In the modern mining industry, the term digital twin is very popular. Yet its meaning is often diluted. It is casually applied to high-resolution 3D visualizations, integrated data platforms, or real-time monitoring dashboards. These tools can be useful. In high-consequence geotechnical environments, however, they are not enough.

​

They are useful digital tools. They can show the state of the ground, but they do not necessarily help manage the risk of the ground.

​

That distinction matters. Most slope failures are not caused by a lack of modelling. Mines already have models, instruments, alerts, and dashboards. The deeper problem is usually the widening gap between what the system shows, what engineers believe, and what the operation is prepared to do.

​

If the concept of a geotechnical digital twin is to be useful, it must be defined with greater precision.

​

Three Levels of Digital Maturity

​

A practical way to clarify the concept is to distinguish three levels of digital representation.

​

Digital Model

A static representation of the system — geological, hydrogeological, or numerical. It may be sophisticated, but it remains a snapshot defined by fixed assumptions. It does not evolve as reality unfolds.

​

Digital Shadow

A model connected to live monitoring data. The representation updates as conditions change, improving visibility. But it remains largely passive. The underlying interpretation, parameters, and uncertainty are rarely revised in a systematic way.

​

Digital Twin

A continuously updated, uncertainty-aware system that supports decisions. It integrates state, prediction, and explicit decision logic. It does not merely display the ground — it continuously reinterprets the ground as new evidence arrives, in a way that makes decisions harder to ignore.

​

The difference is not the amount of data. It is whether the system is allowed to change its own interpretation.

​

A model represents.
A shadow reflects.
A twin updates and supports action.

​

From Representation to Decision

​

What makes a twin meaningful is not graphical fidelity. It is the logic underneath.

​

A true twin requires a digital thread — a continuous, traceable flow linking:

• observation

• interpretation

• modelling

• uncertainty

• and decision

• ​

Without that thread, information remains fragmented. The mine gains visibility, but not coherence.

​

At the center of this system is a loop that has always defined good geotechnical practice:

Observe → Interpret → Update → Decide

​

A digital twin does not invent this loop.
It makes it continuous, explicit, and auditable.

​

The Engine: Updating State and Uncertainty

​

One of the most rigorous ways to enable this continuity is Bayesian data assimilation.

​

In this framework, prior understanding (models, assumptions, uncertainty) is continuously combined with new observations (monitoring data) to produce an updated estimate of the system state.

​

The important point is practical:

New data does not just refresh a display.
It recalibrates the interpretation itself.

​

A shadow shows change.
A twin learns from change.

 

At the same time, Bayesian updating is not the only implementation pathway. The objective is not the method itself, but the ability to continuously update both state and uncertainty.

​

What Changes in Practice

​

When uncertainty is updated rather than hidden, decision-making improves.

• Stabilisation becomes more targeted

• Monitoring becomes more interpretable

• Operational constraints reflect current confidence, not outdated assumptions

• Conservatism becomes selective, not blanket

​

The role of the geotechnical team shifts accordingly.

​

The question is no longer: Has the model been updated?

It becomes: What has changed — and does it require action?

​

Where Most Systems Fail

​

Many current systems still fall short.

​

They provide a window into the slope’s behaviour, but no bridge to action.
They record movement, generate alerts, and produce clean visuals.
But the connection between signal and decision remains weak.

​

They improve visibility.
They do not reliably improve response quality.

Most are still digital shadows with better interfaces.

​

A Necessary Caution

​

A digital twin does not eliminate model risk.

​

If the governing mechanism is misunderstood, if key observations are missing or biased, or if the update framework is built on the wrong conceptual model, the system can become a highly persuasive way of being wrong.

​

A digital twin can increase confidence faster than it increases understanding.

​

Continuous updating does not guarantee correctness.
It guarantees consistency with the assumptions embedded in the system.

​

The value of a twin is therefore not certainty.
It is faster reinterpretation, clearer uncertainty, and more disciplined decisions.

​

State of Practice

​

Geotechnical digital twins are emerging, but unevenly.

​

There are real applications where integrated monitoring and modelling systems have delivered operational value. But most implementations remain closer to advanced digital shadows:

• strong in visualization and data integration

• weaker in continuous state updating

• limited in explicit decision logic

​

The concept is real. Full implementation is still evolving.

​

What a Digital Twin Actually Is

​

A geotechnical digital twin is not:

• a high-fidelity 3D model

• a monitoring dashboard

• a passive mirror of the excavation

​

It is:  

A continuously updated, uncertainty-aware system that supports better decisions as reality evolves.

​

It does not replace engineering judgment. It makes that judgment more current, more traceable, and more defensible.

 

Closing

​

The goal is not to build a better model of the ground. The goal is to reduce the distance between evidence and action. In high-consequence geotechnical systems, that is what matters.

 

A digital twin is valuable only if it changes what engineers and operators actually do as new evidence emerges.

That is the level at which this problem needs to be addressed — linking observation, interpretation, modelling, and uncertainty directly to decision-making.

​

This is the space in which VSKY.GEO operates.