Monday, February 17, 2020

AI and the Future of Predictive Maintenance

Granted, when you're in charge of the world's tallest building, Dubai’s Burj Khalifa and (an assumption on my part) money is no object, you're able to afford the very latest technology.  But even though most campus facilities managers may not be able to relate to the budget their colleagues in Dubai have to work with, the technology they're using--and what it might mean for the future of facilities management--is worth discussing here.

Photo Source HERE
A recent promotional video by Honeywell includes an interview with Bashar Kassab, the senior director of facility management for Burj Khalifa.  On the video, Kassab says, "The people that occupies Burj Khalifa expect nothing but absolute perfectness, so we have to live up to that standard and up to that reputation."  The video focuses on how "new predictive maintenance technologies are enabling the Burj Khalifa to reach a new caliber of operation and create a better experience for the millions of tourists, shoppers and residents who pass through it each year." 

Standing in the building's control room, Kassab continues:

Everything in this building is connected to this room like, you know, in our body everything is connected to the brain. From this room we get to know what is happening in the building. We get to control it, act and react before anything happens.

We have so many things that come to this very room. We have the building-movement monitoring system, so we have sensors that are fixed in various locations in the building and they monitor the building movement. This is one of the information that we receive live. The other information we receive is that live data of wind. This is very essential for our operations, especially for the window cleaning, because we have up-sailors that go outside, and we need to tell them when it is safe to go or unsafe, and they have to stay inside the building. The air conditioning system that keeps this building livable, so to speak. And, if the pumps aren’t working, if the AC units aren’t working, if there is any trouble in the system, literally everything in this building communicates back and forth to this very room.

Burj Khalifa is unique in many ways. First of all, it is the first of its kind to be built at such heights, so the engineers had to be innovative in the ways they do construction in terms of concrete pumping, in terms of drainage and so on and so forth. The design of the building itself is also unique. They had to design it in a wind tunnel. Basically, here we are very close to the sea and the desert, so the wind shifts direction twice a day, and Burj Khalifa being at 828 meters, it acts like a giant sail, so it had to withstand that wind pressure.

In the summer here it can easily reach 50 degrees centigrade—114 Fahrenheit. Sandstorms are basically a form of moving wind with sandblasting into it. So we have different types of facades that are designed specifically for the building, as we have specifically 18 types of building facades fixed in the building. Each is carefully located to design the anticipated pressure of wind. Being also at 828 meters you literally touch clouds. With clouds, comes thunder and lightning, so we have a very intelligent system that is fixed on top of Burj Khalifa. How it works, very simply, it senses the charge in the air. So if the charge is positive, it will energize itself by a negative charge. It will attract lightning to hit it instead of hitting the buildings from the side.

The building's architect, Adrian Smith, co-founder of Smith + Gill Architecture in Chicago, adds:

We’re always thinking about the environment when we design buildings and, you know, Burj Khalifa was no different. It was just that in super-tall buildings you have conditions that are important to take into consideration. For example, the exterior wall weight and the weight of the glass system, or the weight of concrete, or precast or metal skins. All those have to be considered in the holistic design of the building, because the structure is impacted so greatly by the weight of the system itself. Just as the mechanical systems and cooling systems are impacted by how much heat comes into the building at any one time.

Building manager Kassab reiterates the importance of his work and discusses the real "brains" behind the success of the building's operation:

Burj Khalifa is a critical building, and it is very important for the country and for Emaar, the developer of Burj Khalifa. So we deal with the assets with highest criticality in terms of how we maintain them. As an example, if you are an ear, nose and throat doctor, okay. And if you are having an itch in your throat, you’d be able to self-diagnose yourself and self-prescribe the medicine and know at what time you should take it before it actually escalates. Right? So what could we do to make our assets do exactly the same? Diagnose itself, communicate to the technician, before an alarm is even triggered, saying that, “I’m experiencing an abnormality. Please come and fix me before it is a failure.” We went to our service partners, Honeywell, and we shared this idea with them and we said, “How can you help us with this?”

Honeywell's representative, David Trice, explained their approach:

Predictive maintenance gives us a way and a mechanism to keep track of assets and learn from their behavior, so that in advance of a failure or some kind of operational abnormality, we can detect that and see it in advance versus a break/fix model where you wait for it to break and then you send somebody as a result.  They said to us as we were going through the predictive maintenance testing, “We need to see a level of accuracy in your prediction. So come up with a mechanism to show us that your predictions are accurate, and then once you are able to convince us that you have a mechanism for showing the accuracy, show us how you’re going to improve the accuracy over time.”

The scale of this building is one that presents unique challenges. You have to pay very close attention to the minute details in order to drive the efficiencies that you’re looking for. And that’s really where the challenge comes from here. Not so much the size, because we can scale up to meet the size. But it’s really ensuring that we’re focused on trying to solve the right problem to ensure we get that benefit.

In your home, you’ve got the furnace and the compressor that go along with your HVAC. In a large building such as this, you have a chiller and a boiler, and an air-handling unit and fan coil units. And all of these assets in a building like this can number in the hundreds to drive an HVAC system. And so when you change a temperature, these things all turn on. In order to truly optimize, you have to know when to turn them on, how to turn them on and when you want to turn them off. And if you can do that in a more efficient manner, you can drive big savings.

Kassab explains that,

Honeywell came up with algorithms so the machine would understand what is happening. And the machine is learning with time, so it is building that portfolio. So, for example, if the machine decided that I might reach that failure point in six hours, and it fell prematurely outside within maybe two hours, it would understand that OK, this is the threshold, this is the span and this is how I react to what’s that abnormality. 

This has an immediate impact on the cost of maintenance and the reliability of the asset. So we are not impacting the reliability of the asset, we are not impacting the energy efficiency of the asset, yet we reduce the cost of maintenance.  Applying technology allowed us to save roughly 40% of our maintenance activities we perform on the critical assets....

My vision for a future predictive maintenance is to be fully autonomous. So the building will not only, or the asset specifically, will not only self-diagnose itself, but also self-repair itself. So I can go sit in a coffee shop with my phone in hand and the coffee in the other hand and ensuring that the building is running smooth and efficient.

A fully autonomous building, making decisions for itself, learning as it goes along, and controlling the environment of the building's human occupants, sounds like the premise for a science fiction story. But the lessons for facilities managers at all levels are fascinating, and worth discussing.