Join us at NexusCon 2024 - Get your ticket now
"The platform of automated demand management is important for Macerich. The rate pressures that we have seen from utilities have been extraordinary. It’s not just about reducing, but being smarter and more strategic with those reductions. This platform gives us the flexibility and the future proofing that we think we’ll need for the next 5 years.”
—Ryan Knudson
Welcome to Nexus, a newsletter and podcast for smart people applying smart building technology—hosted by James Dice. If you’re new to Nexus, you might want to start here.
The Nexus podcast (Apple | Spotify | YouTube | Other apps) is our chance to explore and learn with the brightest in our industry—together. The project is directly funded by listeners like you who have joined the Nexus Pro membership community.
You can join Nexus Pro to get a weekly-ish deep dive, access to the Nexus Vendor Landscape, and invites to exclusive events with a community of smart buildings nerds.
Sign-up (or refer a friend!) to the Nexus Newsletter.
Learn more about the Smart Building Strategist Course and the Nexus Courses Platform.
Check out the Nexus Labs Marketplace.
Learn more about Nexus Partnership Opportunities.
Episode 162 is a conversation with Ryan Knudson from Macerich and Jon Schoenfeld from Buildings IOT.
Episode 162 features Ryan Knudson from Macerich and Jon Schoenfeld from Buildings IOT and is our 9th episode in the Case Study series looking at real-life, large-scale deployments of smart building technologies. These are not marketing fluff stories, these are lessons from leaders that others can put into use in their smart buildings programs. This conversation explores how Macerich has incorporated Buildings IOT technology into their shopping centers. Enjoy!
You can find Ryan and Jon on LinkedIn.
Introduction (1:18)
Introduction to Ryan (2:18)
What is the project? (3:57)
Vendor team (5:03)
Project details (5:37)
Project results (7:04)
Introduction to Jon (7:40)
Original goals (9:03)
Requirements (11:45)
Other options (16:28)
Phases of operation (19:05)
Supervisory control (21:16)
Utility data (26:45)
Operator engagement (29:54)
Results (32:36)
Lessons learned (41:24)
The business case (51:32)
What would you tell others? (54:28)
Music credits: There Is A Reality by Common Tiger—licensed under an Music Vine Limited Pro Standard License ID: S569626-16073.
Note: transcript was created using an imperfect machine learning tool and lightly edited by a human (so you can get the gist). Please forgive errors!
Ryan Knudson: [00:00:00] The platform of automated demand management is important for MaceRich for a couple of reasons. One, the rate pressures that we have seen from utilities over the last 24 months have been extraordinary. And it's not just about reducing, but being smarter and more strategic about those reductions. And this platform gives us the flexibility and the future proofing that we think we'll need for the next five years.
James Dice: Hey friends, if you like the Nexus podcast, the best way to continue the learning is to join our community. There are three ways to do that. First, you can join the Nexus Pro Membership. It's our global community of smart billing professionals. We have monthly events, paywall deep dive content, and a private chat room, and it's just 35 a month.
Second, you can upgrade from the Pro Membership to our courses offering. It's headlined by our flagship course, the Smart Building Strategist, and we're building a catalog of courses [00:01:00] taught by world leading experts on each topic under the Smart Buildings umbrella. Third and finally, our marketplace is how we connect leading vendors with buyers looking for their solutions.
The links are below in the show notes, and now let's go on to the podcast.
Welcome to the Nexus podcast. This is the latest episode in our series diving into case studies of real life, large scale deployments of smart building technologies. I think this is our 10th. So we're getting into double digits. Uh, we've been doing this for a year and it was really a, our attempt to fill the void in real life case studies and real life stories that we've seen in our industry.
Um, And so we're not here to like create some sort of fluff story. We're here to really, really share the lessons from leaders so other people can put these lessons to use in their smart buildings programs. So today we have a story coming from Mace Rich, which is a leading owner, operator, and developer of top real, uh, retail and mixed [00:02:00] use destinations and major U S markets.
So if you're like me and you're in from Colorado, we have several, uh, shopping malls, shopping destinations here. That are Mace Rich properties. There's probably others in your areas, if you're listening to this or other, other cities in the U S as well. So the voice you heard at the beginning was Ryan Knutson, the vice president, uh, corporate responsibility and sustainability at Mace Rich.
Ryan, can you talk about, uh, introduce yourself and talk about your background a little bit so people can get to know you?
Ryan Knudson: Thanks, James. Uh, as you said, I'm Ryan Knudson. I'm VP of Corporate Responsibility and Sustainability here at Mace Ridge. Uh, I've been here for about 13 years. My career here at Mace Ridge has been unique.
I started, um, in our IT department as The building automation technical expert and really helping Mace Ridge in their first iteration of energy management and sustainability. And I've had the unique opportunity as Mace Ridge has grown its sustainability program [00:03:00] to grow along with it, um, touching different parts of the business from IT and, and, uh, building automation to operations, maintenance and facilities management.
Um, and then, uh, coming. Finally, here to the sustainability team
James Dice: and Ryan, you were on the, on the podcast before it was episode 94, if people want to go back and, and Ryan talks about why data is important in his role in sustainability and, uh, why data is important for the ESG journey as well. So that episode is episode 94.
We'll put a link to that in the show notes. Um, and this conversation, we're kind of diving into this one project that you guys have done. As part of the overall smart buildings program and sustainability program, Ryan. So let's do a couple sort of rapid fire questions to give people context on the project we're talking about.
So, um, What are we talking about here? How do you guys refer to it in house? What's this project called?
Ryan Knudson: So we refer to it in [00:04:00] house as EMS 2. 5 because we are not that clever. But no, it is a continuation of our building management program that we've been doing for 15 years. Started with Basic Tritium Niagara integration, unlocking the buildings, and then moving into standardizing our base building infrastructure with BACnets and, uh, diving into the sequencing of operations.
Um, and now we are at the point of deploying, uh, false detection diagnostics, um, and automated demand management on top of those, uh, investments. We very much view this as a continuation and, uh, a way of deriving new value.
James Dice: Totally. I'm glad you shared the 2. 5 tidbit because it really, it really shows the, the, the, the engineer's ability to name things.
It's [00:05:00] funny, funny to me. Um, so who's your vendor team?
Ryan Knudson: So our project team was made up of, uh, in house, uh, developers. Energy management experts with Mace Rich along with OTI acting as our master systems integrators and Buildings IoT as the platforming and engineering team. Additionally, we had our on staff maintenance team.
Uh, M Core teams acting as the key, uh, doers, uh, of the, of the software.
James Dice: And how many buildings are we talking about here where you've, uh, installed this technology stack?
Ryan Knudson: Sure. So to date, um, and it's, uh, April 25th, 2024 to date, we've, uh, deployed it in eight centers. Um, but we are actively working on a 2024 pipeline to grow that.
Um, with an eye to deploying it to, um, all of our major assets, uh, by the end of 25.
James Dice: And what's the total square footage we're talking about here? [00:06:00]
Ryan Knudson: Yeah, so of those eight centers, we're talking about three of our largest centers. So those eight centers really encompass, uh, almost 10 million square feet of, uh, combined, uh, tenant and common area space.
Because some of these, uh, systems do interface with our tenant systems.
James Dice: . And when did this project start? I mean, I know you talked about the controls upgrades and Niagara and all that stuff, but when did this piece that we're going to talk about today start?
Ryan Knudson: Sure. So we actually started piloting it with Buildings IoT in kind of Q3 of 2022.
Um, but we didn't really decide to move in earnest until Q2 of 23 and really started installing in earnest in Q3 of 23. So we've had eight centers over the last two and a half quarters.
James Dice: And we're going to talk about the like deep dive into the results in a little bit. But can you just like give us big picture?
What results have you seen that make [00:07:00] you kind of want to expand this from the projects you've done so far into the rest of the portfolio?
Ryan Knudson: Sure, so our results vary, uh, as you can imagine based on our building type and the, um, mechanical systems, uh, at each center, but generally speaking, we have seen, um, somewhere in the neighborhood of, uh, close to 100, 000 annually, we're projecting in, um, utility savings, um, which will be kind of carved out in different ways, um, which we'll dive into, um, a little bit later.
James Dice: So we also have, um, so John Schoenfeld here, VP of Energy and Building Technology at Buildings IoT, um, to kind of represent Buildings IoT in this, um, discussion here. Uh, welcome, John. Can you introduce yourself?
Jon Schoenfeld: Thanks, James. Glad to be here. Take care. Uh, name's John Schoenfeld. Uh, I've been with Buildings IOT for, uh, going on eight years.
Uh, prior to that I was a principal at KW [00:08:00] Engineering, where I focused on retro commissioning and energy engineering of large portfolio, uh, clients. Been really a, a great ride here at Buildings iot. I've watched the, the company really transform itself from, uh, a master systems integrator, uh, to a platform developer.
Developing products like an independent data layer, a smart buildings platform, and now getting into advanced supervisory controls with automated demand management. You know, it's been a great eight years and we've had the pleasure of working with Ryan and the Maestrich team for that entire time, and I've really had a lot of success in, you know, taking a collaborative approach to Uh, to the platform development, but also the energy reduction.
James Dice: Totally. All right, let's dive into this case study. So Ryan, can you talk about a little bit, you talked about it a little bit earlier, but why did you guys start down this [00:09:00] path and what were the original goals for, for this program?
Ryan Knudson: Sure. So, um, you know, we were really looking at all of the investments that we've made over the last 10 to 15 years.
And, and John's been a part of that for, for eight, um, in our control system. And how can we. I'm a researcher and I'm trying to leverage all of that technology and all that compute power in new and novel ways to derive new values. One of the key outcomes that I was hoping to achieve was to find a frictionless way of automating our, uh, demand management and controls, um, without the need of human intervention, because as I'm sure everyone can tell you, you can put in the best control system, but if the operators treat it like a light switch, you're not going to get the value.
And I'm not saying that's what was happening with us, but I was looking for something that we could use that wouldn't [00:10:00] require our operators to change their behavior.
James Dice: Totally. Can you, um, I want to circle back on that, but for just a second, but for those of you, the, the, the audience that doesn't know about Mace Rich's bigger goals around net zero, um, really around like, why is the business doing this?
Can you talk about like the bigger picture outcomes real quick?
Ryan Knudson: So Mace Rich has a couple of, uh, stated goals. First is, uh, carbon neutrality by 2030, but then extending beyond that to operational net zero by 2035. Um, and that's going to require, you know, reducing our energy consumption and our, and our scope to emission by 50 to 60%.
Um, and a lot of that can be driven by hard energy assets, such as renewables, but it also requires smart technology inside of the building to heat and cool the assets in a, uh, appropriate fashion for the comfort of our guests. But not, um, overly [00:11:00] so that we're becoming wasteful. Um, we also look for technologies that allow us to become, uh, adaptive to the new, you know, grid, uh, conditions.
As these utilities start to transition. Their energy supply away from fossil fuels and into renewables and, um, you know, various different types of new technologies, our buildings are going to need to be able to adapt to those new, um, supply conditions.
James Dice: So can you talk about like, What the technology stack was like before this project, and specifically relating to what you just said a second ago, what did it require the building operators and the maintenance teams to be doing in order to beat those goals?
It sounds like it needed to be, um, the control systems needed a lot of hand holding to get, you know, What you wanted out of them.
Ryan Knudson: Absolutely. I think, you know, where we were in our journey, um, before we started, uh, with John and his team was, um, we had a, a [00:12:00] pretty standard control system, you know, back net system doing all of the, if, if this, then that logic, uh, rolling up to a Niagara, uh, Jace, uh, and then that was being piped into, um, a supervisory controller, uh, And it was all operating with kind of mixed results, depending on, you know, how sophisticated our operators were.
We had some centers, and we still do have some centers, that are deriving a lot of value from that advanced control system. And then we have some centers that have a more kind of set it and forget it mentality. Um, just based on, you know, their everyday workload. So we were really looking for something that would, um, provide value to all of our, uh, facilities teams, regardless of their sophistication.
James Dice: Totally. And, and John, as you guys think about your, your entire customer base, uh, buildings, IOT, can you talk about kind of what motivated [00:13:00] you to, um, Go into this, this sort of, um, we'll get to it in a second, like what we mean by demand management and all of that. But why did you sort of head down this path to sort of take some of the load off the operator with some of this automation?
Jon Schoenfeld: Yeah, that's a, that's a great question. I mean, we've been delivering fault detection and diagnostics for a long time, you know, more or less since, since I've been at, at buildings IFT, you know, the challenge is to get value of that out of that FDD. Somebody needs to do something with that data. Somebody needs to take an action to fix those issues.
And not every, not every operations team, not every energy team has the time and effort to do that. Um, and so we recognize that we can, we can improve the value that can come from the software by automating. Some optimization strategies. And we decided to start with automated demand management because of, we knew [00:14:00] one, that it could provide value to 90 percent of our customers, but it's, it's also just a societal good.
It's something that allows us to relieve the pressure on the grid. In an automated way, that's going to help propel further investment in renewable energy and further decarbonization of the grid. So it seemed like a great place to start.
Ryan Knudson: And that's exactly, you know, what, what John and I talked about at the inception was again, that, you know, making the buildings, um, adaptive to changing grid conditions without becoming so onerous on our, uh, facilities teams.
They have a day job, which is delivering world class service to our tenants and our guests, not to be energy engineers. And so, leveraging that technology to, be complementary to that strategy was so vitally important to us. And I think we've, we've really achieved that here.
James Dice: Yeah. And I'll just add my, my two cents, you know, John demoed [00:15:00] this to me, it was probably last fall, something like that.
Um, John, and it really, it really hits at this missing piece in the market for me because, you know, when I was at the National Renewable Energy Laboratory, the government and the thought leaders from the, from the national labs We're really talking about, Hey, we really need to have great interactive, efficient buildings.
It's a, it's a acronym that you see everywhere nowadays. And what I really told them and really advocated for when I was there was really how we have this missing piece that is software that integrates all the systems together in the buildings, software that can detect when something is wrong physically, and then software that can also take action and You know, implement these new control sequences.
I felt like it was being sort of talked about at the national lab layer level as this sort of like box that you might install. And it was really this software stack that I feel like we were sort of missing. And so I was really glad. And I told Billings IoT [00:16:00] team this, that like really glad to see them sort of moving in that direction because it really feels like a missing piece in the market.
Speaking of that, so Ryan, as a buyer, you know, we've had several buyer interviews with you among many other buyers across the market, really trying to understand how purchasing decisions happen. What, can you talk about the other options you were considering as you sort of went through this and sort of why you ended up choosing the, what, what buildings IoT has built here?
Ryan Knudson: Um, so we actually looked at a couple of different, um, platforms that purported to do something very similar along with Buildings, IOTs, uh, ADM, and we looked at a couple of things. One, how does it interact with our existing infrastructure, right? Is it bolt on? Is it, you know, we have to rip out half of what we built to make this thing work?
You know, how heavy was the implementation? Because, you know, while the, [00:17:00] you know, initial cost from, you know, B2B may be low, my internal resources. Um, have to manage this. Um, so I, we really took into consideration, um, total cost of ownership. How heavy is it to implement? How does it interact with our existing investments?
And how do we support it long term? Once we understood those components, then we got into now what's the energy saving. Um, and there's a, there's a couple, uh, of opportunities for this platform to reach into new areas, which is attractive to, to me, because while, you know, You may say, Hey, you know, you're doing automated demand management, but there's, you know, heating and cooling and there's, you know, set points and everything.
And those are all things that I think we will look at. It's starting small. Again, going back to the concept of frictionless, right? Developing that buy in at the local level, showing [00:18:00] that these, um, this platform, um, is a force multiplier for their day to day operations. And it won't impact guest comfort. Um, and that, and, and giving us the ability and the runway to then add new features to it.
We looked at, All of those things. How heavy to implement, uh, what is the, uh, support, uh, how does it interact with our existing investments, and then how does it save money today, and how can it save money tomorrow. Given, you know, changes to our landscape.
James Dice: Totally. And can you talk about, like, the, um, main phases of deployment here?
You, you, you touched on them a little bit before, but was there, did you do a pilot, and then you did these eight, eight centers that we're talking about, and then it sounds like there's a rollout that's happening after that. Can you talk about that from the perspective of, you're, you're the program manager, right, Ryan?
You're the one that's going up Upstream and [00:19:00] getting funding from the organization. So how did you navigate this from a procurement perspective?
Ryan Knudson: Yeah, absolutely. So it comes down to partnership, uh, first and foremost. So, um, we partnered or I partnered with internal teams on IT and inside of our facility management group, which was.
Um, you know, kind of easy for me because I sat in those seats before, so I know who to talk to. Um, but it, uh, anyone else on the buyer side, those are the people, those are going to be your key state internal stakeholders. It's IT from a technology deployment perspective and facilities management from a day to day operational perspective.
So once that team was formed, Uh, we then went into a pilot, um, and we piloted a property, uh, in Connecticut for a couple of reasons. It had, um, a mixture of different types of mechanical systems, DX, uh, central plants, um, some lighting control. Um, and we really wanted to kind of touch a lot of the different things that this platform would end up being deployed to on.
And we [00:20:00] let that sit for about six months, just to kind of see how would this thing operate in, in a seasonal basis, um, and how it would impact, um, our, Uh, operation, which is why, you know, we did a pilot in 22, but we didn't really start rolling out in earnest until Q3 of 23, because we really wanted to take our time, um, and not have a situation where everyone gets excited, we deploy something new, and then four months later, it's turned off.
So, um, we really gave it time to, to, to sink in. Um, and then the first, uh, initial deployment, uh, encapsulated about. Five centers, I believe. Um, so we started with one center, we then rolled it out to five. And now this year, I believe we're rolling it out to another four or five centers. Um, so it'll take us ultimately at the end of this year, uh, to 11.
And we're hoping, um, next year to roll it out to the final [00:21:00] about 15 centers. That would be a good candidate.
James Dice: All right. So now I want to ask you guys about the. technical components. Um, so for all of the nerds out there that have been like, well, tell us what they actually did this whole time. So let's first talk about the, the supervisory control component for demand management.
That's really the leading, the headliner here in terms of, um, what you guys did. So John, tell us about that, that piece of it.
Jon Schoenfeld: Yeah. Uh, I mean, before we jump into that, I would like to just talk a little bit about the steps to get there. Cause I think that's really important in order to do this. You need to have connectivity.
You know, Ryan mentioned that the need to coordinate with your IT team is key because we're going to have to install a edge to cloud gateway at every single, uh, at every single site, right? And, and getting buy in from the IT team, getting buy in from the facilities team to make that a smooth process is really important.
And the legwork was done up front, uh, with Mace Search to, for that [00:22:00] to happen. So, that's number one. Uh, then it's, it's about once that's there, the connectivity and then the data modeling. That needs to be done, right? So in order for this to work at scale, we're not programming it at every single facility.
We're modeling the data so that the same strategy can be deployed at every single site. Uh, and so, uh, we model all that data. And then we apply the same FDD we've been doing for eight years. Um, and the FDD is a really important component of our strategy to advanced supervisory controls. I would say that we can't provide advanced supervisory controls without the FDD because it tells us what equipment is functioning so that we know what we can and cannot optimize.
Uh, and that's been a real success for us. Not just in improving the optimization strategy, but also in providing that information to [00:23:00] each site and saying, Hey, these are the equipment that are not quite working and that is preventing optimization on these units while these units are functioning properly and the optimization strategy is successful.
Uh, and we've got great buy in from the operations teams in fixing those issues that we do identify. And that's been a real success story of, of this, this program, uh, from, from day one.
James Dice: And real quick, let me make sure I just restate that back to you for people that don't quite, quite grasp what you're saying here is you might have a fault on an air handling unit somewhere that says, um, I.
My VFD is out or I, I, I, some, some sort of component is out. And, um, you're making that decision at the software layer saying basically, okay, if that's out, then I can't pre cool the space so that we can save demand later in the day. So the, the two strategies and when we, when we think about these two [00:24:00] technologies, we sometimes think about them as, you know, vendor A provides FDD and vendor B provides supervisory control and therefore, like we would need to coordinate, but you guys do both of them.
And so what you're saying is you couldn't do the supervisory control piece because you'd have so many false positives or false, um, You have equipment that's not functional that's being enrolled in these control sequences that it's not capable of actually doing.
Ryan Knudson: Yeah, 100%.
Jon Schoenfeld: That, that's exactly it.
James Dice: Yeah.
Ryan Knudson: And, and, and later on when we get into results, James, you know, there's, there's two specific, uh, centers that I'm going to highlight, um, that really speak to that fault detection diagnostics and that deep layer, uh, level of insight that, um, the platform was able to provide. To find things that were frankly invisible to, uh, our operators, even though we had a, you know, pretty sophisticated EMS system.[00:25:00]
Um, so I think it's, it's important to really emphasize. That kind of, uh, portion of this project, um, because a lot of value will be derived just by finding things, um, that you may not be able to see. And it's not even about like, um, you're ignoring an alarm. It's literally things you may not see.
James Dice: Yeah. And this, John, and those conversations reminded me of this newsletter that we'll link to in the show notes.
And it was basically like all the components required to make grid interaction, or in this case, demand management, which is included in grid interaction work, right? Um, And John, if I can just restate it back to you, the components you required here, you're talking about some sort of device that's on site to integrate with the systems.
Then you have a data layer, and in your case, you guys call it IoT Jetstream. Um, IoT Jetstream is taking all of the equipment, all of the data, Um, not only the ability to figure out which points it can override and [00:26:00] command, but also modeling the data to figure out what does all this data mean? Um, what type of system is this?
All of that. And then the ability to then use FDD to figure out which of those components are working properly, and then supervisory control to basically send commands and implement control sequences in order to reach the outcomes that you need to reach. Yep, 100%. Those, those are all the pieces. What about the utility rate piece of this or the, you would need to understand in each, in each, um, region of the country, what's the rate look like?
Or is there time of day pricing? Can you talk about that piece and how you guys sort of implement, implement this sort of incoming data from the utility?
Jon Schoenfeld: Yeah, absolutely. So you talked a bit about pre cooling and that's an important, important aspect of it. Um, so there's a pre cooling aspect and then there's a.
Uh, a load shed algorithm. Um, so we have a load shed algorithm. Um, it's in [00:27:00] its simplest terms, we are making small adjustments to zone set points in a sophisticated manner to try to minimize impact to occupants, but maximize the load shed, and we're doing that on a five minute basis, constantly making small adjustments.
That whole strategy is independent of the initiation and control of how much we're shedding. So decoupling those two things allows us to then have different initiation strategies depending on what the demand is at a facility. So if we have a facility where we've got a good meter, we know that there's high demand charges, Then we're going to set a set point for the demand and we're going to initiate that load shed based on when load gets to a certain point, close enough to that, to that set point or over that set point.
We're going to start dynamically [00:28:00] making those changes. And keep that demand below the set point. In other facilities where there is no demand charge, we just simply set a schedule. And so during a certain part of the day, we're going to slowly ramp up our load shed and try to shed as much load towards the end of the day.
And so having those two things decoupled allows us to essentially. Match the correct approach with the correct rate, uh, schedule and as well as just the, the specifics of the building. Some of the meters are not maybe working at this facility. All right, we're going to use a scheduled approach here. This particular building, it says solar and a very high peak load.
With a very, um, high peak demand rate. Now we're going to maximize that, that set point approach. So that's really been our strategy and it's really made it. Uh, very easy for us to essentially take the same code, the same strategy for [00:29:00] load shed and apply it to eight different facilities with eight different restructures.
I'd say the other thing that it allows us to do is to be flexible for whatever happens in the future. All of these rate schedules are going to change. Demand response, uh, uh, programs are going to change. Having this ability to, to really change how we deploy the load shed without reprogramming, without changing that load shed strategy allows us to be nimble.
And as, as Ryan said earlier, allows us to be future proof.
James Dice: I think I want to switch gears now and talk about, we've talked about the operator a lot. In this conversation so far. So Ryan, can you talk about, um, Okay, the system's installed. It's going to start taking control of equipment. Um, how have you engaged the local operators?
Again, we talked about earlier, you don't want to expect too much from them. We want to automate as much as possible. You want them to be able to worry [00:30:00] about comfort and cleanliness and user experience as paramount to their jobs. How have you sort of done the sort of, what do you expect of them, but then also the change management to make that actually happen?
Ryan Knudson: Yeah, great question. So, you know, we, we approach it from, uh, A crawl, walk, run perspective. So while we're deploying on the launch pad, the facilities teams are integral to that that project and they're seeing everything. And then as John and his team are starting to do the initial modeling. There's a lot of communication with our facility team through the project around you may see that we are touching that during this time, really giving that transparency so that the operator doesn't think, Whoa, everything's out of control.
Um, but really helping them understand why we're touching things and what we're doing, when we're doing it. And then finally, once the models are [00:31:00] done, um, uh, John, Deploys the algorithm, but without the active control. So kind of a what if, and we're able to show during training, once the system is turned on, this is how we arrived at these conclusions.
Here's what we've been doing in the background that you. Didn't even know what's going on. And, you know, kind of being able to show that transparency, um, of the results while also emphasizing the, um, frictionless impact to their day to day operations, I think has been, uh, what has really made the project, uh, so successful and where the adoption is really driven because there hasn't been Why is my thing turned off, or why is my thing turned on, and I didn't tell it to do that.
Also, on the BMS graphics, we have clear annunciation around what is and is not a [00:32:00] part of our ADM screen. So there's never any ambiguity at the operator level about what is going on and when. And that's all process, right? That's independent of people and that's independent of the tool. That's all just a process, uh, challenge that I think is, is vitally important, um, to, to success.
James Dice: Absolutely. All right. Let's talk through the results of this. We've talked about the tech. We've talked a little bit about the change management piece, which we'll talk more about in a little while with the lessons learned we're going to talk about, but just we can, you guys walk through the results and the value that.
That happened here.
Ryan Knudson: Sure. So, um, we're, we're going to talk about three different specific examples. Um, and we pick these examples to really highlight the desperate nature of the value that can be created through this type of a, of a platform. Um, so the first example that we have is at a regional center in New York.
[00:33:00] Um, and going back to what we were discussing earlier. around fault detection diagnostic and why that's so important as a first step. Um, during deployment, uh, John and his team identified 15 RTUs that were running 24 7. And it wasn't because the operators just had it in occupied. No, they were doing the right thing.
And in the BMS, um, they had it set for occupied and unoccupied. But when John did the data modeling and saw the overnight load profile, we've come to find out that some of the facets between the Niagara system and the base building control system were mismatched. So everything on our Niagara system was saying, this thing is going unoccupied, we've sent the command, everything is great.
But it wasn't actually taking in to these 15 RTUs. So, John and his team, through partnership with OTI, the Master System Integrator, were able to go in, modify the sequence, uh, modify the integration, and [00:34:00] shut those RTUs off per the schedule the operator had already set. And that level of, you know, visibility and deep insight, um, really showed a tremendous amount of value.
And just by that one act alone, before we even get into automated demand management, We've already seen, you know, kind of in the first month, a 20 percent reduction in energy consumption. So, um, that's a, that's a tremendous value before we even start talking about automated demand management. Um, the second, uh, case study that I want to talk about is, um, a product.
Another regional, uh, mall in, uh, Southern California. And we implemented this, uh, uh, platform in Q3 of last year. So right before holiday season, uh, but kind of in the shoulder months of, of California. Um, and since then, we've seen a 1 percent reduction in our energy costs. Now you might be saying, wow, great results, but in that [00:35:00] same exact time period, we saw a 15 percent increase in our rate structure.
So this platform was able to help us avoid a 15 percent cost impact to our utilities, which is tremendous for us. Again, going back to, we're not asking our operators to change any behavior. We're using technology. To insulate us from costs without any sort of behavioral changes. So that's all fine and great.
But let's, how does demand management actually work? And what does that actually do? So I saved that one for last. So at a very large multi use, uh, center in Northern Virginia, Um, we're seeing 250 KW. So for context, about 25 to 30 percent of our total HVAC load Being load shed daily without any impact to comfort or any rebound spikes from like a [00:36:00] demand, uh, response type of implementation.
Um, and this has been going on for the, for the last, um, month and a half to two months, um, so during, you know, kind of the shoulder season of Northern Virginia, and we expect to see these kinds of savings continue on into the summer months, uh, because that system is a hybrid of direct exchange and central plant cooling, so we're able to really Leverage all the desperate systems in one unified way.
But this is the power of automated demand management and, and we really picked these three sites for a lot of reasons, right? They're, they're big centers. Um, you know, the first center that I talked about, we've just done a controls retrofit and this was something that was just a simple You know, um, myths, um, that nobody would have seen and we were able to find it through implementation.
The second example really shows that, you know, grid interactivity and rate structure, [00:37:00] um, decoupling that John's talking about to insulate us from, uh, utility rate increases. And then last but not least is the example of automated demand management where a system is. Fully functional, ready to go, and ready to really leverage the full power of this platform.
When it's ready to go, you see these results almost instantaneously.
James Dice: And John, that was amazing. Thank you for taking us through that, Ryan. The, and I love how you presented the three different types of value. That's really cool. The, um, John, the different ways in which you guys reduce demand. So Ryan talked about 250kw at that site.
Can you talk about the ways in which you make that, that happen?
Jon Schoenfeld: Yeah, so that particular site, um, we have a scheduled Demand reduction that occurs every day. Uh, I believe it starts at like three 30 or four o'clock and it runs towards [00:38:00] essentially close to the end of occupancy. Uh, and we, we schedule this reduction for a couple of different reasons.
One, we have some fairly. Large fluctuations in the meter for that building that make it hard to control the, the, uh, load shed by the meter itself. We're also don't have a, a steep spike in utility demands for this particular building. It's a fairly flat load profile, but they're still paying, uh, energy costs.
Those energy costs are highest, you know, at the three o'clock to eight o'clock hour. You know, not to mention this is also the dirtiest time of the grid as the sun is going dead down and the solar production is the lowest, yet there still is peak, peak energy consumption. So it, this is probably my favorite example right now.
I think I'm going to have another favorite within a [00:39:00] couple months, but if you look at the load profile before and after. I mean, it basically looks like somebody took a big bite out of the load profile of this building because it just, at three o'clock, it just drops off, uh, and it does that until essentially until the end of the day.
And then, and then steadily goes down from there. So really, really powerful how you can see it. It, it make real impact to the, to the load profile of the building.
James Dice: Awesome. All right. So then Ryan, what's, what's next on the horizon? Um, for this, what are you excited about doing now that you've sort of proven the value with these initial use cases?
Ryan Knudson: So beyond rolling it
out to
the other centers, you know, kind of enhancing the platform, I'm very excited to, to work with John on integrating our onsite renewable energy assets. Um, and having those assets. Uh, inform, uh, some sort of optimized startup of our HVAC. Um, you know, John and I talk, uh, extensively about, you know, pre [00:40:00] cooling and the prevailing wisdom of that around rate structures.
But, you know, my, from where I sit, I see that there's tremendous value in informing our startup routines around when production of our solar assets are. And allowing those solar assets to absorb the cost. And that is something that I'm, uh, very keen on. Um, again, Uh, as I said at the, at the top, finding new and novel ways of deriving value from existing investments, uh, using data and using technology.
James Dice: Absolutely. I actually haven't heard anyone connect those two concepts before. Basically, I think everyone thinks about starting up things, uh, starting up devices, HVAC, et cetera, before the sun comes out usually, and what you're talking about is in a retail environment, it feels like. Like stores don't open till later, right?
Is that what you're saying? And so the sun's coming out [00:41:00] at a pretty good time to absorb that, that initial, okay, shit, we got to turn the cooling on right now. Absolutely. That's, that's fascinating to me. Absolutely. It's really cool.
Ryan Knudson: And, and I would much rather have my HVAC running at full tilt, uh, when my solar is running at full tilt, than having my HVAC run at 50 percent when my solar is at zero.
James Dice: That's, that's innovative. Um, all right, let's talk about lessons learned. So I said at the beginning, we don't want this to be a fluff story. We don't want to act like everything went perfectly when you guys are trying to innovate and do stuff that hasn't been done very much in buildings before. So, so feed me, feed me with some, some challenges that you guys saw and sort of how you have been overcoming those challenges.
Ryan Knudson: Sure. So I'll start with mine and it's something that John very politely touched on at the, uh, case study in Northern Virginia around metering. And, you know, Matrix has been monitoring our utility meters, um, at the kind of master level for a while, um, and James, as [00:42:00] you said, you and I talked about data quality for ESG reporting, um, a while ago, um, and John really kind of proved out some of my concerns around the quality of the older investments of meters, and, um, So while at the northern Virginia site, we have this scheduled load shed.
It would be great if we had strong meters. So the lesson learned here is focus on your meters, right? And it always goes back to the meters, right? It always goes back to the data quality. But we're now looking at, um, uh, making some investments in some of the older metering infrastructure, um, to improve those so that we can Really start to accelerate, um, their platform in new and, and novel ways and, uh, expanding on the metering, uh, down at some cases to the major asset level, such as central plants, such as large scale, uh, rooftop units, 100 ton [00:43:00] plus, putting some meters on that, but really expanding our meters and really getting better resolution on our consumption data.
James Dice: Cool. And I just have to give our team a shout out right now because we have an active vendor selection for metering right now. Um, and so people that are listening to this, if you're a metering provider, Ryan's interested, but also we, we are, we are interested as well in terms of finding, finding good, low cost, uh, metering solutions.
Um, so. I want to hit this real quick because people might not people, I know we want to move on to the other challenges, but hitting on this one real quick, people might not like grasp the connection between good meter data and what you guys are doing with demand management. So, John, can you talk about why?
You had to change different, you talked about earlier, two different strategies. If you had good meter data, you do this strategy, bad meter data, you do this strategy, what's the function of meter data for you in this solution?
Jon Schoenfeld: For, uh, and I'll use the, the [00:44:00] Connecticut, um, site as an example, that there are, our, our first pilot some two years ago, uh, they had solar on site.
They had high demand charges. They had a very steep peak that they hit, um, and we're paying quite a bit to the utility for, uh, for those, for that steep peak that was happening at around seven o'clock. So in order to mitigate and reduce that peak, we're essentially using that, uh, that meter data, that, that peak or that measured demand as our, um, our process variable.
That's what we're controlling to. And so as soon as that value goes above our set point, We start ramping up our load shed. We control our load shed just the same way you would control a, uh, a valve or a VFD with a, with a PI loop. Um, and so, uh, [00:45:00] we're able to essentially ramp up and down our load shed to maintain a, a, essentially a maximum demand set point, but that only works.
If the meter is reliable, right? And, and we learned, you know, essentially the, the hard way when the meter goes down, well, it's going to stop working essentially. Um, so, uh, you know, from that, you know, lesson learned, that's where we kind of developed the, the scheduled approach, the scheduled strategy. So we can deploy now everywhere.
Uh, and when, when some of these meters get resolved or get, I should say, uh, get replaced. We'll be able to, you know, flip a switch, change the strategy and, uh, and really focus on, uh, on peak demand where it makes sense.
James Dice: Okay. What's our next challenge that you guys ran into?
Jon Schoenfeld: So I'll take this one. We mentioned the, the fault detection diagnostics and the integration of that, uh, into the [00:46:00] algorithm, but we also need those, um, those things to be fixed, right?
We need those issues to, to, to be resolved so that we can maximize the. The opportunity to maximize the optimization. So a couple of things we did to really help that is, uh, one, I think it goes to, to Ryan's team and, and the Maestrich team and really creating, uh, a strong rollout of this where everyone's on board in what we're doing and what we're trying to accomplish.
So it does not feel like the operators do not feel like this is being done. Being, you know, put on them, but they feel engaged and the, the turnover and the training that we've been doing has been a really successful at each of these, uh, at each of these properties, they seem there, the teams are very much engaged and some of the issues that we've identified have been resolved almost immediately.
Um, [00:47:00] and that's really, really powerful to see. Uh, and so having that engagement and then displaying the information in the right way, right? So we are showing each of the operators here that here the equipment that are part of this strategy Here are the ones that are disabled because maybe they serve a tenant space Maybe they serve an area that you don't want to be part of demand management and here are the ones that are enabled But are disqualified because of some FDD.
So there's already this kind of almost gamification of, all right, how can I get this slice of disqualified units to be as small as possible so we can maximize. Uh, the, the efficiency and maximize the, uh, you know, the return on investment.
Ryan Knudson: So the last challenge, um, and, and we kind of touched on a little bit, um, it's just the disparity in rate structures, the disparity in utility charges, and making sure that the deployment was flexible enough to.[00:48:00]
Meet the unique challenges. Um, as John said, you know, we may have a property that has a TOU and that TOU is, you know, 4 to 6 PM. So that's when we really want to target it. So not being, um, so, uh, uh, broad. In our algorithms, but really kind of tweaking and tuning the base concept for not only the physical, um, makeup of the building, but also the maximum impact of the platform.
And so that took a little, little bit of time and a little, a little bit of, uh, analysis and introspection, um, because. As John will tell you, I always want to go much faster than we end up going. Uh, and he very patiently explains to me how, uh, property A's value proposition cannot be directly translated to property B, uh, because he needs, he and his team need to do the modeling and take into account all of the value [00:49:00] variables.
Um, and, uh, it seemed it more often than not, he is correct and I am just patient.
James Dice: John, can you talk about, you guys do a lot of, um, data modeling leadership, I would call it in the industry, really sort of, um, Trying to take a lead and own and share what you're doing on the data modeling side so that other people can sort of learn from it, adopt it, etc.
Can you talk about some of the modeling challenges? I'm just assuming there are modeling challenges around modeling these different utility rates, um, and including those in the control sequences.
Jon Schoenfeld: Yeah, I mean, I, I think it's a, it's, it's a great question. Um, through the process of developing ADM, we've had to standardize on new data points and new, you know, these new control strategies need ways to be controlled.
And so with our, uh, you know, what you're referring to the, primarily as the ontology alignment [00:50:00] project, we've essentially created these new, these new points, these new attributes. These new parts of our data model, and then we, we publish them, right? So we don't do them in a black box so that only us, only we can use them.
We try to, um, generalize them as much as possible, uh, so that they could be used by anybody. So, you know, if you look right now in, uh, in the OAP, you'll see things like. Loadshed command is, you know, a new point that's available. Loadshed sensor is a new point that's available. Uh, automated demand management enabled is a new point that's available.
These are all part of now the data model, which then makes them available to the end user, to the operators, so they can see how is Loadshed behaving. How do I disable this unit as opposed to this unit? How do I make a change to the maximum set [00:51:00] point that this is going to go, you know, that this, this set point is going to be raised to.
Demand management, the way we do it is probably different than the way others do it, but some of these core concepts are the same. Um, and so there's use of those points regardless of. You know, the, the exact strategy that you deploy.
James Dice: So, so Ryan, can you talk about simply the, the paying for it? The question here, how did you get business approval for this?
Like what's the ROI, uh, calculation look like and, and sort of how do you make the business case?
Ryan Knudson: Sure. So, um, you know, the business case was really built, uh, again, upon, um, looking at, uh, the investments that we had made over the years. And, um, the executives really came to me and said, Hey, rate pressures are so extreme.
How can we leverage all this technology that we've been investing in for years, um, to insulate us from these rate pressures? Or is there anything that we can do? So it [00:52:00] wasn't so much that, um, you know, I had to, So called sell up. It was a lot of, you know, kind of top down pressure around, you know, the last 24 to 30 months of weight escalations and utility costs, um, you know, increase, um, and not throwing more large capital into, you know, onsite generation, but leveraging everything that we've done to date.
And, yeah, and, um, that's when I went to, to John and to Clint at OTI and said, uh, you know, Hey, gents, you know, we've been working together for eight years. We need to come up with something. You know, what, what do we have? What have we been talking about? And are we ready for it? Um, and, uh, I go back to kind of that, that lag between the pilot and full deployment.
That's where we really developed those models. And we really took a hard look, um, at, uh, [00:53:00] that first pilot site. Uh, and I think John had to run the savings model about three or four different times at different points, uh, throughout the, the kind of almost year gap. To really prove out that this, uh, would be frictionless to our operators, um, and would provide values throughout the year and not just, Hey, during the shoulder months, you're going to make all of your savings, but in the summer, it's going to shut off.
And, um, you know, we'll hope we'll pick it up in September. Um, and that's what we built our savings models on. Um, and then we just deployed that saving model. To each center based on size of the property type of system on that goes back to being very strategic and selecting your pilot site, making sure it has a good cross section of the types of systems and buildings that you're going to end up deploying to.
because we were able to then. Use that model across the rest of the portfolio. And, um, today we've, we've seen better [00:54:00] than what we had estimated. Um, frankly, um, because of some of the, the wins that we talked about in the results, uh, that we just weren't expecting to, to find. Fascinating.
James Dice: All right. Last couple of questions here as we, as we sort of close out, these are sort of conclusion sort of questions.
Um, Ryan, if you're sitting down with, you know, you have a peer group that you meet with regularly, other, other sort of people in your, in your position and similar portfolios, what would you tell them in terms of like, here's the playbook to sort of copy this approach and sort of replicate the results we've had?
Is there a five step playbook you could tell them? What would you tell them?
Ryan Knudson: Yeah, I
think step one is to, uh, communicate internally with the Uh, your, your stakeholders, you know, partnering with your I. T. team, making sure that they are ready to onboard, um, such a, uh, this kind of technology and what their privacy and cyber security requirements are, uh, then [00:55:00] working with facility management to, uh, help understand what centers, um, are, uh, Uh, positioned to, you know, deploy.
Um, if we're about to have a major redevelopment, probably not a good time to deploy, you know, advanced algorithms. So, really understanding, um, all of the internal, um, rules and requirements, that's step one. And then step two is, uh, to look at your rate structures. Understand where you think the biggest value is going to be derived, um, and, uh, making sure that you are best positioned to achieve those.
Uh, and then step three, um, is site selection. Um, and really diving into the assets, uh, and working with those local teams to understand what their challenges and their, uh, needs are so that we're not, uh, deploying, um, you know, space age technology on [00:56:00] Flintstone's infrastructure, um, and helping them get to a place to be ready for it.
Uh, and then step four is, you know, deployment. Deployment, active management of your deployment, being engaged, being transparent, making sure that all of those stakeholders that you talk through, step one, step three, are fully informed, um, at every step of the way. And then last but not least, the step five, monitoring and management, going back and doing the M& V, showing the results to all of those stakeholders, right?
Um, and, uh, Building the case for the next group of properties. Building the model for the next deployment, uh, group.
James Dice: Space age technology on Flintstones infrastructure. I'm gonna steal that 100%. Please. I'm definitely gonna steal that. Um, alright, my last question for you guys is, um, What haven't we covered here that you would, that you would want to make sure we covered?
And I'm going to start, I'm going [00:57:00] to start with this. Um, I just want to say thank you, Ryan, because it is really hard to find someone on the buyer side, someone on the building owner side that is willing to talk to, The level of detail in their projects, like you're doing right now, it's even harder to find someone willing to talk to about it at all, but we'll set that aside, I think in order for our industry to transform and accelerate the way that we at Nexus want it to happen.
And that's our mission, right? That's why we're here. We need more people like you, Ryan, that are willing to talk about your projects, and so thank you for doing that. If there are any other people out there like Ryan that want to talk about their projects, either good or bad, we're here. That's what we're here for.
With that, what do you guys think we haven't covered yet that you want to tell someone that's sort of just starting out and doing this? I would just say, um,
Ryan Knudson: Be patient with yourself, be patient with your operators, find a good, uh, core team, um, both internally and [00:58:00] externally. Uh, find, uh, good, uh, scalable, uh, master system integrators and, and systems integrators.
That understand your vision, are on board with your vision of, um, whatever you're trying to achieve. Um, and mostly, um, be ready for unpredictable results. Uh, and Being flexible enough to understand the value of unpredictable results.
Jon Schoenfeld: I couldn't agree more, uh, you know, the, the, the phrase that I keep coming back to is success doesn't run in a straight line.
Uh, and this isn't with automated demand management or advanced supervisory controls. This isn't something that you just pop in and, and walk away and everything works hunky dory. Um, there's, there's real work to be done. And some of [00:59:00] that success comes from that work. And, and honestly, a lot of the fun comes from that work, diving into the details to understand the root cause of that issue so you can solve that issue very quickly and easily and save, save a bunch of energy, save a bunch of money so that the optimization that works from there is.
It's part of the fun of the project and it's part of what makes it impactful. Uh, the last thing I would say that, that, you know, one more kind of, I guess, key thing we haven't talked to, I don't know enough about is the, the openness and transparency of the solution and of the data model, right? At the end of the day.
Um, we're serving MaceReach with a, a transparent and open system where all of their data is available via an independent data layer. And all of it is modeled via an open, [01:00:00] uh, Haystack compliant data model and, and as well as compliant with other data models with the, uh, with the ontology alignment project.
That means that if tomorrow MaceReach finds Some other, you know, strategy related or not related to bolt on to their existing system. They can do it. And they don't have to pay a vendor to recreate the integration, to recreate the data model, it's already done. Uh, and so that, that just makes value for the future that it doesn't go away.
James Dice: Thank you both for sharing this. This is so cool. And I'm excited to hear more about as you, as you scale up.
Rosy Khalife: Okay friends, thank you for listening to this episode. As we continue to grow our global community of change makers, we need your help. For the next couple of months, we're challenging our listeners to share a link to their favorite Nexus episode on LinkedIn with a short post about why you [01:01:00] listen. It would really, really help us out.
Make sure to tag us in the post so we can see it. Have a good one.
"The platform of automated demand management is important for Macerich. The rate pressures that we have seen from utilities have been extraordinary. It’s not just about reducing, but being smarter and more strategic with those reductions. This platform gives us the flexibility and the future proofing that we think we’ll need for the next 5 years.”
—Ryan Knudson
Welcome to Nexus, a newsletter and podcast for smart people applying smart building technology—hosted by James Dice. If you’re new to Nexus, you might want to start here.
The Nexus podcast (Apple | Spotify | YouTube | Other apps) is our chance to explore and learn with the brightest in our industry—together. The project is directly funded by listeners like you who have joined the Nexus Pro membership community.
You can join Nexus Pro to get a weekly-ish deep dive, access to the Nexus Vendor Landscape, and invites to exclusive events with a community of smart buildings nerds.
Sign-up (or refer a friend!) to the Nexus Newsletter.
Learn more about the Smart Building Strategist Course and the Nexus Courses Platform.
Check out the Nexus Labs Marketplace.
Learn more about Nexus Partnership Opportunities.
Episode 162 is a conversation with Ryan Knudson from Macerich and Jon Schoenfeld from Buildings IOT.
Episode 162 features Ryan Knudson from Macerich and Jon Schoenfeld from Buildings IOT and is our 9th episode in the Case Study series looking at real-life, large-scale deployments of smart building technologies. These are not marketing fluff stories, these are lessons from leaders that others can put into use in their smart buildings programs. This conversation explores how Macerich has incorporated Buildings IOT technology into their shopping centers. Enjoy!
You can find Ryan and Jon on LinkedIn.
Introduction (1:18)
Introduction to Ryan (2:18)
What is the project? (3:57)
Vendor team (5:03)
Project details (5:37)
Project results (7:04)
Introduction to Jon (7:40)
Original goals (9:03)
Requirements (11:45)
Other options (16:28)
Phases of operation (19:05)
Supervisory control (21:16)
Utility data (26:45)
Operator engagement (29:54)
Results (32:36)
Lessons learned (41:24)
The business case (51:32)
What would you tell others? (54:28)
Music credits: There Is A Reality by Common Tiger—licensed under an Music Vine Limited Pro Standard License ID: S569626-16073.
Note: transcript was created using an imperfect machine learning tool and lightly edited by a human (so you can get the gist). Please forgive errors!
Ryan Knudson: [00:00:00] The platform of automated demand management is important for MaceRich for a couple of reasons. One, the rate pressures that we have seen from utilities over the last 24 months have been extraordinary. And it's not just about reducing, but being smarter and more strategic about those reductions. And this platform gives us the flexibility and the future proofing that we think we'll need for the next five years.
James Dice: Hey friends, if you like the Nexus podcast, the best way to continue the learning is to join our community. There are three ways to do that. First, you can join the Nexus Pro Membership. It's our global community of smart billing professionals. We have monthly events, paywall deep dive content, and a private chat room, and it's just 35 a month.
Second, you can upgrade from the Pro Membership to our courses offering. It's headlined by our flagship course, the Smart Building Strategist, and we're building a catalog of courses [00:01:00] taught by world leading experts on each topic under the Smart Buildings umbrella. Third and finally, our marketplace is how we connect leading vendors with buyers looking for their solutions.
The links are below in the show notes, and now let's go on to the podcast.
Welcome to the Nexus podcast. This is the latest episode in our series diving into case studies of real life, large scale deployments of smart building technologies. I think this is our 10th. So we're getting into double digits. Uh, we've been doing this for a year and it was really a, our attempt to fill the void in real life case studies and real life stories that we've seen in our industry.
Um, And so we're not here to like create some sort of fluff story. We're here to really, really share the lessons from leaders so other people can put these lessons to use in their smart buildings programs. So today we have a story coming from Mace Rich, which is a leading owner, operator, and developer of top real, uh, retail and mixed [00:02:00] use destinations and major U S markets.
So if you're like me and you're in from Colorado, we have several, uh, shopping malls, shopping destinations here. That are Mace Rich properties. There's probably others in your areas, if you're listening to this or other, other cities in the U S as well. So the voice you heard at the beginning was Ryan Knutson, the vice president, uh, corporate responsibility and sustainability at Mace Rich.
Ryan, can you talk about, uh, introduce yourself and talk about your background a little bit so people can get to know you?
Ryan Knudson: Thanks, James. Uh, as you said, I'm Ryan Knudson. I'm VP of Corporate Responsibility and Sustainability here at Mace Ridge. Uh, I've been here for about 13 years. My career here at Mace Ridge has been unique.
I started, um, in our IT department as The building automation technical expert and really helping Mace Ridge in their first iteration of energy management and sustainability. And I've had the unique opportunity as Mace Ridge has grown its sustainability program [00:03:00] to grow along with it, um, touching different parts of the business from IT and, and, uh, building automation to operations, maintenance and facilities management.
Um, and then, uh, coming. Finally, here to the sustainability team
James Dice: and Ryan, you were on the, on the podcast before it was episode 94, if people want to go back and, and Ryan talks about why data is important in his role in sustainability and, uh, why data is important for the ESG journey as well. So that episode is episode 94.
We'll put a link to that in the show notes. Um, and this conversation, we're kind of diving into this one project that you guys have done. As part of the overall smart buildings program and sustainability program, Ryan. So let's do a couple sort of rapid fire questions to give people context on the project we're talking about.
So, um, What are we talking about here? How do you guys refer to it in house? What's this project called?
Ryan Knudson: So we refer to it in [00:04:00] house as EMS 2. 5 because we are not that clever. But no, it is a continuation of our building management program that we've been doing for 15 years. Started with Basic Tritium Niagara integration, unlocking the buildings, and then moving into standardizing our base building infrastructure with BACnets and, uh, diving into the sequencing of operations.
Um, and now we are at the point of deploying, uh, false detection diagnostics, um, and automated demand management on top of those, uh, investments. We very much view this as a continuation and, uh, a way of deriving new value.
James Dice: Totally. I'm glad you shared the 2. 5 tidbit because it really, it really shows the, the, the, the engineer's ability to name things.
It's [00:05:00] funny, funny to me. Um, so who's your vendor team?
Ryan Knudson: So our project team was made up of, uh, in house, uh, developers. Energy management experts with Mace Rich along with OTI acting as our master systems integrators and Buildings IoT as the platforming and engineering team. Additionally, we had our on staff maintenance team.
Uh, M Core teams acting as the key, uh, doers, uh, of the, of the software.
James Dice: And how many buildings are we talking about here where you've, uh, installed this technology stack?
Ryan Knudson: Sure. So to date, um, and it's, uh, April 25th, 2024 to date, we've, uh, deployed it in eight centers. Um, but we are actively working on a 2024 pipeline to grow that.
Um, with an eye to deploying it to, um, all of our major assets, uh, by the end of 25.
James Dice: And what's the total square footage we're talking about here? [00:06:00]
Ryan Knudson: Yeah, so of those eight centers, we're talking about three of our largest centers. So those eight centers really encompass, uh, almost 10 million square feet of, uh, combined, uh, tenant and common area space.
Because some of these, uh, systems do interface with our tenant systems.
James Dice: . And when did this project start? I mean, I know you talked about the controls upgrades and Niagara and all that stuff, but when did this piece that we're going to talk about today start?
Ryan Knudson: Sure. So we actually started piloting it with Buildings IoT in kind of Q3 of 2022.
Um, but we didn't really decide to move in earnest until Q2 of 23 and really started installing in earnest in Q3 of 23. So we've had eight centers over the last two and a half quarters.
James Dice: And we're going to talk about the like deep dive into the results in a little bit. But can you just like give us big picture?
What results have you seen that make [00:07:00] you kind of want to expand this from the projects you've done so far into the rest of the portfolio?
Ryan Knudson: Sure, so our results vary, uh, as you can imagine based on our building type and the, um, mechanical systems, uh, at each center, but generally speaking, we have seen, um, somewhere in the neighborhood of, uh, close to 100, 000 annually, we're projecting in, um, utility savings, um, which will be kind of carved out in different ways, um, which we'll dive into, um, a little bit later.
James Dice: So we also have, um, so John Schoenfeld here, VP of Energy and Building Technology at Buildings IoT, um, to kind of represent Buildings IoT in this, um, discussion here. Uh, welcome, John. Can you introduce yourself?
Jon Schoenfeld: Thanks, James. Glad to be here. Take care. Uh, name's John Schoenfeld. Uh, I've been with Buildings IOT for, uh, going on eight years.
Uh, prior to that I was a principal at KW [00:08:00] Engineering, where I focused on retro commissioning and energy engineering of large portfolio, uh, clients. Been really a, a great ride here at Buildings iot. I've watched the, the company really transform itself from, uh, a master systems integrator, uh, to a platform developer.
Developing products like an independent data layer, a smart buildings platform, and now getting into advanced supervisory controls with automated demand management. You know, it's been a great eight years and we've had the pleasure of working with Ryan and the Maestrich team for that entire time, and I've really had a lot of success in, you know, taking a collaborative approach to Uh, to the platform development, but also the energy reduction.
James Dice: Totally. All right, let's dive into this case study. So Ryan, can you talk about a little bit, you talked about it a little bit earlier, but why did you guys start down this [00:09:00] path and what were the original goals for, for this program?
Ryan Knudson: Sure. So, um, you know, we were really looking at all of the investments that we've made over the last 10 to 15 years.
And, and John's been a part of that for, for eight, um, in our control system. And how can we. I'm a researcher and I'm trying to leverage all of that technology and all that compute power in new and novel ways to derive new values. One of the key outcomes that I was hoping to achieve was to find a frictionless way of automating our, uh, demand management and controls, um, without the need of human intervention, because as I'm sure everyone can tell you, you can put in the best control system, but if the operators treat it like a light switch, you're not going to get the value.
And I'm not saying that's what was happening with us, but I was looking for something that we could use that wouldn't [00:10:00] require our operators to change their behavior.
James Dice: Totally. Can you, um, I want to circle back on that, but for just a second, but for those of you, the, the, the audience that doesn't know about Mace Rich's bigger goals around net zero, um, really around like, why is the business doing this?
Can you talk about like the bigger picture outcomes real quick?
Ryan Knudson: So Mace Rich has a couple of, uh, stated goals. First is, uh, carbon neutrality by 2030, but then extending beyond that to operational net zero by 2035. Um, and that's going to require, you know, reducing our energy consumption and our, and our scope to emission by 50 to 60%.
Um, and a lot of that can be driven by hard energy assets, such as renewables, but it also requires smart technology inside of the building to heat and cool the assets in a, uh, appropriate fashion for the comfort of our guests. But not, um, overly [00:11:00] so that we're becoming wasteful. Um, we also look for technologies that allow us to become, uh, adaptive to the new, you know, grid, uh, conditions.
As these utilities start to transition. Their energy supply away from fossil fuels and into renewables and, um, you know, various different types of new technologies, our buildings are going to need to be able to adapt to those new, um, supply conditions.
James Dice: So can you talk about like, What the technology stack was like before this project, and specifically relating to what you just said a second ago, what did it require the building operators and the maintenance teams to be doing in order to beat those goals?
It sounds like it needed to be, um, the control systems needed a lot of hand holding to get, you know, What you wanted out of them.
Ryan Knudson: Absolutely. I think, you know, where we were in our journey, um, before we started, uh, with John and his team was, um, we had a, a [00:12:00] pretty standard control system, you know, back net system doing all of the, if, if this, then that logic, uh, rolling up to a Niagara, uh, Jace, uh, and then that was being piped into, um, a supervisory controller, uh, And it was all operating with kind of mixed results, depending on, you know, how sophisticated our operators were.
We had some centers, and we still do have some centers, that are deriving a lot of value from that advanced control system. And then we have some centers that have a more kind of set it and forget it mentality. Um, just based on, you know, their everyday workload. So we were really looking for something that would, um, provide value to all of our, uh, facilities teams, regardless of their sophistication.
James Dice: Totally. And, and John, as you guys think about your, your entire customer base, uh, buildings, IOT, can you talk about kind of what motivated [00:13:00] you to, um, Go into this, this sort of, um, we'll get to it in a second, like what we mean by demand management and all of that. But why did you sort of head down this path to sort of take some of the load off the operator with some of this automation?
Jon Schoenfeld: Yeah, that's a, that's a great question. I mean, we've been delivering fault detection and diagnostics for a long time, you know, more or less since, since I've been at, at buildings IFT, you know, the challenge is to get value of that out of that FDD. Somebody needs to do something with that data. Somebody needs to take an action to fix those issues.
And not every, not every operations team, not every energy team has the time and effort to do that. Um, and so we recognize that we can, we can improve the value that can come from the software by automating. Some optimization strategies. And we decided to start with automated demand management because of, we knew [00:14:00] one, that it could provide value to 90 percent of our customers, but it's, it's also just a societal good.
It's something that allows us to relieve the pressure on the grid. In an automated way, that's going to help propel further investment in renewable energy and further decarbonization of the grid. So it seemed like a great place to start.
Ryan Knudson: And that's exactly, you know, what, what John and I talked about at the inception was again, that, you know, making the buildings, um, adaptive to changing grid conditions without becoming so onerous on our, uh, facilities teams.
They have a day job, which is delivering world class service to our tenants and our guests, not to be energy engineers. And so, leveraging that technology to, be complementary to that strategy was so vitally important to us. And I think we've, we've really achieved that here.
James Dice: Yeah. And I'll just add my, my two cents, you know, John demoed [00:15:00] this to me, it was probably last fall, something like that.
Um, John, and it really, it really hits at this missing piece in the market for me because, you know, when I was at the National Renewable Energy Laboratory, the government and the thought leaders from the, from the national labs We're really talking about, Hey, we really need to have great interactive, efficient buildings.
It's a, it's a acronym that you see everywhere nowadays. And what I really told them and really advocated for when I was there was really how we have this missing piece that is software that integrates all the systems together in the buildings, software that can detect when something is wrong physically, and then software that can also take action and You know, implement these new control sequences.
I felt like it was being sort of talked about at the national lab layer level as this sort of like box that you might install. And it was really this software stack that I feel like we were sort of missing. And so I was really glad. And I told Billings IoT [00:16:00] team this, that like really glad to see them sort of moving in that direction because it really feels like a missing piece in the market.
Speaking of that, so Ryan, as a buyer, you know, we've had several buyer interviews with you among many other buyers across the market, really trying to understand how purchasing decisions happen. What, can you talk about the other options you were considering as you sort of went through this and sort of why you ended up choosing the, what, what buildings IoT has built here?
Ryan Knudson: Um, so we actually looked at a couple of different, um, platforms that purported to do something very similar along with Buildings, IOTs, uh, ADM, and we looked at a couple of things. One, how does it interact with our existing infrastructure, right? Is it bolt on? Is it, you know, we have to rip out half of what we built to make this thing work?
You know, how heavy was the implementation? Because, you know, while the, [00:17:00] you know, initial cost from, you know, B2B may be low, my internal resources. Um, have to manage this. Um, so I, we really took into consideration, um, total cost of ownership. How heavy is it to implement? How does it interact with our existing investments?
And how do we support it long term? Once we understood those components, then we got into now what's the energy saving. Um, and there's a, there's a couple, uh, of opportunities for this platform to reach into new areas, which is attractive to, to me, because while, you know, You may say, Hey, you know, you're doing automated demand management, but there's, you know, heating and cooling and there's, you know, set points and everything.
And those are all things that I think we will look at. It's starting small. Again, going back to the concept of frictionless, right? Developing that buy in at the local level, showing [00:18:00] that these, um, this platform, um, is a force multiplier for their day to day operations. And it won't impact guest comfort. Um, and that, and, and giving us the ability and the runway to then add new features to it.
We looked at, All of those things. How heavy to implement, uh, what is the, uh, support, uh, how does it interact with our existing investments, and then how does it save money today, and how can it save money tomorrow. Given, you know, changes to our landscape.
James Dice: Totally. And can you talk about, like, the, um, main phases of deployment here?
You, you, you touched on them a little bit before, but was there, did you do a pilot, and then you did these eight, eight centers that we're talking about, and then it sounds like there's a rollout that's happening after that. Can you talk about that from the perspective of, you're, you're the program manager, right, Ryan?
You're the one that's going up Upstream and [00:19:00] getting funding from the organization. So how did you navigate this from a procurement perspective?
Ryan Knudson: Yeah, absolutely. So it comes down to partnership, uh, first and foremost. So, um, we partnered or I partnered with internal teams on IT and inside of our facility management group, which was.
Um, you know, kind of easy for me because I sat in those seats before, so I know who to talk to. Um, but it, uh, anyone else on the buyer side, those are the people, those are going to be your key state internal stakeholders. It's IT from a technology deployment perspective and facilities management from a day to day operational perspective.
So once that team was formed, Uh, we then went into a pilot, um, and we piloted a property, uh, in Connecticut for a couple of reasons. It had, um, a mixture of different types of mechanical systems, DX, uh, central plants, um, some lighting control. Um, and we really wanted to kind of touch a lot of the different things that this platform would end up being deployed to on.
And we [00:20:00] let that sit for about six months, just to kind of see how would this thing operate in, in a seasonal basis, um, and how it would impact, um, our, Uh, operation, which is why, you know, we did a pilot in 22, but we didn't really start rolling out in earnest until Q3 of 23, because we really wanted to take our time, um, and not have a situation where everyone gets excited, we deploy something new, and then four months later, it's turned off.
So, um, we really gave it time to, to, to sink in. Um, and then the first, uh, initial deployment, uh, encapsulated about. Five centers, I believe. Um, so we started with one center, we then rolled it out to five. And now this year, I believe we're rolling it out to another four or five centers. Um, so it'll take us ultimately at the end of this year, uh, to 11.
And we're hoping, um, next year to roll it out to the final [00:21:00] about 15 centers. That would be a good candidate.
James Dice: All right. So now I want to ask you guys about the. technical components. Um, so for all of the nerds out there that have been like, well, tell us what they actually did this whole time. So let's first talk about the, the supervisory control component for demand management.
That's really the leading, the headliner here in terms of, um, what you guys did. So John, tell us about that, that piece of it.
Jon Schoenfeld: Yeah. Uh, I mean, before we jump into that, I would like to just talk a little bit about the steps to get there. Cause I think that's really important in order to do this. You need to have connectivity.
You know, Ryan mentioned that the need to coordinate with your IT team is key because we're going to have to install a edge to cloud gateway at every single, uh, at every single site, right? And, and getting buy in from the IT team, getting buy in from the facilities team to make that a smooth process is really important.
And the legwork was done up front, uh, with Mace Search to, for that [00:22:00] to happen. So, that's number one. Uh, then it's, it's about once that's there, the connectivity and then the data modeling. That needs to be done, right? So in order for this to work at scale, we're not programming it at every single facility.
We're modeling the data so that the same strategy can be deployed at every single site. Uh, and so, uh, we model all that data. And then we apply the same FDD we've been doing for eight years. Um, and the FDD is a really important component of our strategy to advanced supervisory controls. I would say that we can't provide advanced supervisory controls without the FDD because it tells us what equipment is functioning so that we know what we can and cannot optimize.
Uh, and that's been a real success for us. Not just in improving the optimization strategy, but also in providing that information to [00:23:00] each site and saying, Hey, these are the equipment that are not quite working and that is preventing optimization on these units while these units are functioning properly and the optimization strategy is successful.
Uh, and we've got great buy in from the operations teams in fixing those issues that we do identify. And that's been a real success story of, of this, this program, uh, from, from day one.
James Dice: And real quick, let me make sure I just restate that back to you for people that don't quite, quite grasp what you're saying here is you might have a fault on an air handling unit somewhere that says, um, I.
My VFD is out or I, I, I, some, some sort of component is out. And, um, you're making that decision at the software layer saying basically, okay, if that's out, then I can't pre cool the space so that we can save demand later in the day. So the, the two strategies and when we, when we think about these two [00:24:00] technologies, we sometimes think about them as, you know, vendor A provides FDD and vendor B provides supervisory control and therefore, like we would need to coordinate, but you guys do both of them.
And so what you're saying is you couldn't do the supervisory control piece because you'd have so many false positives or false, um, You have equipment that's not functional that's being enrolled in these control sequences that it's not capable of actually doing.
Ryan Knudson: Yeah, 100%.
Jon Schoenfeld: That, that's exactly it.
James Dice: Yeah.
Ryan Knudson: And, and, and later on when we get into results, James, you know, there's, there's two specific, uh, centers that I'm going to highlight, um, that really speak to that fault detection diagnostics and that deep layer, uh, level of insight that, um, the platform was able to provide. To find things that were frankly invisible to, uh, our operators, even though we had a, you know, pretty sophisticated EMS system.[00:25:00]
Um, so I think it's, it's important to really emphasize. That kind of, uh, portion of this project, um, because a lot of value will be derived just by finding things, um, that you may not be able to see. And it's not even about like, um, you're ignoring an alarm. It's literally things you may not see.
James Dice: Yeah. And this, John, and those conversations reminded me of this newsletter that we'll link to in the show notes.
And it was basically like all the components required to make grid interaction, or in this case, demand management, which is included in grid interaction work, right? Um, And John, if I can just restate it back to you, the components you required here, you're talking about some sort of device that's on site to integrate with the systems.
Then you have a data layer, and in your case, you guys call it IoT Jetstream. Um, IoT Jetstream is taking all of the equipment, all of the data, Um, not only the ability to figure out which points it can override and [00:26:00] command, but also modeling the data to figure out what does all this data mean? Um, what type of system is this?
All of that. And then the ability to then use FDD to figure out which of those components are working properly, and then supervisory control to basically send commands and implement control sequences in order to reach the outcomes that you need to reach. Yep, 100%. Those, those are all the pieces. What about the utility rate piece of this or the, you would need to understand in each, in each, um, region of the country, what's the rate look like?
Or is there time of day pricing? Can you talk about that piece and how you guys sort of implement, implement this sort of incoming data from the utility?
Jon Schoenfeld: Yeah, absolutely. So you talked a bit about pre cooling and that's an important, important aspect of it. Um, so there's a pre cooling aspect and then there's a.
Uh, a load shed algorithm. Um, so we have a load shed algorithm. Um, it's in [00:27:00] its simplest terms, we are making small adjustments to zone set points in a sophisticated manner to try to minimize impact to occupants, but maximize the load shed, and we're doing that on a five minute basis, constantly making small adjustments.
That whole strategy is independent of the initiation and control of how much we're shedding. So decoupling those two things allows us to then have different initiation strategies depending on what the demand is at a facility. So if we have a facility where we've got a good meter, we know that there's high demand charges, Then we're going to set a set point for the demand and we're going to initiate that load shed based on when load gets to a certain point, close enough to that, to that set point or over that set point.
We're going to start dynamically [00:28:00] making those changes. And keep that demand below the set point. In other facilities where there is no demand charge, we just simply set a schedule. And so during a certain part of the day, we're going to slowly ramp up our load shed and try to shed as much load towards the end of the day.
And so having those two things decoupled allows us to essentially. Match the correct approach with the correct rate, uh, schedule and as well as just the, the specifics of the building. Some of the meters are not maybe working at this facility. All right, we're going to use a scheduled approach here. This particular building, it says solar and a very high peak load.
With a very, um, high peak demand rate. Now we're going to maximize that, that set point approach. So that's really been our strategy and it's really made it. Uh, very easy for us to essentially take the same code, the same strategy for [00:29:00] load shed and apply it to eight different facilities with eight different restructures.
I'd say the other thing that it allows us to do is to be flexible for whatever happens in the future. All of these rate schedules are going to change. Demand response, uh, uh, programs are going to change. Having this ability to, to really change how we deploy the load shed without reprogramming, without changing that load shed strategy allows us to be nimble.
And as, as Ryan said earlier, allows us to be future proof.
James Dice: I think I want to switch gears now and talk about, we've talked about the operator a lot. In this conversation so far. So Ryan, can you talk about, um, Okay, the system's installed. It's going to start taking control of equipment. Um, how have you engaged the local operators?
Again, we talked about earlier, you don't want to expect too much from them. We want to automate as much as possible. You want them to be able to worry [00:30:00] about comfort and cleanliness and user experience as paramount to their jobs. How have you sort of done the sort of, what do you expect of them, but then also the change management to make that actually happen?
Ryan Knudson: Yeah, great question. So, you know, we, we approach it from, uh, A crawl, walk, run perspective. So while we're deploying on the launch pad, the facilities teams are integral to that that project and they're seeing everything. And then as John and his team are starting to do the initial modeling. There's a lot of communication with our facility team through the project around you may see that we are touching that during this time, really giving that transparency so that the operator doesn't think, Whoa, everything's out of control.
Um, but really helping them understand why we're touching things and what we're doing, when we're doing it. And then finally, once the models are [00:31:00] done, um, uh, John, Deploys the algorithm, but without the active control. So kind of a what if, and we're able to show during training, once the system is turned on, this is how we arrived at these conclusions.
Here's what we've been doing in the background that you. Didn't even know what's going on. And, you know, kind of being able to show that transparency, um, of the results while also emphasizing the, um, frictionless impact to their day to day operations, I think has been, uh, what has really made the project, uh, so successful and where the adoption is really driven because there hasn't been Why is my thing turned off, or why is my thing turned on, and I didn't tell it to do that.
Also, on the BMS graphics, we have clear annunciation around what is and is not a [00:32:00] part of our ADM screen. So there's never any ambiguity at the operator level about what is going on and when. And that's all process, right? That's independent of people and that's independent of the tool. That's all just a process, uh, challenge that I think is, is vitally important, um, to, to success.
James Dice: Absolutely. All right. Let's talk through the results of this. We've talked about the tech. We've talked a little bit about the change management piece, which we'll talk more about in a little while with the lessons learned we're going to talk about, but just we can, you guys walk through the results and the value that.
That happened here.
Ryan Knudson: Sure. So, um, we're, we're going to talk about three different specific examples. Um, and we pick these examples to really highlight the desperate nature of the value that can be created through this type of a, of a platform. Um, so the first example that we have is at a regional center in New York.
[00:33:00] Um, and going back to what we were discussing earlier. around fault detection diagnostic and why that's so important as a first step. Um, during deployment, uh, John and his team identified 15 RTUs that were running 24 7. And it wasn't because the operators just had it in occupied. No, they were doing the right thing.
And in the BMS, um, they had it set for occupied and unoccupied. But when John did the data modeling and saw the overnight load profile, we've come to find out that some of the facets between the Niagara system and the base building control system were mismatched. So everything on our Niagara system was saying, this thing is going unoccupied, we've sent the command, everything is great.
But it wasn't actually taking in to these 15 RTUs. So, John and his team, through partnership with OTI, the Master System Integrator, were able to go in, modify the sequence, uh, modify the integration, and [00:34:00] shut those RTUs off per the schedule the operator had already set. And that level of, you know, visibility and deep insight, um, really showed a tremendous amount of value.
And just by that one act alone, before we even get into automated demand management, We've already seen, you know, kind of in the first month, a 20 percent reduction in energy consumption. So, um, that's a, that's a tremendous value before we even start talking about automated demand management. Um, the second, uh, case study that I want to talk about is, um, a product.
Another regional, uh, mall in, uh, Southern California. And we implemented this, uh, uh, platform in Q3 of last year. So right before holiday season, uh, but kind of in the shoulder months of, of California. Um, and since then, we've seen a 1 percent reduction in our energy costs. Now you might be saying, wow, great results, but in that [00:35:00] same exact time period, we saw a 15 percent increase in our rate structure.
So this platform was able to help us avoid a 15 percent cost impact to our utilities, which is tremendous for us. Again, going back to, we're not asking our operators to change any behavior. We're using technology. To insulate us from costs without any sort of behavioral changes. So that's all fine and great.
But let's, how does demand management actually work? And what does that actually do? So I saved that one for last. So at a very large multi use, uh, center in Northern Virginia, Um, we're seeing 250 KW. So for context, about 25 to 30 percent of our total HVAC load Being load shed daily without any impact to comfort or any rebound spikes from like a [00:36:00] demand, uh, response type of implementation.
Um, and this has been going on for the, for the last, um, month and a half to two months, um, so during, you know, kind of the shoulder season of Northern Virginia, and we expect to see these kinds of savings continue on into the summer months, uh, because that system is a hybrid of direct exchange and central plant cooling, so we're able to really Leverage all the desperate systems in one unified way.
But this is the power of automated demand management and, and we really picked these three sites for a lot of reasons, right? They're, they're big centers. Um, you know, the first center that I talked about, we've just done a controls retrofit and this was something that was just a simple You know, um, myths, um, that nobody would have seen and we were able to find it through implementation.
The second example really shows that, you know, grid interactivity and rate structure, [00:37:00] um, decoupling that John's talking about to insulate us from, uh, utility rate increases. And then last but not least is the example of automated demand management where a system is. Fully functional, ready to go, and ready to really leverage the full power of this platform.
When it's ready to go, you see these results almost instantaneously.
James Dice: And John, that was amazing. Thank you for taking us through that, Ryan. The, and I love how you presented the three different types of value. That's really cool. The, um, John, the different ways in which you guys reduce demand. So Ryan talked about 250kw at that site.
Can you talk about the ways in which you make that, that happen?
Jon Schoenfeld: Yeah, so that particular site, um, we have a scheduled Demand reduction that occurs every day. Uh, I believe it starts at like three 30 or four o'clock and it runs towards [00:38:00] essentially close to the end of occupancy. Uh, and we, we schedule this reduction for a couple of different reasons.
One, we have some fairly. Large fluctuations in the meter for that building that make it hard to control the, the, uh, load shed by the meter itself. We're also don't have a, a steep spike in utility demands for this particular building. It's a fairly flat load profile, but they're still paying, uh, energy costs.
Those energy costs are highest, you know, at the three o'clock to eight o'clock hour. You know, not to mention this is also the dirtiest time of the grid as the sun is going dead down and the solar production is the lowest, yet there still is peak, peak energy consumption. So it, this is probably my favorite example right now.
I think I'm going to have another favorite within a [00:39:00] couple months, but if you look at the load profile before and after. I mean, it basically looks like somebody took a big bite out of the load profile of this building because it just, at three o'clock, it just drops off, uh, and it does that until essentially until the end of the day.
And then, and then steadily goes down from there. So really, really powerful how you can see it. It, it make real impact to the, to the load profile of the building.
James Dice: Awesome. All right. So then Ryan, what's, what's next on the horizon? Um, for this, what are you excited about doing now that you've sort of proven the value with these initial use cases?
Ryan Knudson: So beyond rolling it
out to
the other centers, you know, kind of enhancing the platform, I'm very excited to, to work with John on integrating our onsite renewable energy assets. Um, and having those assets. Uh, inform, uh, some sort of optimized startup of our HVAC. Um, you know, John and I talk, uh, extensively about, you know, pre [00:40:00] cooling and the prevailing wisdom of that around rate structures.
But, you know, my, from where I sit, I see that there's tremendous value in informing our startup routines around when production of our solar assets are. And allowing those solar assets to absorb the cost. And that is something that I'm, uh, very keen on. Um, again, Uh, as I said at the, at the top, finding new and novel ways of deriving value from existing investments, uh, using data and using technology.
James Dice: Absolutely. I actually haven't heard anyone connect those two concepts before. Basically, I think everyone thinks about starting up things, uh, starting up devices, HVAC, et cetera, before the sun comes out usually, and what you're talking about is in a retail environment, it feels like. Like stores don't open till later, right?
Is that what you're saying? And so the sun's coming out [00:41:00] at a pretty good time to absorb that, that initial, okay, shit, we got to turn the cooling on right now. Absolutely. That's, that's fascinating to me. Absolutely. It's really cool.
Ryan Knudson: And, and I would much rather have my HVAC running at full tilt, uh, when my solar is running at full tilt, than having my HVAC run at 50 percent when my solar is at zero.
James Dice: That's, that's innovative. Um, all right, let's talk about lessons learned. So I said at the beginning, we don't want this to be a fluff story. We don't want to act like everything went perfectly when you guys are trying to innovate and do stuff that hasn't been done very much in buildings before. So, so feed me, feed me with some, some challenges that you guys saw and sort of how you have been overcoming those challenges.
Ryan Knudson: Sure. So I'll start with mine and it's something that John very politely touched on at the, uh, case study in Northern Virginia around metering. And, you know, Matrix has been monitoring our utility meters, um, at the kind of master level for a while, um, and James, as [00:42:00] you said, you and I talked about data quality for ESG reporting, um, a while ago, um, and John really kind of proved out some of my concerns around the quality of the older investments of meters, and, um, So while at the northern Virginia site, we have this scheduled load shed.
It would be great if we had strong meters. So the lesson learned here is focus on your meters, right? And it always goes back to the meters, right? It always goes back to the data quality. But we're now looking at, um, uh, making some investments in some of the older metering infrastructure, um, to improve those so that we can Really start to accelerate, um, their platform in new and, and novel ways and, uh, expanding on the metering, uh, down at some cases to the major asset level, such as central plants, such as large scale, uh, rooftop units, 100 ton [00:43:00] plus, putting some meters on that, but really expanding our meters and really getting better resolution on our consumption data.
James Dice: Cool. And I just have to give our team a shout out right now because we have an active vendor selection for metering right now. Um, and so people that are listening to this, if you're a metering provider, Ryan's interested, but also we, we are, we are interested as well in terms of finding, finding good, low cost, uh, metering solutions.
Um, so. I want to hit this real quick because people might not people, I know we want to move on to the other challenges, but hitting on this one real quick, people might not like grasp the connection between good meter data and what you guys are doing with demand management. So, John, can you talk about why?
You had to change different, you talked about earlier, two different strategies. If you had good meter data, you do this strategy, bad meter data, you do this strategy, what's the function of meter data for you in this solution?
Jon Schoenfeld: For, uh, and I'll use the, the [00:44:00] Connecticut, um, site as an example, that there are, our, our first pilot some two years ago, uh, they had solar on site.
They had high demand charges. They had a very steep peak that they hit, um, and we're paying quite a bit to the utility for, uh, for those, for that steep peak that was happening at around seven o'clock. So in order to mitigate and reduce that peak, we're essentially using that, uh, that meter data, that, that peak or that measured demand as our, um, our process variable.
That's what we're controlling to. And so as soon as that value goes above our set point, We start ramping up our load shed. We control our load shed just the same way you would control a, uh, a valve or a VFD with a, with a PI loop. Um, and so, uh, [00:45:00] we're able to essentially ramp up and down our load shed to maintain a, a, essentially a maximum demand set point, but that only works.
If the meter is reliable, right? And, and we learned, you know, essentially the, the hard way when the meter goes down, well, it's going to stop working essentially. Um, so, uh, you know, from that, you know, lesson learned, that's where we kind of developed the, the scheduled approach, the scheduled strategy. So we can deploy now everywhere.
Uh, and when, when some of these meters get resolved or get, I should say, uh, get replaced. We'll be able to, you know, flip a switch, change the strategy and, uh, and really focus on, uh, on peak demand where it makes sense.
James Dice: Okay. What's our next challenge that you guys ran into?
Jon Schoenfeld: So I'll take this one. We mentioned the, the fault detection diagnostics and the integration of that, uh, into the [00:46:00] algorithm, but we also need those, um, those things to be fixed, right?
We need those issues to, to, to be resolved so that we can maximize the. The opportunity to maximize the optimization. So a couple of things we did to really help that is, uh, one, I think it goes to, to Ryan's team and, and the Maestrich team and really creating, uh, a strong rollout of this where everyone's on board in what we're doing and what we're trying to accomplish.
So it does not feel like the operators do not feel like this is being done. Being, you know, put on them, but they feel engaged and the, the turnover and the training that we've been doing has been a really successful at each of these, uh, at each of these properties, they seem there, the teams are very much engaged and some of the issues that we've identified have been resolved almost immediately.
Um, [00:47:00] and that's really, really powerful to see. Uh, and so having that engagement and then displaying the information in the right way, right? So we are showing each of the operators here that here the equipment that are part of this strategy Here are the ones that are disabled because maybe they serve a tenant space Maybe they serve an area that you don't want to be part of demand management and here are the ones that are enabled But are disqualified because of some FDD.
So there's already this kind of almost gamification of, all right, how can I get this slice of disqualified units to be as small as possible so we can maximize. Uh, the, the efficiency and maximize the, uh, you know, the return on investment.
Ryan Knudson: So the last challenge, um, and, and we kind of touched on a little bit, um, it's just the disparity in rate structures, the disparity in utility charges, and making sure that the deployment was flexible enough to.[00:48:00]
Meet the unique challenges. Um, as John said, you know, we may have a property that has a TOU and that TOU is, you know, 4 to 6 PM. So that's when we really want to target it. So not being, um, so, uh, uh, broad. In our algorithms, but really kind of tweaking and tuning the base concept for not only the physical, um, makeup of the building, but also the maximum impact of the platform.
And so that took a little, little bit of time and a little, a little bit of, uh, analysis and introspection, um, because. As John will tell you, I always want to go much faster than we end up going. Uh, and he very patiently explains to me how, uh, property A's value proposition cannot be directly translated to property B, uh, because he needs, he and his team need to do the modeling and take into account all of the value [00:49:00] variables.
Um, and, uh, it seemed it more often than not, he is correct and I am just patient.
James Dice: John, can you talk about, you guys do a lot of, um, data modeling leadership, I would call it in the industry, really sort of, um, Trying to take a lead and own and share what you're doing on the data modeling side so that other people can sort of learn from it, adopt it, etc.
Can you talk about some of the modeling challenges? I'm just assuming there are modeling challenges around modeling these different utility rates, um, and including those in the control sequences.
Jon Schoenfeld: Yeah, I mean, I, I think it's a, it's, it's a great question. Um, through the process of developing ADM, we've had to standardize on new data points and new, you know, these new control strategies need ways to be controlled.
And so with our, uh, you know, what you're referring to the, primarily as the ontology alignment [00:50:00] project, we've essentially created these new, these new points, these new attributes. These new parts of our data model, and then we, we publish them, right? So we don't do them in a black box so that only us, only we can use them.
We try to, um, generalize them as much as possible, uh, so that they could be used by anybody. So, you know, if you look right now in, uh, in the OAP, you'll see things like. Loadshed command is, you know, a new point that's available. Loadshed sensor is a new point that's available. Uh, automated demand management enabled is a new point that's available.
These are all part of now the data model, which then makes them available to the end user, to the operators, so they can see how is Loadshed behaving. How do I disable this unit as opposed to this unit? How do I make a change to the maximum set [00:51:00] point that this is going to go, you know, that this, this set point is going to be raised to.
Demand management, the way we do it is probably different than the way others do it, but some of these core concepts are the same. Um, and so there's use of those points regardless of. You know, the, the exact strategy that you deploy.
James Dice: So, so Ryan, can you talk about simply the, the paying for it? The question here, how did you get business approval for this?
Like what's the ROI, uh, calculation look like and, and sort of how do you make the business case?
Ryan Knudson: Sure. So, um, you know, the business case was really built, uh, again, upon, um, looking at, uh, the investments that we had made over the years. And, um, the executives really came to me and said, Hey, rate pressures are so extreme.
How can we leverage all this technology that we've been investing in for years, um, to insulate us from these rate pressures? Or is there anything that we can do? So it [00:52:00] wasn't so much that, um, you know, I had to, So called sell up. It was a lot of, you know, kind of top down pressure around, you know, the last 24 to 30 months of weight escalations and utility costs, um, you know, increase, um, and not throwing more large capital into, you know, onsite generation, but leveraging everything that we've done to date.
And, yeah, and, um, that's when I went to, to John and to Clint at OTI and said, uh, you know, Hey, gents, you know, we've been working together for eight years. We need to come up with something. You know, what, what do we have? What have we been talking about? And are we ready for it? Um, and, uh, I go back to kind of that, that lag between the pilot and full deployment.
That's where we really developed those models. And we really took a hard look, um, at, uh, [00:53:00] that first pilot site. Uh, and I think John had to run the savings model about three or four different times at different points, uh, throughout the, the kind of almost year gap. To really prove out that this, uh, would be frictionless to our operators, um, and would provide values throughout the year and not just, Hey, during the shoulder months, you're going to make all of your savings, but in the summer, it's going to shut off.
And, um, you know, we'll hope we'll pick it up in September. Um, and that's what we built our savings models on. Um, and then we just deployed that saving model. To each center based on size of the property type of system on that goes back to being very strategic and selecting your pilot site, making sure it has a good cross section of the types of systems and buildings that you're going to end up deploying to.
because we were able to then. Use that model across the rest of the portfolio. And, um, today we've, we've seen better [00:54:00] than what we had estimated. Um, frankly, um, because of some of the, the wins that we talked about in the results, uh, that we just weren't expecting to, to find. Fascinating.
James Dice: All right. Last couple of questions here as we, as we sort of close out, these are sort of conclusion sort of questions.
Um, Ryan, if you're sitting down with, you know, you have a peer group that you meet with regularly, other, other sort of people in your, in your position and similar portfolios, what would you tell them in terms of like, here's the playbook to sort of copy this approach and sort of replicate the results we've had?
Is there a five step playbook you could tell them? What would you tell them?
Ryan Knudson: Yeah, I
think step one is to, uh, communicate internally with the Uh, your, your stakeholders, you know, partnering with your I. T. team, making sure that they are ready to onboard, um, such a, uh, this kind of technology and what their privacy and cyber security requirements are, uh, then [00:55:00] working with facility management to, uh, help understand what centers, um, are, uh, Uh, positioned to, you know, deploy.
Um, if we're about to have a major redevelopment, probably not a good time to deploy, you know, advanced algorithms. So, really understanding, um, all of the internal, um, rules and requirements, that's step one. And then step two is, uh, to look at your rate structures. Understand where you think the biggest value is going to be derived, um, and, uh, making sure that you are best positioned to achieve those.
Uh, and then step three, um, is site selection. Um, and really diving into the assets, uh, and working with those local teams to understand what their challenges and their, uh, needs are so that we're not, uh, deploying, um, you know, space age technology on [00:56:00] Flintstone's infrastructure, um, and helping them get to a place to be ready for it.
Uh, and then step four is, you know, deployment. Deployment, active management of your deployment, being engaged, being transparent, making sure that all of those stakeholders that you talk through, step one, step three, are fully informed, um, at every step of the way. And then last but not least, the step five, monitoring and management, going back and doing the M& V, showing the results to all of those stakeholders, right?
Um, and, uh, Building the case for the next group of properties. Building the model for the next deployment, uh, group.
James Dice: Space age technology on Flintstones infrastructure. I'm gonna steal that 100%. Please. I'm definitely gonna steal that. Um, alright, my last question for you guys is, um, What haven't we covered here that you would, that you would want to make sure we covered?
And I'm going to start, I'm going [00:57:00] to start with this. Um, I just want to say thank you, Ryan, because it is really hard to find someone on the buyer side, someone on the building owner side that is willing to talk to, The level of detail in their projects, like you're doing right now, it's even harder to find someone willing to talk to about it at all, but we'll set that aside, I think in order for our industry to transform and accelerate the way that we at Nexus want it to happen.
And that's our mission, right? That's why we're here. We need more people like you, Ryan, that are willing to talk about your projects, and so thank you for doing that. If there are any other people out there like Ryan that want to talk about their projects, either good or bad, we're here. That's what we're here for.
With that, what do you guys think we haven't covered yet that you want to tell someone that's sort of just starting out and doing this? I would just say, um,
Ryan Knudson: Be patient with yourself, be patient with your operators, find a good, uh, core team, um, both internally and [00:58:00] externally. Uh, find, uh, good, uh, scalable, uh, master system integrators and, and systems integrators.
That understand your vision, are on board with your vision of, um, whatever you're trying to achieve. Um, and mostly, um, be ready for unpredictable results. Uh, and Being flexible enough to understand the value of unpredictable results.
Jon Schoenfeld: I couldn't agree more, uh, you know, the, the, the phrase that I keep coming back to is success doesn't run in a straight line.
Uh, and this isn't with automated demand management or advanced supervisory controls. This isn't something that you just pop in and, and walk away and everything works hunky dory. Um, there's, there's real work to be done. And some of [00:59:00] that success comes from that work. And, and honestly, a lot of the fun comes from that work, diving into the details to understand the root cause of that issue so you can solve that issue very quickly and easily and save, save a bunch of energy, save a bunch of money so that the optimization that works from there is.
It's part of the fun of the project and it's part of what makes it impactful. Uh, the last thing I would say that, that, you know, one more kind of, I guess, key thing we haven't talked to, I don't know enough about is the, the openness and transparency of the solution and of the data model, right? At the end of the day.
Um, we're serving MaceReach with a, a transparent and open system where all of their data is available via an independent data layer. And all of it is modeled via an open, [01:00:00] uh, Haystack compliant data model and, and as well as compliant with other data models with the, uh, with the ontology alignment project.
That means that if tomorrow MaceReach finds Some other, you know, strategy related or not related to bolt on to their existing system. They can do it. And they don't have to pay a vendor to recreate the integration, to recreate the data model, it's already done. Uh, and so that, that just makes value for the future that it doesn't go away.
James Dice: Thank you both for sharing this. This is so cool. And I'm excited to hear more about as you, as you scale up.
Rosy Khalife: Okay friends, thank you for listening to this episode. As we continue to grow our global community of change makers, we need your help. For the next couple of months, we're challenging our listeners to share a link to their favorite Nexus episode on LinkedIn with a short post about why you [01:01:00] listen. It would really, really help us out.
Make sure to tag us in the post so we can see it. Have a good one.
Head over to Nexus Connect and see what’s new in the community. Don’t forget to check out the latest member-only events.
Go to Nexus ConnectJoin Nexus Pro and get full access including invite-only member gatherings, access to the community chatroom Nexus Connect, networking opportunities, and deep dive essays.
Sign Up
Nexus Labs exists to move the smart buildings industry forward, without the fluff.
On Wednesday, we send out the renowned (and free) Nexus Newsletter.