Around 30 years ago, the video game publisher, Maxis, released SimCity. As the name implies, the goal was to create and run a virtual city. From building roads to power plants, adding different housing zones and even constructing an underground waterpipe system, every detail had to be meticulously thought out. For a city to thrive, it had to be sustainable, safe and economically sound. Gamers unknowingly became urban planners who had to get to grips with tax, infrastructure, gentrification and maintenance.
Everything had to work because when disaster struck, which in the world of SimCity could be earthquakes, fires or even UFOs, you needed a resilient backup plan, income and skilled virtual inhabitants. “That is, for me, a digital twin in plain English,” says Johan Potgieter, cluster industrial software lead at Schneider Electric. “To see the future and not make costly mistakes.”
“The vision is really like SimCity. In other words, a gamified representation of an asset,” says Douglas Ackerman, director of digital engineering at WSP in Africa. “It’s a virtual representation in three dimensions in an environment that has real-time metrics.” Ackerman says that in a working digital twin, you can move around assets, get insights into how these assets are functioning and use the data to make predictions. But to do this, the system has to have IoT sensors or cameras that continuously feed information back to the twin so there’s an up-to-date view of what’s happening in real-time. “It’s not only how things are performing now, but looking at trends to help the client be proactive in addressing possible issues in due course,” says Ackerman.
Digital twins are moving towards more autonomous behaviour, with systems capable of adjusting themselves.
Johan Potgieter, Schneider Electric
According to Fortune Business Insights, the digital twin industry will be worth $384.79bn by 2034. It’s tech that has already been adopted by different industries, like agriculture, smart cities, healthcare and manufacturing. In the property sector, there are a number of existing building management systems (BMS) like Honeywell Enterprise Buildings Integrator and Siemens’ Desigo CC that provide real-time telemetry. Ackerman’s area of speciality is transportation and infrastructure (TI), which is, he says, the least mature field. “This puts us in a pure R&D space where we have to develop the software from scratch, filling the massive gaps left by vendors that simply haven’t come to the party.”
In TI, digital twins are challenging for many reasons. “Think about 500km of rail line. How do you build a SimCity version of that?” he asks. Another issue is interoperability. When systems don’t speak to each other, bridging the gap between disparate and often proprietary systems can be technically challenging. For Ackerman, this means standardising protocols, something that’s already happening in BMS. And then there’s maintenance – life cycle management. Digital twins are complicated systems with many moving parts.
They need to be built, deployed and maintained. “The physical assets are in flux to a certain extent, which means the system will need to expand and improve over time,” says Ackerman. “But what we’ve found is that once a client has their hands on a proper SimCity, they want more. It leads to a lot of downstream work.”
Think about 500km of rail line. How do you build a SimCity version of that?
Douglas Ackerman, WSP in Africa
WSP in Africa developed a digital twin of South32’s manganese rail infrastructure in the Northern Cape, using bi-weekly drone data and 3D modelling to track construction progress against the approved design.
A digital twin exposing infrastructure upgrade requirements is exactly why this technology makes sense. In Africa, Potgieter says, there’s a tendency to run everything until it breaks and hope for the best. There are also many businesses that choose cheaper brands or grey imports over premium products and, as a result, assets like IoT devices have to be changed more often. Digital twinning helps with life cycle management as you can change parameters in the field. “If I run an asset at 70% normally and change it to 92% to boost production for a certain month, how long can I run that asset before I have to maintain it?” asks Potgieter. “Everything costs money.”
THE COLOUR OF THE PLAYING FIELD
For digital twins, whether a project is greenfields or brownfields makes a big difference to how complex the work becomes. Greenfields projects are simpler because digital twins can be planned from the start, with control over design data, asset information and sensors as the project is delivered. Brownfields projects, on the other hand, are harder to manage. They often involve incomplete records, legacy systems and infrastructure that was never designed to be digitally connected. In many cases, teams first have to capture what actually exists, using tools like drones or LiDAR to scan assets and rebuild accurate geometry, before performance data can even be layered on.
Johan Potgieter, cluster industrial software lead at Schneider Electric, says greenfields are the exception, not the rule. Of course, new, purpose-built sites are technically simpler for digital twins, but in South Africa, he says that barring datacentres, very little is being built. “I have a water board that still works on eight-bit system. And just to change that out will take us five years.”
The good news is that operational value can usually be seen within the first few months as teams begin identifying unusual behaviour, predicting component failures and adjusting maintenance plans accordingly. In sectors like mining and heavy industry, that kind of visibility can feed into safety planning, helping operators understand how far they can push an asset before it becomes a risk. Instead of waiting for something to fail, engineers can watch patterns build up over time – temperature spikes, vibration anomalies, subtle shifts in performance – and intervene before those warning signs turn into incidents. “A system like this is a safety enabler of the highest order,” says Ackerman.
This early‑warning approach is now being improved by vision AI, which applies pattern recognition to video and turns it into another data source for the digital twin. Instead of relying on someone sitting in a control room and watching a camera feed, the system analyses footage automatically and learns what normal movement looks like onsite. When something changes, like a vehicle entering a zone that should be empty, equipment cycling more often than usual, or a sequence of actions that has previously led to an incident, the event can be flagged before it becomes a problem. “While you’re looking elsewhere, there’s something with full eyes on what is going on,” says Ackerman. The twin remains a simulation, but vision AI gives it a constant view of behaviour on the ground, adding another layer of safety and situational awareness.
Video can tell you that something unusual is happening, but on its own, it does not tell you where that risk sits in the asset itself. The digital twin’s role is to place behaviour picked up on camera in context, tying it back to the physical world it represents.
“You need three things,” says Ackerman. “Three-dimensional geometry, metadata about the geometry and real-time sensory data.” When these three elements are combined, a digital twin starts to function like the city building games we grew up with, where all the systems were easily accessible. In the real world, most industries have never had that kind of consolidated view. Production metrics sit in one system, maintenance logs in another, safety observations in a third. With a digital twin, all of those pieces can come together in the same environment. “And now, digital twins are moving towards more autonomous behaviour, with systems capable of adjusting themselves, identifying issues automatically and learning from ongoing operations,” says Potgieter.
It’s not a game, but the logic is the same – every move counts, only this time the consequences are real.
STEP BY STEP
One of the biggest misconceptions around digital twins is how they work in an African context. Many assume you can simply switch one on, and start optimising, but without stable power, reliable connectivity, access to cloud, AI and the right skills in place, the whole idea collapses. “Everyone says they love AI, but nobody knows the containment, especially in a digital twin environment, that must be practiced,” says Johan Potgieter, cluster industrial software lead at Schneider Electric. At his company, the process of implementing a fully functional digital twin happens in stages. “There is construction, building and planning. And then there’s the lifecycle of that specific business,” he says. “And the most important aspect that has surfaced in the last 10 years is the decommissioning of that asset in order to assess environmental impact.”
1. Site design
It begins with a piece of land. Using building management software from RIB Software, Schneider sketches the planned future state on top of it – buildings, foundations, cables, pipes and flows of people. The model pulls in wind, temperature, local hazards and CO2 targets to find the smartest, cleanest way to turn that ground into a working site.
2. Going electric
Once the virtual site exists, the electrical layer is built in software. An electrical digital twin traces how power will flow from the grid, where demand spikes, how overloads or penalties might hit the bottom line. Before a single cable is ordered, the model reveals the most resilient and cost‑effective way to energise the operation.
3. Industrial flow
With land and power in place, the industrial heartbeat is added: conveyors, pumps, crushers, valves, instruments. Different brands and configurations drop into the model like pieces in a strategy game. The twin estimates output, energy use, wear and tear and total cost of ownership, long before any equipment is installed.
4. Virtual vs real world
Once the mine, factory or water board is running, real‑time data starts to feed the model, turning it into a live operational twin. Every change in the field – a speed adjustment, a new setpoint, a maintenance action – can first be tried virtually. The twin becomes a safe space for “what if” experiments that predict production, safety and asset‑life effects.
5. Life cycle
The same model stretches from when the first concrete is poured to the project’s completion. It helps plan maintenance, upgrades and, eventually, decommissioning: environmental impact, carbon footprint, clean‑up costs and even what happens to people and skills when operations wind down. The digital twin becomes a way to design not just a plant, but a legacy.
* Article first published on www.itweb.co.za
Share
