2/4 - Is the heavy investment required by Digital Twin worth it?

A look into simulation games and smart cities

Digital strategy as an overarching transformation pillar

Singapore has repeatedly been ranked among the best smart cities in the world. Their digital strategy revolves around several pillars with a wide variety of use cases:

• Healthcare: widespread use of video consultations, wearable IoT devices to monitor patients’ health, and development of AI powered alert systems for the elderly which monitors and learns peoples’ regular movements to notify a caregiver when urgent care might be required.
• Transportation: autonomous fleet to help the city’s elderly and disabled residents stay mobile, open data to facilitate transport planning, and contactless payment for public commuting.
• Public services: development of public apps for residents to report municipal issues, receive alerts, monitor their energy consumption, access public information, as well as a digital ID system.
• Education: initiatives (like AI for everyone and AI for industry) support the upskilling of 12?000 professionals and students in AI.?

?Another key pillar concerns urban planning and building, which is essential for Singapore’s development being a small city-state with a land area of just over 728km2. In this perspective, many entities including the Prime Minister’s Office and the Government Technology Agency, collaborated to create Virtual Singapore, a 3D city model and collaborative data platform that copies and simulates the features of the actual city – in short, a digital twin of Singapore.

?Along with corresponding data points, there are more than 100 TB of data covering anything from the tree coverage, the city’s evolving infrastructure to individual buildings – down to their roofs, facades, and windows. Virtual Singapore is crammed with real-time, dynamic data which can be used to simulate and test new solutions to urban planning problems in a virtual ecosystem. It is also semantically enriched, meaning that the twin understands real-world meaning and context of the information it processes – when looking at a representation of a building, it knows what kind of building it is, what kind of walls it has and so on.

?Virtual Singapore brings life to the game SimCity, but with actual, real-world data and buildings, it can provide opportunities far beyond simple entertainment. It notably offers 4 major features:

• Virtual experimentation – i.e., examining the coverage areas of 4G/5G networks, visualizing existing landscape against ongoing/future renovation projects.
• Virtual Test-Bedding – i.e., modeling, and simulating crowd dispersion to establish evacuation procedures during an emergency.
• Planning and decision-making – i.e., analyzing transport flows and pedestrian movement patterns, analyzing how ambient temperature and sunlight vary throughout the day to optimize urban planning.
• Research & Development – i.e., allowing researchers to innovate and develop new technologies or capabilities, adapting solar panels to customized buildings that have a higher potential for solar energy production ?

?Great challenges – Investment

Virtual Singapore was a significant investment, it cost $73m, and the development of the tool was spread over more than 5 years. The city also had to collaborate with various companies that brought in their own expertise. Singapore was able to develop the solution despite the associated high cost and complexity because there was a strong political drive – and that requirement stays true even for smaller-scale digital twin projects.

Indeed, deploying a digital twin requires a considerable investment, in workforce, in time, and in money. Digital twins are complex digital assets, that require capabilities in model-based systems engineering, modelling, data architecture, Artificial Intelligence, and Machine Learning to name a few.

Some technical aspects of a digital twin are shown below:

Developing such an asset requires either upskilling existing workforce, hiring externally, or as it happens most often, set up partnerships or M&A operations. Besides IT capabilities, digital twins also require a good command of operational technology (related to urban planning, architecture, decommissioning, etc.), are often multi-year projects, but if used well can become a foundational building block to widespread digital transformation. To reach that stage one should start by defining the requirements of the digital twin: what will be its end-applications and who will be its end-users? What data will they need to procure? How will it be accessed?

Once the requirements are detailed, every scenario needs to be modelized (examples include construction of a new building, road, facility; monitoring of energy consumption and 4G/5G coverage; visualization of bumps on the sidewalk to account for PRM needs, etc.) and account for a way to update each datapoint in real-time. Architecture also needs to be continuously updated to provide for new applications and features.

Digital Twins require an extensive upfront investment to cover the cost to hire or train key talent, as well as an anchored vision to be able to steer its build-up over the years.?

Even greater challenges – Data

Digital twins leverage a substantial amount of data, hence raising various issues related to data collection, security, governance, and standardization.

• Data collection is the first step towards building and maintaining a digital twin. It requires having cost-efficient means to collect (through IoT sensors and actuators, cameras, and satellite imagery) and store data (on the cloud or on premise). One key challenge is to restrict the number of datapoints to monitor and the number of scenarios covered to limit the cost of the exponential data aggregation. For instance, it is necessary to precisely define the level of details that ought to go into the virtual rendering. It is possible to model every door, window, lamppost, and tree, but the more there is, the more costly it becomes and the harder it is to maintain.
• Data security is also an essential challenge, as the digital twin aggregates a huge variety of data, some of which potentially confidential. A lot of entry points for the data being fed into the digital twin, increases the number of devices that can be vulnerable to cyberattacks. Maintaining state-of-the-art data security is essential, as erroneous or missing data might threaten the integrity of the overall system, as well as eroding trust in the insights generated. Cyberattacks would also be detrimental as they can lead to sensitive data leaks (governments need to create conditions for such twins to respect the privacy of citizens and confidentiality of enterprise data IP).
• Data governance needs to be very well thought through from project inception. Since the digital twin aggregates a huge variety of data, it is necessary to define who will feed data into the model, who will be able to access what type of data, and at what price if applicable.
• Data standardization is necessary to simplify data architecture of the digital twin, to streamline access to it and to promote interoperability. ?Such standards are being created at company- and industry-levels but there is currently no clear consensus on the matter.

Deploying a digital twin requires upfront investment and incremental scaling

Singapore has developed a national strategy to become a smart city-nation and Virtual Singapore is a key element supporting this. But given the significant advantages they provide, even smaller-scale projects may require the use of digital twins. Regardless of project size, similar challenges are to be expected in terms of digital capabilities, upfront required investment, and time-to-market. Considerable CAPEX needs to be invested early on, which will eventually be offset by lower long-term CAPEX and OPEX, improving overall ROI. Starting small, with limited reach and use cases at first, is recommended, but unlocking maximal value only comes with scaling and adding complexity to the twin, by creating a platform logic.

Authors

Vanessa Lyon, Managing Director and Senior Partner, Mika?l Le Mouellic, Managing Director and Partner, Jean-Christophe Laissy, Partner and Director, Laurent Alt, Associate Director, Alexandre Toureh, Senior Associate, Victoria Guérendel, Associate

Related link