GridAPM Demo: TransformerAPM Workbench
A guided product demo of TransformerAPM, the first GridAPM workbench for transformer diagnostics, AI-assisted reasoning, health indexing, and engineering reporting.
TransformerAPM is the first GridAPM product direction: a workstation-oriented evaluation concept for power transformer engineers who need to turn diagnostic evidence into reliable maintenance decisions.
The product is designed for utility engineers, asset performance teams, transformer service organizations, and industrial power users managing transformer fleets. It brings transformer diagnostic data, health index logic, AI-assisted reasoning, and engineering reporting into one review workflow.
What the demo shows
The demo is organized around the reliability workflow an engineer would naturally follow:
- Open a fleet view.
- Select a transformer.
- Review diagnostic evidence.
- Understand the health and risk drivers.
- Ask the AI layer to organize and explain the evidence.
- Verify assumptions and recommended actions.
- Generate a report for maintenance planning or internal review.
This is not intended to be a generic enterprise dashboard. TransformerAPM is centered on power transformer decisions.
Fleet and asset overview
The workbench starts with a fleet view that helps teams answer practical questions:
- Which transformers need attention first?
- Which assets are healthy, monitored, or elevated risk?
- Which evidence supports that classification?
- Which action should be reviewed next?
- Who approved the latest recommendation?
The fleet view is not the final decision. It is a triage layer that helps engineers move from a portfolio of assets to the evidence that needs review.
Diagnostic evidence
TransformerAPM is designed to organize multiple evidence streams:
- Dissolved gas analysis and gas generation trends.
- Oil quality and moisture records.
- Partial discharge and PRPD evidence.
- SFRA curve comparisons.
- Electrical test records.
- Bushing and tap changer snapshots.
- Acoustic emission or monitor data when available.
- Inspection notes and maintenance history.
The software should keep the origin and quality of each record visible. A laboratory DGA report, an online monitor reading, and a manually entered note should not be treated as identical evidence.
Health index and risk queue
The demo includes a health-index workflow that combines diagnostic inputs into an explainable asset-health view. The score is not the product. The explanation is.
For every risk queue entry, TransformerAPM should show:
- The evidence that changed.
- The diagnostic category involved.
- The trend direction and confidence.
- Related operating or maintenance context.
- Suggested follow-up action.
- Engineer review state.
That structure is more useful than a black-box score because it helps teams decide whether to monitor, inspect, test, plan maintenance, or escalate.
AI-assisted reasoning
The AI layer is framed as an engineering assistant. Its job is to organize evidence, detect patterns, explain risk drivers, and draft review-ready recommendations.
Examples of bounded AI tasks:
- Summarize DGA changes across the last six samples.
- Compare a PRPD signature with prior test campaigns.
- Identify whether a maintenance event changed the interpretation.
- Draft a report section with cited evidence.
- List assumptions that require engineer confirmation.
The engineer remains responsible for approval. This matters because transformer reliability decisions affect safety, reliability, outage planning, and capital allocation.
Standards-aware reporting
TransformerAPM is built around reporting, not only visualization. A useful report should include:
- Asset identity and configuration.
- Data sources reviewed.
- Diagnostic findings.
- Health and risk explanation.
- Standards-aware interpretation context.
- Recommended next actions.
- Confidence and uncertainty notes.
- Engineer signoff and decision history.
The goal is to reduce manual report preparation while improving traceability.
Pilot evaluation
An early pilot should be narrow and measurable. GridAPM can evaluate the workbench using a selected group of transformers and approved diagnostic records.
Good pilot questions include:
- Can engineers find and review evidence faster?
- Are recommendations traceable enough for internal review?
- Does the risk queue match expert judgment?
- Which data sources are most important?
- Which reports reduce manual work?
- What security and deployment constraints matter before scale-up?
Why offline-first matters
Utilities and industrial operators often work with sensitive operational data. An offline-first workstation path can help teams evaluate transformer AI workflows without forcing immediate enterprise integration or cloud data sharing.
Over time, GridAPM can support broader integration. The first product direction should prove value in the workflow that matters most: transformer evidence review and maintenance decision support.
Request a GridAPM pilot to evaluate TransformerAPM with a focused reliability workflow.
