AI Policy

Overview

Delta-X Research develops diagnostic software for transformer dissolved gas analysis (DGA). Our products, including the Transformer Oil Analyst (TOA) software and Monitor Watch online monitoring platform, incorporate analytical algorithms that assist utility engineers in interpreting complex gas data and making maintenance decisions.

This policy describes how computational methods, including those that may be described as artificial intelligence, are used within our products, and the principles that guide our approach.

How We Use AI and Advanced Analytics

Reliability-based DGA (R-DGA)

At the core of our products is R-DGA, a methodology developed over decades of research. R-DGA analyses the rate and history of dissolved gas generation in transformer oil to produce quantitative risk assessments, including:

• Cumulative Severity (CSEV): a metric reflecting the total fault energy deposited over a transformer’s operating history.
• Hazard Factor (HF): a metric reflecting the current rate of fault gas generation.

These calculations are based on peer-reviewed research and validated against international standards including IEEE C57.104 and IEC 60599. They are deterministic and reproducible — the same input data always produces the same result.

Fault Classification

TOA applies established fault classification methods, including the Duval Triangle and Pentagon, gas ratio analysis, and threshold-based condition assessment. These methods are drawn directly from published IEEE, IEC, and CIGRE standards and technical brochures.

Fleet Screening and Risk Ranking

TOA uses CSEV and HF metrics to rank transformers within a fleet by risk, helping asset managers prioritise inspection and maintenance resources. This ranking is based on each transformer’s individual DGA history and is updated automatically as new sample data is entered.

Our Principles

Transparency

We are committed to making our analytical methods understandable to the engineers who rely on them. The R-DGA methodology is documented in published research, and TOA’s outputs include the underlying data and calculations that produced each result.

Human Oversight

Our software is designed to support human decision-making, not replace it. TOA provides analysis, risk rankings, and alerts — but maintenance decisions remain with the qualified engineers and asset managers who understand the full operational context of each transformer.

Accuracy and Validation

Our analytical methods are grounded in peer-reviewed research and validated against decades of field data. We continuously refine our approach based on new research, standards updates, and feedback from utilities using our software.

Data Privacy

Transformer DGA data processed by our software is treated as confidential. We do not share individual customer data between organisations, use customer data for purposes beyond the contracted service, or sell data to third parties. See our Privacy Policy for full details on how we handle personal information.

What We Do Not Do

• We do not use generative AI to produce diagnostic results or maintenance recommendations.
• We do not use customer data to train machine learning models shared across customers without explicit consent.
• We do not make autonomous maintenance decisions — our software informs decisions made by qualified professionals.

Evolving Technology

As analytical methods and computational capabilities continue to advance, Delta-X Research will evaluate and, where appropriate, incorporate new techniques into our products. Any changes to our analytical approach will be communicated to customers and documented in product release notes.

Our commitment to transparency, accuracy, and human oversight will remain constant regardless of the specific technologies we employ.

Questions

If you have questions about how Delta-X Research uses AI or advanced analytics in its products, please contact us.