Transformer diagnostics beyond a single score. Connect DGA, oil condition, uncertainty and trends to a reviewable Action Plan, with methods appropriate to mineral oils and esters.
New release in preparationNew release in preparation. Contact us to discuss your data and assessment workflow.
Transformer duty, insulating fluid and sampling history affect how results should be read. Useful diagnosis starts by keeping that context with the evidence.
Ronin AI is transformer diagnostic and maintenance decision-support software. Deterministic engines interpret applicable diagnostic methods; probabilistic calculations show sensitivity to measurement uncertainty. Health assessment, historical trends and action planning add context. The trained ML model is one component, not the entire system.
Transformer fleets are managed on assumptions, not data. Too many assets, too little time, and diagnostic reports that sit in spreadsheets instead of driving decisions. Ageing infrastructure and a shrinking expert workforce have made reactive maintenance untenable. "Old" is not the same as "at risk", but most fleets still spend as if it were.
Lab reports arrive as spreadsheets and PDFs, then stall. The signal you paid for never becomes a decision.
Reading DGA correctly takes a specialist most teams no longer have in the room.
Budgets follow age or failure, not actual risk. The wrong transformer gets the attention.
Ronin AI checks which assessment route is appropriate for the fluid and available measurements. Missing gas data is not treated as evidence that the gas is absent. Where a method cannot support a conclusion, the result should state the limitation rather than manufacture certainty.
Evidence coverage, uncertainty and diagnostic classification are different quantities. They should be read together, not confused with a probability that the transformer will fail.
This existing illustration shows a condition-ordered asset list using sample values. It is not a live diagnostic calculation or a preview of the new release interface. A maintenance priority also requires criticality, operating duty and engineering review.
Illustrative example · not a live analysisAsset history helps turn individual laboratory results into a reviewable worklist. The illustration below belongs to the earlier presentation; it does not show the new Ronin AI interface.
Example assets ordered by assessed condition. Sample scores are out of 100, not failure probabilities.
Ronin AI brings data preparation, diagnostic methods, uncertainty, historical analysis and reporting into one workflow. The engineer retains responsibility for operating and maintenance decisions.
Upload DGA, oil and nameplate data. A health index per unit, one at a time or across a whole batch.
A per-asset report with the health index, the Duval Triangle for DGA and a plain interpretation.
Review triggered actions with supporting findings and references. Ronin AI adds dated planning and explicit cost assumptions where available.
Compare dated samples and eligible gas trends. Ronin AI distinguishes measured change, data gaps and changes of assessment method.
Organise assets into fleets and rank the whole population by severity, worst first.
Share assets with their health indices as A4 or A3 PDF, or export to Excel for your own reports.
Prepare the data, interpret it for the right fluid, examine its history and review the next actions. This is the workflow of the forthcoming Ronin AI release.
Upload supported laboratory files and let the parser prepare the measurements. Review values, units, sample dates and asset identity before submitting the analysis.
Check field mappings, fluid identity and flagged or missing measurements against the source report. Ambiguous layouts need correction; scanned documents depend on OCR availability and source quality.
Read applicable deterministic diagnostics alongside the condition assessment. Mineral oil, natural esters and synthetic esters follow dedicated assessment routes, not one universal model.
Where available, probabilistic calculations show how classifications respond to measurement uncertainty. The trained health-index model is used where applicable. Neither an index nor a simulation frequency is a calibrated probability of failure.
Examine the asset history, changes in diagnostic findings and eligible gas-generation trends. A single sample cannot show how the evidence has evolved.
Check sample dates, units, fluid and assessment method before comparing results. Missing measurements are not zero, and a change in method must not be mistaken for physical deterioration.
Examine proposed actions linked to diagnostic findings and evidence gaps. Review timing and stated cost assumptions with the engineer responsible for the asset.
The Action Plan supports maintenance review. It does not issue autonomous operating instructions or optimise decisions from a claimed probability of failure. Loading, criticality and site constraints remain part of engineering judgement.
Prepare an Asset Health Report or Action Plan for review, and export selected analysis data to PDF or Excel. Keep findings connected to the available evidence and its limitations.
Asset Health Report, Action Plan and selected analysis exports. A dedicated fleet health report is not included in the forthcoming release scope.
Two anonymised programmes, told as the customer stated the problem and what the health index made possible. Sectors only, no names.
The challenge: a generation utility with an obsolete generator step-up fleet could not tell which units actually carried the risk, and kept spending on the wrong ones.
A 4-year health-index trend flagged the risk in time for a one-week planned outage, instead of an unplanned failure and a two-year replacement lead time.
The challenge: a steel group running electric arc furnaces (over 1.1M tonnes a year) needed to know where the reliability budget should go across a large, mixed fleet.
Over 100 assets classified by risk and priority, so testing went where it mattered and planned outages fell.
Figures from real Seetalabs deployments, reported by sector and anonymised. Your results depend on your data and your fleet.
Colour always comes with a label and the value, so a report stays readable in greyscale and no button is ever mistaken for a "Critical" asset.
Ronin AI combines deterministic diagnostic engines and probabilistic uncertainty analysis with a trained health-index model where applicable. These components answer different questions: what the evidence indicates, how stable the classification is, and how the assessed condition changes. The method used and its limitations matter as much as the number.
The international reference for transformer health index and fleet prioritisation. RONIN's model is anchored to it, so the output maps back to a published method.
ISO/IEC 42001 provides a reference for AI management systems. Product features alone do not demonstrate organisational conformity; no certification is claimed here.
AI Act classification depends on intended purpose and deployment, including whether a system is a safety component. Ronin AI's separation of methods, inputs and limitations supports readiness work; it is not a declaration of conformity or certification.
The historical R² figure describes the fit of one trained health-index model to its evaluation targets. It is not the accuracy of the complete Ronin AI system, a probability of failure, or the chance that an individual diagnosis is correct. A higher R² does not automatically imply overfitting: that requires evidence from independent validation.
Ronin AI separates measured inputs, diagnostic rules and uncertainty assumptions. A result can be computable yet unsuitable for interpretation. Missing evidence, method disagreement and fluid-specific applicability therefore remain visible instead of being compressed into one reassuring score.
No dark patterns and no fine print. Here is exactly where your data lives, what it is used for, and how it is kept apart.
Your data sits on a VPS in the European Union and is handled under the GDPR.
Your DGA, oil and asset data are never used to train or tune the model. Your fleet stays yours.
No behavioural tracking. Account data is minimal: name, email and an optional company. No photos.
Transformer records can contain commercially or operationally sensitive information. Access permissions, customer separation and retention requirements should be agreed for the deployment; asset data should not be assumed non-sensitive.
The application uses HTTPS for data transfer. Hosting, access controls and the boundaries of backend services should be reviewed against your organisation's security requirements.
Access is firewalled and the scoring service is reachable only through the application, not directly.
If you already run oil treatment, testing or field service, RONIN becomes a diagnostic service you sell to your own clients. You bring the samples and the relationship; RONIN turns the data into a health index and a ranked fleet your customer can act on. A software add-on, no hardware to stock.
Plain-language verdicts for the industrial technician who is not a specialist; the depth on click for the asset manager and the engineer. You do not need to be a DGA expert to know which transformer is at risk.
Since 1999 I have had one job that changed name about ten times: I enter a field I do not know, learn it fast, and dig until I find what the specialists stopped seeing. Industrial compliance, then ecodesign, then artificial intelligence. In 2019 I pointed the same method at power transformers.
Court-appointed technical expert at the Court of Turin. Taught Industrial Standardisation and Engineering at Politecnico di Torino. Founder of Ecotp and Quantic. Independent expert in the Plenary of the European Commission's Code of Practice for general purpose AI.
Seetalabs is founder-led and lean by design, backed by a deliberate network of power engineers, transformer testing specialists and utility asset managers across Europe, Asia-Pacific and the Americas.
HAKUTAKU is the planned expert advisory layer for RONIN. It is not included in the currently available application. Its scope and release date will be confirmed after implementation and validation.
The planned role is to help engineers examine evidence and references. Diagnostic calculations remain the responsibility of the validated engines, not generated text.
Discuss the roadmapAsset-specific advisory is a development direction, not an available feature or a promise of autonomous operation.
Usage-based and transparent. No hardware to buy, no hidden line items, and no opaque enterprise quote you have to chase.
You pay for the predictions you run, the assets you actually assess. No hardware, no seats you do not use, no hidden costs.
The more assets you assess, the less each one costs. Scoring a whole fleet is cheaper per asset than checking a single unit.
Larger volumes, partner and white-label terms are available on request, on the same transparent, usage-based basis.
No "contact us" wall. You see how it works, and try it free, before anyone asks you for a number.
Discuss your transformer data and assessment workflow with us. Ronin AI is in final release preparation; existing customers can continue using their current application access.
Ronin AI: release preparation. Existing customer access remains available.
Onboarding assistant. No sales calls, the AI guides you.
LinkedIn profile