that we're dealing with structures and assets that have lifespans far beyond what these other industries have. So, an iPhone might have a three year lifespan or a computer might have a five. We're talking about buildings that are 50, 100 years life. So, the stuff that we're putting into these buildings has to be able to be supported for that long. And we're still supporting stuff that's 15, 20 years old and not even halfway through its life cycle, as far as these actual buildings go. So, I think that's going to be one of the big stumbling blocks of how do we overcome this? How do we build systems that are compatible with legacy, that are not going to end up with these same problems moving forward?
When it comes to sentiment, there's really two levels of sentiment we deal with. We deal with facility managers. We deal with people who are on the ground, turning wrenches. We deal with people who know what they know, like the systems that can make their jobs easier and don't want to put risk into the equation when it comes to this fancy new technology made my life easier. It might not. So, that sentiment is really hard for us to overcome. This comes back to this whole user experience, getting the data to the people in a way that they can consume it, where they can take action and that action actually does things to make their lives easier.
And then we have a completely different type of person that we deal with, which is more of the C-suite, more of the "Oh, this technology is going to make our facilities more efficient, faster. This technology is going to make sure that these investments we're making into our buildings are going to over the course of the next 30 years, provide a better ROI." These people don't have the baggage of our industry and the timelines we're associated with. So, I feel like that marketing acceptance is really—it depends who you talk to, it depends about their perspective. We are saddled with this burden of having to build really great technology that does both things. It has to make facility managers, operational peoples, it has to make their lives easier. It has to provide a better material return on investment than the alternative of status quo technologies. So, that's really what we're up against.
John Falcioni:
Bob, same industry, different market. What's your sentiment on adopting these new technologies in your space?
Bob Tonner:
Yeah. When Stephen was talking, he was talking about some of the old buildings and stuff, and I was having flashback memories of the early 90s when we first started developing controls for HVAC and manufacturing them. And we were one of the first companies to introduce microprocessors for embedded controls, for furnaces and boilers, and that type of thing. And we were heavy into oil heat in the northeast US, on Long Island and places like that, which we still are. But I remember some of the techs and companies saying in their industry, they have an insatiable desire to be second, and that is let someone else test it all out first. And I heard another fellow say, "You know what? I'm happy to use any technology as long as my grandfather used it first."
So, there's still a lot of that type of sentiment. People are saying, "We've been doing it this way forever and it pays the bills and gets the job done. Don't want to change it now."
So, there's that. And the other end of the spectrum, we have the young techs, people who are brought up in the age of computers and smartphones and that type of thing, and they're happy to embrace technology. So we have different ends of the spectrum as far as the types of people.
Our personal experience at Aviexx has been that we get an immediate excitement and enthusiasm from the service companies. They absolutely, quite literally go gaga over what we can do for them, they can see it immediately. And so they immediately start putting it on test homes, on their own homes, et cetera. But when it finally comes to the point where they want to start to deploy it on their customer base and we're talking residential, light commercial, and so they might have thousands and thousands of customers, the problem we're running into is they don't know how to communicate that to their customer base.
So, we don't get to sell Aviexx or our technology directly to the end user. We go through the service company and the service companies are a reactive business model. That is, they wait for somebody to call with a problem. And so, the adoption of this type of proactive technology requires a complete paradigm shift in thinking and the way we approach the industry. So, that's a big challenge. That's what we're seeing and I assume that Stephen and Michael are seeing a similar kind of response, that is thinking to prevent problems, preventive maintenance, that type of thing, rather than reacting to problems that happen eventually.
But that all aside, we are seeing there is a thirst, a thirst and a hunger for this type of technology. And there's an absolute necessity because of certain factors that are playing out in the market these days, that demand this type of a technological solution. We're talking about things like, I don't know if it's happening in your end of the industry, Stephen and Michael, but in the residential, it's the shortage of qualified service personnel. There's just not enough people coming into the industry.
So every day, through attrition, we lose more and more qualified service techs, and yet there's more buildings going up. There's more homes being built. There's more HVAC being installed. So, how then are these two diverging lines going to ever be reconciled? More buildings, more equipment, less people to work on them? The only solution is technology and it's the Internet of Things, especially for HVAC.
Michael Cooper:
Yeah, Bob. No, I agree. I think the Internet of things is that next step, as we mentioned, and I think we're all dealing with, I guess as baby boomers retire and the evolution of the workforce, how do we leverage technology to get us to the other side? I think to Stephen's point earlier, in that the challenge was sometimes upgrading to these systems is—when the reality is some of the DDC and some of the more 20-year-old technologies are, they're pretty reliable and they work pretty well, and there's certainty in that. And so, you have to always have with any new technology, what's the business case? What are we bringing to owners? That's going to put them over the top, so to speak, to push them and say, "You know what? It's worth that capital investment because the ROI, or return on mission is there." Be it through certainly efficiency and upgrades due to behavioral modifications, or impacts in how the system operates, or is there an impact to the end user that provides a return on investment?
If you look in the healthcare environment, certainly there's metrics that drive reimbursements for hospitals that deal with outcomes, and also deal with the experience of patients. So, with anything, with any new technology or new system, how does it impact that owner from a financial standpoint or performance standpoint? And some of that goes to, we have to make certain that we communicate that message, otherwise it is difficult for that owner to transition without having a return on investment or return on mission.
Bob Tonner:
Yeah, the ROI argument is massive. It's huge. And from our perspective, one of the things—well, we have the luxury of being an electronics manufacturing and design company. But from the get go, we've thought about that whole ROI thing. So, we've basically taken the approach is how can we design products in systems that are so inexpensive and reliable, that the ROI factor becomes a no-brainer, that is just make it so inexpensive and quick to install that the tech is in and out of the basement in 15 minutes? That is if it's a northern home, if they're in Florida, they're in and out of the attic in 15 minutes and the system's up and running for under a hundred bucks, that type of thing.
I would imagine that would be exceedingly difficult for your customer base to be in and out in a very short period of time. But that's something that you have to think about right from the beginning in the design of the technology. How do we make this low cost when we're setting up to design it? And how do we design it so it can be installed quickly, easily, and also not having to have the highest caliber, highest trained people, expensive people, installing it? Maybe these can be juniors, people who are new to the trade, that type of thing. So, that way we can get that ROI much faster and much more effective.
ASHRAE Journal:
Seeking a job in the engineering field or searching for the most qualified engineers? The ASHRAE Career Center connects opportunities and candidates. Employers: post new jobs, review resumes, and manage recruiting. Go to jobs.ashrae.org.
John Falcioni:
So, when you talk to your customers, do you talk to them about a proactive approach versus a reactive approach to monitoring HVAC systems? Is that part of the conversation you have with your customers?
Michael Cooper:
Absolutely. I mean, that has to be part of it, John, because to Bob's point, there's less people in the industry. And if at the end of the day, we're always fighting fires it's really difficult to provide the outcomes. And so we found with the clients where we're using remote-based monitoring services, especially in the higher ed and healthcare environments, that that's been a significant game changer for us in being able to address the issue before it becomes an issue, quite honestly. Simple items that may not even get noticed by an end user, where you may have simultaneous heating and cooling of a VAV box, where the space may or may not be ... Maybe within comfort of sorts but at the end of the day, the fault detection recognizes that we have a hot water control valve that's stuck, or an actuator not working, or something of that nature.
And if we can pick that up before it becomes a hot or cold call, and we can address it with the resources we have in a proactive manner, now we're providing outcome. So, that's been our mantra the last several years as we've implemented those strategies in those systems and really, to mitigate some of the issues that Bob brought up relative to manpower. I don't see that changing anytime soon. We have to work a little smarter, not harder. So no, absolutely. That's part of, from an asset management standpoint that Bernhard executes, that's one of our key items that we have to have as a tool in our toolbox.
Bob Tonner:
Yeah, I agree. The other thing that we are seeing is what we call limits to growth. That is a lot of HVAC companies, there's some very, very large nationals who are in every city across the US and Canada, but then you've got a huge number that are mom and pop type size companies, maybe small organizations of one to eight service techs and a small number of trucks.
Now, how does that business owner grow their business? If they want to grow their business, typically they have to buy another truck, equip it with tools and then they have to go out and find somebody to put in that truck that they can put on the road. So, they have to put this huge investment of before they can start taking on new customers. And that is very limiting, especially if you get into the northeast US, where there's a lot of delivered fuels. There's a lot of oil heat on Long Island and Massachusetts, and that ties into the Internet of Things as well.
Before you can take on another 5,000 customers, you've got to buy another truck and you've got to put a person in it and fill it full of oil, or propane, or whatever the case might be. The idea of using technology like we're discussing today, allows companies to grow their business without adding equipment, without adding personnel, because the equipment is just terribly expensive and the personnel just can't be found.
So, by using technology, now you can manage your customers in a more timely fashion, get things organized, far more efficiencies can be found, having your expensive people spending less time on the interstate, trying to get between calls and more time in the attic, in the basement or in the equipment room, fixing the equipment, getting in, getting out and billing the customer. So, when we look at the word proactive, we look at it proactive from equipment maintenance perspective, which we were just discussing, but also proactive in business growth is how do you grow your business? You got to think ahead and technology is the way to grow your business if you're in the HVAC service industry, I suppose other industries as well. So yeah, that's something we think is very important.
And then the other thing is in residential, I don't know if this is the case in your industry, your part of the market, but in residential, the relationship between the HVAC service company and the homeowner is so spotty, that is it happens once every five to six years or so, there's a problem, they call in for help. The service company shows up, they do their work, they move on, but there's no ongoing relationship. And so, there's no opportunity to build trust at that trust-based relationship between the consumer and the service company and the provider. The idea then, is that when it's time for that homeowner put in a new HVAC system, who do they call? Well, right now, they've forgotten about that person that was in their basement five years ago. Now what they're doing is they're googling, they're asking friends and family, who did you use to put in your HVAC system? The idea of an Internet of Things relationship is that the service company is connected to the homeowner on an ongoing basis. And that is, they have not a daily relationship, but a monthly relationship in the form of reports, or custom apps, that type of thing, so the homeowner or the light commercial customer knows who their service provider is. And when it comes time to get new equipment, they've already got that trust relationship with Acme HVAC, or whomever is their service provider.
And then, the benefit of our type of technology that we've been discussing here, is that when it comes time for new HVAC system, whether it's a boiler, heat pump, air-conditioning, furnace, whatever the case might be, now the service company has got actionable historical data. They can help pull up the data and say, "What did this house actually—or this light commercial establishment, what did they actually burn through in cooling costs and heating costs over the last one, two, three years while we've been monitoring? And let's see what we can really put into this."
One of the things we've found in the early adopters of our technology was that most heating systems, forced air heating systems are grossly oversized. My own home has a 60,000 BTU gas furnace in it and it's almost ready for replacement, but with the benefit of the IoT technology that we're using on it, we're able to see that the worst winter, and we're in Canada, the worst day in February, we burnt through 20,000 BTUs in an hour. My furnace is already 300% too big. So, I can see now that a heat pump would be far—a heat pump with maybe a electric heat backup, would be far more than adequate for our house. So, what we're saying is that proactive has got different flavors to it. There's proactive on repair, there's proactive on preventative maintenance. Then there's proactive in growing your business and proactive in getting this sale three, four, five years down the road, when it's time for new equipment.
John Falcioni:
I like where you're going with this.
Stephen Heard:
I was just going to say, I think that the idea of proactive and reactive is going to be evolving as this technology continues to take hold. Especially when we talk about what we can do with proactiveness, with IoT and edge computing versus the current status quo. Because right now, our definition of reactive is wait till it breaks and then fix it, versus proactive is what we try to get a heads up.
But with the idea of edge computing, where we can actually put intelligence, we can actually delegate intelligence down to the equipment level, we can have it watch the equipment itself. We can redefine proactiveness as not as even wait until a human can get involved. Proactiveness can be the equipment itself can build its own intelligence to be its own proactive response to a situation. So, it can actually say, "Well, this is a situation I've not encountered before of myself." I can watch myself. I can detect this anomaly. That is a new behavior that I've never seen myself exhibit before. I can push out a notification to an engineer to say, "Come, what am I even doing? What's going on right now? Can you please help me?"
And the idea is that this comes back to the needle in the haystack situation, where our engineers are not yet spending 100% of their time solving problems. Our technical folks are not yet at that point. They're still spending a lot of their time, just even trying to identify where the problems are. So now, instead of just saying, "Well, we think these are the 20 things we think we might find." Which is the traditional approach to fault detection using the kind of telemetry data that comes out of these IoT devices. Nowadays, we might say, "Well, we think we might find these 20 things. Let's just see if any of these things light up on a grid somewhere. And oh yeah, we found this, we found that."
Instead we can start leveraging edge computing. We can start leveraging machine learning. We can start saying, "Can you just tell us when something's off?" You yourself as a piece of equipment, we're going to bake the intelligence into you for you to tell us when something's broken, even in a situation where we might not be able to predict because we've never seen it before. So, I think that this redefinition of proactiveness is really going to drive the next wave of improvement in our ability to scale operations and to make our engineers and our technicians even more effective.
Bob Tonner:
What you were saying about baking those thresholds, let's call it, but that intelligence at the local edge node, one of the challenges that we've found is that our technology isn't self-learning yet. So, what requires right now is for the installer to set the upper and lower thresholds of various types of things that we're monitoring. And back to what are some of the challenges in the industry. And that is, they've never been asked this before.
Service techs have never been asked to know what the upper and lower thresholds are of say liquid refrigerant approach, or the suction line approach, that type of thing. They really don't know. And so we've seen a lot of them. This is a problem we've been experiencing, is they're just accepting the default settings that come with our edge computing devices, and sometimes they're not suitable. One of the things that, when we developed our technology, our research was done in Canada. We figured that's pretty cold enough. And then our summer research was done in Florida, which was hot and humid. We thought we had really good default settings built into our system and then allowed the users to tighten that up.
But of course, one of our early adopters is in Phoenix, Arizona. We found that summers in Phoenix, Arizona are very, very different and have different thresholds than Florida does, even though the temperatures are similar. So, that proactivity that we're talking about, having the edge computing device be able to figure out when something's gone wrong, right now is still limited by humans that currently don't know it. But I think the answer is probably where you and I are thinking this is going to go, with some type of an AI solution where the edge computing, maybe with some help from the cloud, or maybe we have enough power locally, can start to figure out what is acceptable, what's not acceptable and set those upper and lower thresholds automatically. That's I think maybe where we're going to get to, I don't know if you guys are there?
Stephen Heard:
Well, I think it absolutely is when it comes to this sort of analysis that we're talking about. We're once again, limited by our human capacity. Because ultimately, we're talking about dimensional analysis that's far beyond what can fit in our brains. What can fit in a spreadsheet, what can fit in even really complicated equations. We're talking about if you actually took the kind of regressions that would be able to analyze this data with even four, five or six inputs, you're ending up with equations that are a page long.
So, that's one of the big advantages of trying to leverage machine learning in these area. Machine learning can do that. They can be the equation in a way that can handle dozens, or hundreds, or even thousands of inputs. And we can deploy these machine learning models that can identify patterns that our human brains don't know about and can't even pick up on. And this is one of the reasons why IoT in my mind is actually important just to get engaged, to get installed, to get the data out.
Because once we have the data out and into a central repository, it's almost like gold. It's almost like putting it into the bank because in 10 years we might realize, "Oh my gosh, there's this dynamic that we never even knew existed." And we need to be able to reanalyze every single hospital in the nation to say, "Can we identify this dynamic? Can we implement this new energy efficiency strategy based on this data? Okay, well, wouldn't you know, we didn't collect that data point because we didn't think it was important," because our human brains didn't think it was important at the time.
So, I think that the idea of us trying to get past the limitations of our understanding of these systems and really lean into machine learning to identify these patterns we don't know about, is something that we're going to be able to use again, to get break through the next wall of operational scalability and being able to do this on the edge in real time using an IoT system, or being able to share a model between the cloud and IoT, or being able to train a model on the edge and then transport it amongst your different IoT devices because they're all interconnected through the cloud. It's going to be a game changer for us being able to actually scale this stuff out.
Bob Tonner:
Yeah. Back to the original question about the proactive versus reactive. I'm assuming you guys are doing this as well, but one of the things that we find is really appealing, it actually blows the mind of our customer base is the idea of being able to remotely diagnose equipment and that is remotely activate. And before you even roll a truck before you even put an expensive person in a truck with set of tools, try to figure out what's wrong before you deploy somebody. And so, to be able to sit down, either you sit down at a desktop computer and activate the equipment, try to change some of the variables remotely, or in our case, the service techs if it's a small organization, maybe the business owner him or herself is on the road in a truck.
And we like to say, they can be at the drive through, with a coffee in one hand and run a diagnostic test on one of their customers' homes and get a really good idea of what's wrong and they can see the equipment, they can see, "Oh, it's a 10 microfarad capacitor on a blower and I'm pretty sure I don't have that in the truck." So they can stop at the wholesaler on the way through to that home, make sure they've got the part on the truck. And that's another, I mean this word proactive that we've been talking about, there's so many ways of slicing that word up and looking at it from so many perspectives. Proactive and making sure you've got the right parts. You know what you're doing before you get there, you've got the right person being deployed, just reducing your overall costs of doing your business, it's huge.
Stephen Heard:
Yeah. And I think there's a different direction that we're also coming at it with proactiveness versus reactiveness, which is about, like you're saying, about being able to fix physical things that are wrong with a piece of equipment. But oftentimes in Michael and I's world, we're dealing with issues with programming or control and trying to optimize that. We almost run into an elite athletic type situation where we're trying to get gains at 5% or 2% or 1% or half a percent at a time. And one of the things that a proper, scaled IoT, cloud-driven solution would be able to provide you that a traditional building automation system would not be able to provide you, would be the concept of being able to essentially remotely update the control algorithms, the control schemes.
For example, one of the things that we're really pushing right now is this concept of being able to hot swap control algorithms. So, you start out with the standard control algorithm. You might have two or three different modifications. You don't really know which one's going to be the most efficient, so you pick the one you think it's going to be most efficient and you start running it. But then you run the other control algorithms in the cloud and when you're doing automatic analysis on these to figure out which one is the most efficient at a given time. So after a span, after 30 minutes of, "Oh, this alternative actually has presented itself to be more efficient." You hot swap because this is the sort of thing that you could do with an IoT, with an off-the-shelf IoT system, is simply remotely changed the control algorithm. And at that point you're then using empirical data to drive, "Well, now we're going to switch to this one, it's 2% more efficient." And then you use that one until a third maybe, or the original itself becomes more efficient and you swap it back out.
In the same way, you can actually take the same approach with machine learning because one of the things is machine learning takes data to learn. So if you can do this with a static algorithm, and then in parallel in the cloud, run a machine learning algorithm, learning the data and stumbling its way through trying to control this stuff, eventually it's going to get smart and start identifying patterns and being able to run the systems when that happens, you just swap it. And this is not a tech going out on site. This is not a person going and programming it with their laptop and coming up with some sort of proprietary block language and drawing lines on a screen. This is real technology, real code, getting automatically swapped from the cloud to the edge in that same transition we talked about. This is the kind of thing that's really going to be able to get us from 95% of potential, to 98% of potential, to 99% of potential. That's going to really make almost that elite performance gains that we're going to need to be able to maintain edge in our markets. And it brings a whole different dimension of what we call proactive when it comes to controlling the systems.
Michael Cooper:
And I tell you, therein lies the opportunity and the challenge. Because Stephen, what you just described and you and I have talked about this previously, that ability to understand the situation and be dynamic in making those type of changes is not done at the field level, by the boots on the ground. So, we as an industry have to evolve for that to be able to occur and as well, continue to provide the boots on the ground, the information.
So, that to me is really the opportunity as well as the challenge all at the same time. But if we can, as the industry evolves, take that, move it forward. To your point. I think that's the difference between the pro athlete versus playing rec ball and trying to find that little bit more efficiency along the way.
Bob Tonner:
It's interesting that you should mention the improvement. The parallel in our industry is that we deal directly with manufacturers of the equipment. And one of the things I learned back in the 90s was I'd be in the laboratories of various HVAC manufacturers. And they would say things like, “Well, we have to prepare our product for 10,000 operational cycles a year.” And I'd say, “Where did that number come from?” And they'd look at each other blindly like, “Well, we don't know it. That's what we all say, ‘10,000 cycles a year.’” And it became abundantly clear to me that nobody really knew what happened to their equipment once it left the factory.
I confirm that with our OEM customers this day. Basically when a furnace, or a AC, or a boiler leaves the factory floor and gets loaded on a truck, they don't hear about that piece of equipment again until there's a problem. So, they really don't have any hard data to develop the next generation of equipment, it's just gone. And the only information they get is complaints, warranty calls, that type of thing, and they don't necessarily get the straight goods on what really happened. Often these failures are clouded by poor information about the installation, et cetera, et cetera. So, what we think IoT will do for the industry is allow the continuous live feedback to the OEMs so they can see their equipment operating and seeing like, "Oh wow, we didn't design for that."
John Falcioni:
Gentlemen, I do have a couple of other questions that I think are significant. So, let me just ask, a lot of the conversation has been about disruptive changes to the HVAC industry and the implications of IoT. Is there such a thing as an HVAC killer app for IoT?
Stephen Heard:
I certainly have my thoughts on what it would take to be a killer app, but you got to think about disruptive technology versus sustaining, or non-disruptive technology. And one of the key things with disruptive technology is as often doesn't work as well at the beginning, right? It often is not as great. It's a generational improvement, but it's not an actual literal improvement until it gets adoption. I think that we have to overcome this barrier of adoption.
So in my mind, all of the killer apps are about simply that. How do we make this technology easier to adopt? How do we make it easier to work? How do we reduce zero to 60 time? How do we take it and make it automatic, “automagic”? People use this “automagic” word, that's kind of an eye roll, but it really makes a complete difference.
So, the idea is that there's so many applications in our industry that are overly cumbersome, that we say are standardized and say are scalable, but really aren't. So, for example, you could easily build a piece of technology that you go and you plug into a mainframe, syncs up, walk out to a terminal, has an IO board on it, plug whatever random wires you want in whatever random thing, and just flip it on and walk away. And it could be at that level of complication to get this set up. You go back to the mainframe and by the time you're there, it's already synced up. It's already figured out what all the inputs are based on the signals that it's getting, and it's already mapped everything and it already has fault detection and a graph and you can already get a ping on your phone saying that, "Hey, the discharge air temperature on this terminal is lower than it should be."
I mean, that level of automatic working, that level of “automagic” setup and ability to make things happen smoothly and quickly is doable. It's doable because other industries are already doing it, this exact thing. So, once we can get that level of adoption where you can, just like you were saying earlier, we can get these younger techs that are used to playing video games, but not used to turning wrenches and have them come up and start plugging things in, and it just works the way that they would expect it because it's the same way that every app they've ever used works. I mean, that is going to be the killer app for HVAC controls. It's going to be when people don't even realize what's going on, it just works.
John Falcioni:
So, a couple of other points that have come up in the course of the conversation. You've spoken a great length about the merits of the IoT and edge computing, but when is edge computing not a good solution, what do you do then?
Bob Tonner:
We have safety and programmable systems. So, if you're controlling things like gas valves, burners, equipment that where a malfunction could result in a dangerous operation, either danger to humans or to equipment, then I think that we have to keep a significant firewall in between the equipment and the cloud. That's my perspective. I think that's what I would see as the limitation for edge computing, meaning that when it comes to safety functions, for now I think those things have to be isolated from the cloud until we get to a point where it can be proven, safe, and effective to download new operating programs from the cloud to a piece of equipment. I don't see that right now, personally, but who knows? It could happen.
Stephen Heard:
Yeah, I agree. I think it's going to be important not to let the tail wag the dog, when it comes to the technology here, this IoT approach should be used to simplify things and not make things more complicated. So, if there's things that are already built simple for a real reason, like security, or just extreme reliability, I don't think there's necessarily a reason to introduce all this complexity.
For a lot of applications, it makes a lot of sense because we're actually making things simpler, I believe, but just like Bob said, in matters of extreme safety, security, there's definitely limitations to IoT in those areas that it is connected to the cloud. It is hackable. Foreign agents can get in and get into systems and do nefarious things. So, I mean, that's just a reality of the situation. It's a risk that you have to balance like anything else.
John Falcioni:
So, it looks like the opportunities brought on by industry 4.0 are terrific, but the adoption challenges may be just as great. So, how about some final comments, Michael, we'll start with you, then Stephen and Bob.
Michael Cooper:
Well, I guess John, we touched on this earlier, as the industry evolves. I think the opportunity for us to take the data that we have that's already in the technologies we're using today to provide a higher level efficiency and a higher level of thermal comfort in the commercial building space is tremendous. So, that opportunity, I think certainly outweighs some of the hurdles and challenges that we need to utilize and the opportunities to be more proactive. I think it's exciting.
I think I have another 20 years in the industry at a minimum, so I'm really looking forward to see how that evolves over the next, even two to four years. Because as others said earlier, I don't think this is going to be a generational change of the building or the built environment. The change is going to happen more rapidly as the technology continues to evolve. So, really looking forward to it.
Stephen Heard:
I definitely agree. I think it's going to be really important to make sure that we're talking and having these conversations and continually to head in the right direction. I mean ultimately, the goal of any of this technology is to save energy. Is to save energy, reduce greenhouse gas emissions, lower cost. And ultimately, we're up against the second law of thermodynamics here. You can only save so much, there's a maximum and this technology has the potential to be way oversold about what it can do and way oversold about the benefits it can accomplish. And even the applications, just like we were talking about, because it's not perfect everywhere.
So, I think that the challenge in my mind is keeping your eyes on the ball, focusing on energy efficiency, making sure you're adopting technology in ways that are actually improving operations, actually improving energy efficiency. And then just making sure that you're continually headed in that direction. I think it's much more important for us to be heading in the right direction than it is for us to be in any particular situation because that's going to change. The adoption challenges are really being addressed by just having these conversations and just even making the attempts.
Bob Tonner:
Well, from our perspective in the residential and light commercial, energy efficiency is an important thing, but for us, I think it's more operational efficiency that we're very excited about, that is using this technology to vastly improve the operational efficiency of small, medium, and even large HVAC service companies, so that they can get the job done faster, better, cheaper, and with a better relationship with their customers. We're so excited to be in the industry, in this technology, because we just think the time is now for the approach we're taking. And we're just seeing the vast majority of people are, like I said earlier, hungering and thirsting for this type of thing. And we're just thrilled to be part of it.
John Falcioni:
I want to thank our guests, Michael Cooper, Stephen Heard and Bob Tonner for being with us today. From all of us at ASHRAE Journal, I want to thank you for tuning in today. I'm John Falcioni, join us next time for another conversation.
ASHRAE Journal:
The ASHRAE Journal Podcast team is editor, John Falcioni; producer and associate editor, Chadd Jones; assistant editor, Kaitlyn Baich; and associate editors, Tani Palefski and Rebecca Matyasovski. Copyright, ASHRAE. The views expressed in this podcast are those of individuals only and not of ASHRAE, its sponsors or advertisers. Please refer to ASHRAE.org/podcast for the full disclaimer.