Certification of Digital Twin Models for Regulatory Use – Detailed Explanation With Case Laws

1. Meaning

A Digital Twin is a computer-based representation of a real physical asset or system. In the energy sector, a digital twin can represent:

a power station;

transmission line;

substation;

electricity distribution network;

battery-storage facility;

renewable-energy plant; or

an entire electricity grid.

Certification of Digital Twin Models for Regulatory Use means establishing legal and technical procedures to determine whether a digital twin is sufficiently accurate, secure, reliable and transparent to be used by a regulator or public authority for making or supporting regulatory decisions.

For example, a regulator could use a digital twin to simulate whether a proposed transmission project will create congestion or whether a power plant will affect system reliability.

2. Why Certification Is Necessary

A digital twin is not the physical infrastructure itself. It is a model of that infrastructure.

If the model contains incorrect information, the regulator may make an incorrect decision.

For example:

Physical grid → incorrect sensor data → inaccurate digital twin → incorrect simulation → regulatory decision based on unreliable information.

Therefore, certification must establish confidence in the relationship between the physical system and its digital representation.

3. Main Certification Requirements

A regulatory digital twin should generally satisfy several requirements.

Accuracy

The model should represent relevant physical characteristics accurately.

Validation

Its predictions should be tested against real-world observations.

Data Integrity

The data entering the model must be reliable and protected against manipulation.

Traceability

Regulators should know:

where the data came from;

when it was collected;

who changed it; and

which model version was used.

Cybersecurity

The digital twin must be protected against unauthorised modification.

Reproducibility

Another qualified expert should be able to examine the methodology and reproduce important results.

4. Difference Between Ordinary Software and Regulatory Digital Twins

A digital twin used for internal engineering may have relatively limited legal consequences.

A digital twin used by a regulator is different.

Suppose a regulator uses a model to determine whether a proposed electricity project should receive approval.

The model may influence:

environmental decisions;

network-access decisions;

infrastructure planning;

tariffs;

reliability assessments; or

licensing decisions.

Therefore, certification must address administrative-law requirements, not only engineering accuracy.

5. Administrative Law

The South African Constitution requires public administration to be lawful, reasonable and procedurally fair.

Section 33 is particularly important, together with the Promotion of Administrative Justice Act 3 of 2000.

If a digital twin substantially influences a regulatory decision, affected persons should be able to understand the important basis of the decision sufficiently to challenge it where appropriate.

In Affordable Medicines Trust v Minister of Health, the Constitutional Court emphasised the importance of lawful and properly structured regulatory discretion.

In Democratic Alliance v President of South Africa, the Court reinforced the requirement that exercises of public power must satisfy rationality.

These cases are not digital-twin cases, but their principles are highly relevant.

A regulator cannot simply say:

“The digital twin produced this result.”

The authority remains responsible for the legality and rationality of the final decision.

6. Model Validation

Certification should require comparison between the digital twin and the physical asset.

For example, a transmission-network twin might predict:

electricity flows;

voltage levels;

congestion;

equipment temperature; and

failure risks.

These predictions should be compared with actual operational data.

Certification should establish acceptable levels of:

accuracy + uncertainty + error tolerance.

Importantly, certification should also identify circumstances in which the model becomes unreliable.

7. Model Updating

Energy infrastructure changes continuously.

A digital twin may become inaccurate because:

equipment is replaced;

new generation is connected;

demand changes;

network topology changes;

sensors are upgraded; or

operating rules change.

Therefore, certification should not be permanent.

A suitable system is:

initial certification → validation → deployment → continuous monitoring → periodic recalibration → recertification.

A major change to the underlying infrastructure should trigger a review.

8. Evidence and Regulatory Decisions

Digital twins can provide powerful evidence, but their results should not automatically be treated as unquestionable facts.

Regulators should distinguish between:

measured physical information;

modelled information;

assumptions;

forecasts; and

uncertainty.

This distinction becomes particularly important where digital-twin results are used to justify:

infrastructure investment;

electricity tariffs;

environmental approvals;

grid-access decisions; or

enforcement action.

9. Electricity Regulation in South Africa

The Electricity Regulation Act 4 of 2006 provides the principal statutory framework for electricity regulation.

NERSA may need increasingly sophisticated modelling tools to assess:

electricity demand;

generation adequacy;

transmission requirements;

tariff applications;

system reliability; and

market conditions.

Certification of digital twins can therefore improve regulatory decision-making, but it should operate within existing statutory powers.

A digital twin should support the regulator's statutory function; it should not create a new regulatory power by itself.

10. Electricity Reliability

Eskom Holdings SOC Ltd v Vaal River Development Association is relevant by analogy because the Constitutional Court considered electricity supply and the public consequences of electricity interruptions.

Digital twins could help regulators evaluate:

system resilience;

congestion;

equipment failure;

maintenance requirements;

extreme-weather risks; and

future electricity shortages.

But certification becomes particularly important because an inaccurate model could create false confidence about grid reliability.

11. Environmental and Climate Regulation

Digital twins can also model environmental effects.

For example, a digital twin of an electricity system may estimate:

greenhouse-gas emissions;

water consumption;

air pollution;

climate vulnerability; and

environmental impacts of alternative infrastructure.

In Fuel Retailers Association of Southern Africa v Director-General: Environmental Management, Mpumalanga, the Constitutional Court emphasised integrated environmental and socio-economic decision-making.

In Earthlife Africa Johannesburg v Minister of Environmental Affairs, climate-change impacts were recognised as relevant to environmental decision-making concerning energy infrastructure.

These are analogical authorities, not direct digital-twin cases.

They support the principle that technological models should provide reliable evidence for environmental decision-making rather than becoming a substitute for legal judgment.

12. Transparency and Public Participation

Where a digital twin substantially influences an important regulatory decision, transparency becomes significant.

Authorities should, where legally appropriate, disclose:

the purpose of the model;

important assumptions;

relevant data sources;

uncertainty levels;

validation methods; and

material limitations.

Trade secrets and cybersecurity concerns may justify protecting some technical information, but confidentiality should not make meaningful regulatory review impossible.

13. Cybersecurity

Digital twins are connected to real infrastructure and therefore can become attractive targets for cyberattacks.

A malicious actor could manipulate the digital model so that the regulator receives false information.

Certification should therefore include:

access controls;

encryption;

authentication;

secure data feeds;

tamper detection;

audit logs;

backup systems; and

incident-response procedures.

South Africa's Cybercrimes Act 19 of 2020 forms part of the broader legal environment concerning cyber-related conduct.

14. Conclusion

Certification of Digital Twin Models for Regulatory Use ensures that digital representations of energy infrastructure are sufficiently reliable to support public regulatory decisions.

A strong certification framework should provide:

model accuracy → independent validation → data integrity → uncertainty assessment → cybersecurity → version control → auditability → transparency → continuous monitoring.

The cases Affordable Medicines Trust, Democratic Alliance v President, Eskom v Vaal River Development Association, Fuel Retailers Association, and Earthlife Africa provide useful principles by analogy. There is limited direct South African case law specifically dealing with certified digital twins in energy regulation.

The central principle is that a digital twin can assist regulatory judgment but cannot replace legal accountability. If a regulator relies on a digital model, it must still demonstrate that the decision was made under lawful authority, was rational and procedurally fair, and was based on sufficiently reliable evidence. Certification therefore provides a bridge between advanced digital modelling and traditional administrative, electricity and environmental law, making digital-twin technology more trustworthy for long-term energy governance.