Boosting Campus Operational Efficiency Through the Power of Autodesk Tandem Digital Twins

RAYMOND ISSA: Welcome to BLD4024, Boosting Campus Operational Efficiency Through the Power of Autodesk Tandem Digital Twins. You presenters today are myself, Raymond Issa, Distinguished Professor and Director of Center for Advanced Construction Information Modeling in the Rinker School of Construction Management at the University of Florida. Frank Phillips, Director of UF Business Affairs Technical services. And Isabelle Southern, a PhD student in the Center for Advanced Construction Information Modeling.

Our agenda for today is Current Status of Facilities, Why Digital Twins, Building Digital Twins, and Moving Forward. Current Status of Facilities. The University of Florida has over 900 buildings in this facility portfolio. It has $1.6 billion in deferred maintenance backlog that needs to be taken care of.

The first building on campus was built in 1906. As we all know, 60% to 80% of life cycle cost of a building is the operations and maintenance facet of the building. Why do we want to use digital twins? Opportunities and applications. First of all, we'd like to manage all buildings remotely.

We'd like to monitor what's going on with all our systems 24/7. Digital twins is one way to do that. We'd like to diagnose and predict problems before they arise and to work out inefficiencies and maintenance needs. And more importantly, we want to stay informed so that we can make better choices to reflect the institution's needs.

This graphic represents how we go from the physical asset to the digital twin. Transformation process looks at the physical asset. We collect data from drones, sensors, video devices. And using IoT Platform, we do data analytics in the cloud. And we feed that the digital twin, same thing with the video. We feed that to the digital twin and extract data for from it.

So the digital twin acts as a framework that allows us to integrate all the data streams associated with a building. In return, we can send information back from the digital twin to the building, for example, to activate actuators, maybe to turn on a motor or turn off a motor in the building, or anything else that's on or off or actionable.

In this next image shows you the digital twin maturity continuum. We start with existence twin, which, in essence, is a building information model. LOD 500. It has the properties and location of the components. Then we go to status twin that exemplifies the condition of the building. All the readings, all the sensors are integrated with that model. Then we go to operational twin that adds work and cost data to it.

And then we get to the simulation twin. And the simulation twin is of great importance since it's let us look at what-if situations. What-if situations, what if we change certain parameter? What happens? And based on the results of these what-if situations, we can decide on whether or not to implement those in the real building. And then, finally, we have the cognitive digital twin.

The cognitive digital twin allows us to use AI and machine learning to predict problems with the building operations, as well as diagnose problems and prevent them from happening. So our aim at the University of Florida is insightful operations. Digital twins drive campus-wide information consistency.

So in terms of operations and maintenance document organization, the digital twin framework is used to organize those, fall detection and disaster mitigation. Again, the digital twin provides a framework to display those maintenance work order and management, space track and asset health monitoring, tracking HVAC performance, leverage knowledge or informed decision making, and finally, holistic view of the performance.

The digital twin integrates all this data and provides contextual knowledge as well as allow us to be contextually aware of what's going on with our building or buildings. Building Digital Twins Case Studies and Use Cases. Isabelle?

ISABELLE SOUTHERN: So here in Heavener Hall, we are showing the ability to create and name views that highlight different trades, areas, or features of the building. We can view our realistic architecture outside and inside. We have the ability to view internal systems and trace the movement of air. For example, here you may have received a work order for an inoperable chilled beam.

You're able to find out which chilled beam it is, what air-handling unit services that chilled beam, and decide whether the air-handling unit or the chilled beam is the issue. You are able to view all of your systems serviced by a specific air-handling unit or a chosen VAV. As shown here, you are able to isolate the rooms that are being affected by a VAV.

You are able to trace fire protection and locate emergency valve connections from the exterior. And we are able to focus to an individual sprinkler. You are able to filter your assets by warranty, manufacturer, floor, type, room, and room type, and any other parameter that you have chosen to view your assets.

You're able to view our fire-rated walls and doors, and also view floor-by-floor performance, which display heat maps for any parameter that you have populated in the model that reflect the real-time performance. So by choosing each asset, you are able to view your live performance or go back and view historic performance points to gain better understanding of your assets.

RAYMOND ISSA: Thank you, Isabelle. Frank?

FRANK PHILLIPS: Most of our buildings on campus are older buildings where we do not have full MEP. But we are looking at other ways to use this. And in Tigert Hall, our main administration building, we brought in our space data. And this is setting up a filter so that we can filter to information about space, filter to a college and see the college distribution. And once that is accomplished, we can color code by college.

And from there, we can cluster to see the groupings by college. In this case, we then filter based on some of our space data to remove our nonassignable space, leaving us with all the usable space in the building that the colleges occupy, once again, in this case, their administrative units. From there, we can filter to a floor. And from there, we can drill in and see individual information about a room.

We've also built a number of links into the system, this going back to our space tracking system where the information originates and, next, to our maintenance management system showing us all the historical work orders related to this individual space.

We can then go back into the default view of the space and look at space adjacencies across the college units. And then return back to the complete distribution across the building. And in this building, this is one of our main research buildings in our medical school.

And this has a precanned filter that, by default, does a color coding that we set up before the video. And from here, we can filter this further to look at research productivity that we're bringing out of another system and into Tandem. And we can, from there, set up individual cards that filter directly into the information that we want to be shown so that we are not having to look at all of the peripheral data about a given space.

So here we're adding information about the research productivity, the college involved.

And we've color coded each room by its productivity index. And from there, we can set up a card to show us the information specific to the room that we want to see in the card. So using the asset properties we pull in against the research space, we can take our space indices and our colleges and add those to a very specific card so that we don't have to filter through the full property panel.

So now, as we move from room to room, we get a very targeted view of the information that we want to share with our constituencies to manage the space within the buildings. And as before, we can further filter this to an individual college.

And from there, we can cluster and color code by the departments within the college. And our card still contains the filtered information that we set up earlier. And as before, we can restrict it to a floor. And from there, we can return to the adjacency view in the building to understand how the space correlates to each other.

RAYMOND ISSA: Isabelle?

ISABELLE SOUTHERN: So now that you have seen what we built, here's a brief outline on how we built the digital twin with the real-time streaming. From the building automation system, we extract all points that were being tracked in the building and input them as parameters in the Autodesk Tandem environment.

We then created a facility template that contains all of these different parameters in OmniClass format, which is the desired classification system for UF facility and asset management applications. Our live streams were achieved through API connections, which allowed for the real-time reflection of performance.

On the other hand, all of our information from the asset, the document management, and the space management system are imported in the model using a Tandem Connect integration. And the result is a very user-friendly interface where all of our different views can be filtered for our different applications, as seen before. The Autodesk Tandem environment allows for the realization of all of these streams of information shown on the far left and the far right of our implementation workflow.

As we focus on the far left side of the workflow, here shows the payloads received from the building automation system. The BAS, which are mapped to the appropriate parameters, created in Tandem using webhook connections. Below the red arrow shows the output stream inside Tandem for the different parameters over time.

Our data is received in Tandem in intervals of 15 minutes for our applications. But we are able to also pull historical information at any time.

RAYMOND ISSA: Frank?

FRANK PHILLIPS: One of the significant components to dealing with the digital twin is being able to deploy meaningful information to the personnel that need it to maintain and operate the facilities. On the left side of this screen is a web document that we deploy through a QR code that would be attached to a piece of equipment.

We're using the same data stream to populate asset information against a piece of equipment inside of Tandem, but also augmenting that with additional information that links out to the property asset record, as well as its work history and other information. And also, we include cost information in this so that at, at a click from Tandem, you get a cost summary of the equipment's history.

In this screen, this is leveraging our space management system data very much like we saw from the asset side. On the top is information out of our space management system where we have the floor plan of the building based on our Autodesk Revit model. And on the bottom is one of the views that we saw earlier where we are color coding this by the departments.

But then you can see that we have a commonality in the data set where the information that we are pushing into Tandem can be accessed through a 3D view, whereas we're seeing it more textually in our space management system. And once again, we include links to all the other internal systems that we support.

RAYMOND ISSA: Frank?

FRANK PHILLIPS: One of the key elements that we're working on getting connected to Tandem is the document management system, whether that be based on a piece of equipment, as we see here, or any other information that we happen to be tracking, it's very critical that the metadata be accurate and tagged out across all of our systems so that we have commonality in how we refer to equipment so that we can access the important information for the people that need it to operate and maintain the buildings.

RAYMOND ISSA: Opportunities and Applications, Utilizing Technical Capability for Facilities Management. Well, one of the most important things we do with our digital twins is real-time monitoring of how the facilities are operating and working. Another important component is work order management. We want to know the total cost of ownership of all the components in the building.

And we want to track those, so we can have historical data. Frank talked about the space optimization. We'd like to know how much each space, for example, each laboratory gets used, what kind of overhead it brings in, what kind of research it brings in. So when we plan another lab space building, we know what labs we want to add to it and, of course, strategic planning for buildings on campus, for changes in buildings, and all that.

The value of the near real-time monitoring for Hebner Hall, for example-- this is an example right here. We have the number of AHUs, VAVs, FCUs, all the components, plant maintenance. We have the plant maintenance per-year column. We have the third column represents the number of inspections per day since we're monitoring.

Sensors are read at every 15 minutes at a 15-minute rate. So that would be the daily number of times we touch this equipment to evaluate these operations. And then we translated that into a yearly number. And so you can see, using such a system, we're nearing 10 million observations of how the system's working.

And so that can give us an accurate state of where on the components of the building we can look at the quantification of failure impact and use that to prioritize our repairs for the building. For example, if we have replacing the filters every month, that's a standard maintenance. Well, by using the sensor, we can know the pressure in front of the filter and the pressure behind the filter.

And that might lead us not have to change it every month, but change it as needed if it makes economic sense. So those are different things we're exploring using these digital twins. Moving Forward, Challenges and Future of Digital Twins. Isabelle?

ISABELLE SOUTHERN: So what moved the industry from standard BIM to this idea of a digital twin was the potential for connecting the entire ecosystem. Our standard BIM does continue to enhance the project process through design and construction. But when we start speaking about facility management, we encounter some obstacles.

So here at the University of Florida, we do use several different systems for facility and asset management, as shown here on the right, that all speak in different ways. Autodesk standard provides a way to integrate all of these different interfaces into one environment, enhancing information, accessibility, and our ability to make decisions.

So shown here is the network of the different systems that are utilized by UF to make the digital twin a comprehensive and functional model. Each introduces its own level of complexity, which is adequately managed by Tandem by way of a single integrated information stream.

Some hurdles that may be present in the development of the digital twin include inaccurate models and unmonitored changes, especially as we work with existing and older facilities. Asset information or even models may be very old and outdated.

For example, we experience connectivity and clash instances, which makes it hard to gather all of the updated information to make a living model. Also, integrating all of our facility and asset management system into the model using Tandem Connect is a unique and a challenging process.

As a common name for assets may not be there in the model, when we received the model at project handover or other bugs in the system may make it hard to share our information between our asset and facility management systems and the model.

RAYMOND ISSA: Frank?

FRANK PHILLIPS: So as Isabelle just pointed out, there are very many challenges that we have on our quest to get to a good state with digital twins on our campus. And one of those key being user adoption. And so we have a lot of work to do on making this useful to our end users. And some of that is identifying what is useful to them, getting them trained.

And one of the big efforts that we have going today is the information collection for existing buildings, but also reworking what we need from project delivery and handover, not only for new major construction, but also for the change management process through renovations and maintenance activities. And then, once we get everything wrapped up and we can execute this on a recurring basis, our digital twins will be alive, useful, and living like the existing buildings.

So it's very critical that these processes be both identified, documented, and executed with a recurring methodology.

RAYMOND ISSA: Isabelle?

ISABELLE SOUTHERN: So as we move forward, we are focused on perfecting the implementation workflow for creating our digital twin in Tandem, so it does apply to all UF buildings, old and new. We are also using this workflow to validate our models against the existing-built environment and develop a digital twin execution plan that is built into every new project.

Further, we are focused on refining our user interface in Tandem for a fault detection notification system. So we are able to alert our users and our facility management employees based on minimum and maximum performance values that are appropriate for each parameter.

And finally, we are aiming to enhance communication between industry professionals by utilizing artificial intelligence and machine learning for fault detection and other insights using the digital twin. So if we refer back to our five levels of digital twin maturity, our goal is the most advanced level, a cognitive twin.

And finally, overall, UF does aim to digital twin the entire campus in a Campusverse, where our insights are leveraged for any chosen facility with the ultimate goal to enhance student experience through efficient facility and asset management. So using geographic information systems and other 3D-scanning technologies, this aim is currently being executed. And we look forward to further executing it with Autodesk.

So I'd like to thank everyone for your time. For more information on how our work fits into the aim of Autodesk, please feel free to reach out. We're very excited about the motivation for the future that Autodesk shares with us. And we look forward to expanding this information beyond our status twin and beyond Heavener Hall.

So we are ultimately aiming at an industry-applicable tool, using a digital twin at a very broad scope. And we would not have been able to accomplish any of this without our team at Autodesk and at the University of Florida. So all of these names here on this slide, it goes to show that it does take an army as an industry as we move forward to a future using digital twins. Thank you.