A Paradigm Shift in Building Operations : Digital Twin with BIM and AI-based Data Analytics

JANGSOON KIM: Hi. My name is Jangsoon Kim. I'm a project manager at Samsung C&T. And I'm working on project related to digital twin technology. First of all, I'd like to thank everyone for joining this session. Today's topic of my presentation is a paradigm shift in building operations using digital twins. Then, without further ado, let's begin.

Let me first briefly introduce myself. I majored in civil engineering at university and received a master's degree in steel structural engineering. After joining Samsung C&T, I worked as an engineer at bridge, marine, and port construction sites. Later, I took on project building related to BIM, providing technical support for project sites. And now, I'm working on planning and developing digital twin technology, and taking it into the commercialization stage.

You might think I have a pretty unique career path, as the work I'm doing now is quite different from where I started. But I'm really excited about the fact that I'm doing the job that I can do best now. This is a safe harbor disclaimer. Please note that this presentation may include some forward-looking statements. Before going into detail, let's first watch a preview video of THYNC, a digital twin developed by C&T.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- Once a digital twin starts thinking on its own, synchronized with reality, what would future operation of buildings look like? THYNC presents a new vision for building operation. THYNC utilizes various modeling methods to virtualize real-world spaces, and builds fit-for-purpose database and analysis models. Its simulation results support problem solving and decision making in reality. Here's how THYNC will bring paradigm shift for building operation.

THYNC allows us to comfortably look around every part of the building. It maintains the moderate level of comfort for users, also with an AI-based data, analysis, and simulation. Wasted energy during vacancies is tracked and reduced in real time by usage.

[MUSIC PLAYING]

Operating guidelines enhance the facility operational efficiency and optimize energy use. Analysis of building facility status enables performance preservation and management efficiency enhancement. Layout simulation allows for easy and fast optimal space planning. And evacuation simulation helps predict optimal evacuation routes and time. THYNC delivers sustainable building operations, ensuring comfort and safety for occupants.

Now, with infinite expandability enabled by THYNC, bring your own imagination of building operation into reality. In sync with your thinking, building operation digital twin, THYNC.

[END PLAYBACK]

JANGSOON KIM: So there was a preview of THYNC. In this session, I'd like to introduce the development process of digital twin that can be utilized for building operations, as shown in this video, and highlight its various features. Let me start with a question. Why did Samsung C&T, a construction company, develop a digital twin? We, as a construction company, are involved in every stage of building's life cycle.

With our involvement in design, construction, and expertise in operation and maintenance, we have deep understanding of the entire process. This allows us to build the most efficient and sophisticated digital twins. And that's how Samsung C&T started taking its first step in the building operation sector.

Starting from there, we can build physical spaces in the real world and digital spaces virtually at the same time. This helps us provide customized services that meet the needs of our clients, such as building owners and operations companies. Using digital twins, our clients can optimize the energy use of a building and improve operational efficiency during the operation stage.

In fact, our ultimate goal is to bring a paradigm shift in building operations through digital twins. With that, let me introduce our digital twin brand THYNC, and how it will help us achieve our goal.

Today's agendas include an overview, development process, and each feature along with demo videos. First, on overview, development background. I'd like to begin by explaining how we started the development. In our project 2022, we implemented a level-two digital twin that can monitor the operation status of facilities in residential complexes. Through this project, we verify the individual technological components of the digital twin. And we realized that sustaining the utilization of digital twins requires a clear purpose of use, and laboratory simulation is required to fully realize their true value.

So we decided to develop a digital twin technology by applying simulation features to building operations, especially energy management, with the goal of optimizing the energy use of a building. We've been conducting a demonstration project for global engineering center, which is our headquarters building. So we develop a digital twin for tower B for about 10 months. We aim to demonstrate energy waste assessment of the target tower to predict potential energy savings and verification of energy reduction by optimizing facility operations.

As real-time data are required for energy analysis, we installed measuring instruments and IoT sensors in an area of about 100,000 square meters for data collection. We are currently upgrading the analysis model and verifying energy reduction for A-spec facilities by applying optimal operating guidelines to the digital twin. Next, introduction of our digital twin brand, THYNC.

THYNC is a combination of the words, think, and synchronize. It means that the shape and data of a physical building in the real world are replicated and synchronized to the virtual world. Based on the AI-based analysis and thinking process, it supports optimal decision making for building operations. THYNC will become one of our brand identities, which represents the digital twin technology. Now, let me tell you about the development process.

This slide shows the development architecture of a digital twin. This was cited from the paper in the footnote below. The most important part of creating a digital twin is to define data [INAUDIBLE] for the purpose of use and to build an analysis model. You design control points for the data to be collected, and build infrastructure of transmission networks. And through 3D modeling, you can visualize the data in a space, increasing the sense of reality. Data and 3D models are the two pillars of a digital twin. And once established, they are integrated and interlinked.

And then, you develop an engine for simulation. After all, you have to perform each layer of the architecture as defined here to build an optimized restructuring. So let me walk you through the architecture step by step. Let's start with the data acquisition and transmission layer. First, you must define the data.

The type of data varies depending on which feature you are going to implement. For real-time energy management, especially energy analysis and simulation, data such as indoor environment data about temperature, humidity, and the air quality power usage, facility operations data, including cooling and heating aspect, and weather forecast data from the weather station are required in real time.

Collecting such data requires analysis of the building information about building attributes like energy, and certification, and operation and maintenance in order to design control points for the data to be collected. Next, you have to build an infrastructure for data transmission networks to collect data.

In our demonstration project, 77 IoT sensors for indoor environment have been installed in the offices. The data from these sensors is transmitted to a database server every five minutes through WebSocket API and collect power usage data. About 1,000 power meters have been installed on the panel boards in the room on each floor. And the data is transmitted and collected every five minutes through multiple TCP and IP.

We also collect facility operations data in building automation system, or bus, by linking the BACnet protocol. The data is transmitted to the database server every 15 minutes. In the demonstration project, about 2,200 pieces of operational data are being collected in real time. Each piece of data is collected to a database server by interlinking IoT sensors, measuring instruments, and building system protocols, and use for each proposed.

Collected data are transmitted to the AWS cloud for analysis and processing. The data is then visualized in a 3D game engine and finally provided to users in the form of services. After the data acquisition and transmission is completed, you move on to 3D modeling. The space for [INAUDIBLE] can be formed using various methods. For buildings, we typically use a BIM-based 3D model. And through photogrammetry and 3D scanning, we can create models of large outdoor spaces and individual object components, such as facilities and equipment in the mechanical room.

As for the level of detail for a 3D model, we achieved level-three based on the CityGML 3.0 standard. Since the main goal of such visualization elements of a 3D model is to improve visibility of various pieces of information, users can determine the level of detail for the 3D model according to each purpose of use. For efficient 3D modeling, choosing the appropriate modeling methods and tools is very important. Samsung C&T, using its own technical expertise, has conducted highly-efficient 3D modeling through various methods.

By dividing the outdoor and indoor spaces of the building, we classify targets of modeling in detail, and selected appropriate modeling methods according to the characteristics of each target. Modern tools are [INAUDIBLE] by modeling method. For example, you can use AutoCAD for referenced drawings, Revit and 3ds Max for model creation, Navisworks for model review, and ReCap when using point cloud, which is 3D scan data.

Using mainly Autodesk products, we were able to perform efficient and optimized modeling. The next layer of the architecture is data and model integration. After you complete data collection and 3D modeling, the data and 3D model have to be integrated and interlinked. Once data pre-processing and API generation are completed in the previous stage, the data in a 3D model can be integrated in the final 3D game engine.

Design UI/UX of a wireframe are implemented, and the data API and 3D model are mapped according to functional specifications. By mapping the data to the dashboard where buttons are operated, where numbers are displayed, a digital twin gets created. In the data model integration layer, the most important part is to develop a digital-- to develop a knowledge engine. When it comes to THYNC, It includes an energy simulation engine for energy analysis using AI-based optimization algorithm, machine learning, and deep learning models.

Each solution is connected and interacts with each other through a systematic pipeline ranging from data collection to final solution development. So that was the entire process of developing a digital twin. As I mentioned earlier, after the digital twin is developed, we provide it to users through various features.

The four basic features include energy management, which is core feature, asset management, space management, and virtual tour. Users can add features of their choice in the form of modules, which enables expandability. Then, let's take a look at them one by one. The first feature I will show you is THYNC Energy. Let's watch a demo video first.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- The energy management feature identifies energy waste, and reduces energy through optimized operation. A 3D model appears in the center of the screen with dashboards on both sides. Indoor and outdoor spaces of the building are created using various modeling methods, and data is visualized through mapping. The dashboards allow the monitoring of energy-related figures, including energy usage, and waste. First, on the left dashboard, you can monitor the cumulative energy usage of the entire building by period and purpose.

It predicts energy usage based on the actual usage, and wasted energy is identified for optimal usage suggestion. Also, carbon emissions and energy bills are calculated, enabling efficient energy management. Energy management solutions notify results from energy simulations in real time, helping building operations companies optimize energy use.

Next, on the right, you can check detailed energy usage by floor. Let's move to the fifth floor. On the left, you can see the fifth floor's energy usage and indoor comfort level. The six zones on the fifth floor are divided by the locations of six air handling units. If you select a zone, you can check its energy usage and indoor environment status, allowing detailed energy analysis within each floor.

Now, let's take a look at the details of an energy management solution. First, an indoor environment assessment solution is provided when the temperature, humidity and air quality levels exceed an optimal threshold. After checking the details of the solution, move to the applicable zone. You can see that the indoor temperature and humidity have exceeded the threshold. And the PMV indicator shows an slightly warm sign.

After reviewing the solution, an operator would check the operation status of the cooling and heating system, and change the operation setting to maintain comfort for occupants. Lighting energy usage has been detected in a vacant space on the southwest side of the eighth floor. This notification indicates energy waste in an unoccupied space of the building. Using the Hidden Markov Model, electric heat power usage is used for data input to determine whether the space is occupied or not.

It determines that energy use in a vacant space is a waste, and notifies users. Users can check the status of energy waste and turn off lights to save energy. In the case of a next day operation suggestion, the HVAC operation status of the day is analyzed to provide a suggestion for optimal operation of the next day. Data such as weather forecasts, time for pre-heating and cooling, and the start time of the applicable floor, is comprehensively analyzed.

The optimal operation guide is applied to individual air handling units controlled by the building automation system. This reduces the start-up time for the pre cooling and heating of air handling units, which helps save energy.

[END PLAYBACK]

JANGSOON KIM: How was it? As such, THYNC Energy offers optimal energy management for buildings through AI-based simulation using digital twins. THYNC Energy is a feature that conducts real-time monitoring of the current energy usage, and predicts and manages the future energy usage. And by continuously applying the operation guidelines from simulations to the actual facility operations, it helps reduce and optimize energy use.

In addition to the AI-based data analysis and simulations, we have developed our own physics analysis engine. This engine can be used in a digital twin for energy analysis. The central part of this THYNC energy is to provide energy management solutions. As you can see on this slide, it offers 10 different solutions. There are many focused on maintaining a moderate level of comfort for occupants, while reducing energy consumption.

And the simulation results are provided to users in real time. And the solutions include data collection, assessment of indoor comfort level for occupants, energy waste assessment, and solutions for optimal facility operations. Next, the second feature is THYNC Asset. Let's first watch the demo video.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- Digital twins enable efficient asset management within the building. Preview videos allow you to review the functions of each asset in advance. The asset list includes the equipment from the mechanical room and IoT sensors, or measuring instruments, throughout the building, which can be searched by name and object ID. You can intuitively comprehend the building's complex facility systems through a 3D model, and check the real-time status of each facility.

If you click on the specific equipment, you can see its locations within the building. For the management of a facility asset, its basic specifications, parts replacement, and repair history are recorded and managed through an asset history card. In addition, facility operations data linked to the building automation system can be collected in real time and monitored. This data helps identify the operation status and immediately detect issues for maintenance.

Also, the indoor environment data collected from IoT sensors can be monitored in real time. You can analyze the collected data to assess the indoor comfort level or identify IoT sensor issues. You can also check the real-time power usage and maximum power using power meters, and monitor energy usage by purpose.

[END PLAYBACK]

JANGSOON KIM: By using the second feature, like in the video, you can effectively manage building assets. THYNC Asset is a feature that integrates the information about building assets into digital twin and manages. By interlinking facility operation status and measurement data from IoT sensor and measuring instruments, you can conduct real-time monitoring.

Also, by checking abnormal conditions of assets in real time, you can immediately respond when issues happen. This will help preserve the performance of assets and prevent their aging. Next, the third feature is THYNC Space. It has two types of simulation, layout and evacuation simulations. Let's check out the layout simulation first.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- Layout simulation allows for optimal space planning. First, select the floor for space planning. Let's move to the fifth floor. A 3D model appears in the middle of the screen. On the left, there is a dashboard showing the object library and asset information. On the right, you can check layout results, such as layout status and lists. Once the number of workstations and meeting rooms is determined, the library can be arranged according to the layout structure, and space dimensioning is possible.

Results of layout can be viewed on the right. First, the available area means the size of an area where layout is possible relative to the total area. You can check the number of staff and meeting rooms according to the library layout, and predict energy usage accordingly. Also, the library material builds are automatically drafted. Once the simulation is complete, you can save the layout results.

[END PLAYBACK]

JANGSOON KIM: Just as you saw in the video, THYNC Space utilizes virtual spaces built in a digital twin to help solve problems and support decision making through role-based simulation. In particular, layout simulation allows you to easily and quickly set up an optimal space plan in a 3D space.

We expect that the layout simulation before moving will increase the operational efficiency for the users, including the building operations company, or general [INAUDIBLE] staff of the tenant company. The other type of simulation is evacuation simulation. Let's check out how it works.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- The evacuation simulation provides optimal routes and predicts evacuation time in case of an emergency. If you select the location of a fire, you can conduct a rule-based evacuation simulation. A fire alarm will show you the when and where the fire broke out. And if you click the emergency guide button, you can check Emergency Guidelines and locations of fire extinguishers. You can view an optimal evacuation route from the fire location and have a virtual tour along the route.

Next, let me show you a simulation that predicts the evacuation time based on each floor occupancy. You can set the occupant numbers for each floor, and run a simulation. Once everyone has gathered at a designated assembly point following the evacuation route, the simulation gets completed.

[END PLAYBACK]

JANGSOON KIM: Evacuation simulation provides optimal evacuation routes in case of an emergency and location information about fire protection facilities, emergency exits, and others. You can also conduct a simulation to predict evacuation time by setting different numbers of occupants on each floor. This can help users prepare against disasters in advance, contributing to securing safety for occupants. THYNC Space offers the ultimate space measurement through various simulations.

As it can satisfy the various needs of users, I think its expandability is infinite. Last but not least, the first feature is virtual tour. Let me first take you to a virtual tour exploring every part of [INAUDIBLE] building.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- Experience a virtual tour through digital twins. Users can choose between two different tour modes. An auto tour automatically takes users along the designated route. A free tour allows users to move freely as they wish. Users can select the space they want to visit, including office areas and mechanical rooms that are available with the program currently developed. Check the operating manual and click the Start button to start a virtual tour. This is a mechanical room where HVAC, cooling, and heating, and other equipment are located.

Various pumps are located here. And you can also see the motor control center. Additionally, there are four absorption chillers, and heaters, and a turbo chiller, which is the main cooling facility. As such, you can virtually experience the key spaces of a building without visiting them in person.

[END PLAYBACK]

JANGSOON KIM: The virtual tour feature can be used for building [INAUDIBLE] and guidance. Automatic and free tours from a first-person perspective, [INAUDIBLE] new employees or visitors to virtually explore major spaces of the building. This gives them great convenience as they can enjoy the tour without visiting in person. So now, how about joining us in bringing your own imagination of building operations into reality?

Samsung C&T THYNC can make this happen by synchronizing your thinking, and thereby create a paradigm shift in building operations. So that's all for my presentation today. Thank you.