eRTMAC-NWIS (Nearby Wells Intelligence System): An AI-Powered Offset Well Knowledge and Decision Support Platform for Drilling Operations
Oil India Limited · Smart Automation · Software
One of the better-specified industry problems in this block — build on the public Volve or Norwegian well reports, make the depth-triggered alert deterministic, and be candid that the corpus is foreign.
What it actually is
When a drilling crew hits trouble at a certain depth, the odds are a nearby well hit the same trouble years ago — but that knowledge sits in old completion reports and in the memory of whoever was on site. The ask is a platform that reads those historical documents, puts the nearby wells on a map, and warns the active crew before they reach a depth where neighbouring wells ran into problems.
What to build
A decision-support platform that ingests historical well completion and daily drilling reports through OCR and information extraction, structuring the events that matter — mud losses, kicks, stuck pipe, casing programmes, cementing practice, formation tops — into a queryable per-well record; a map showing offset wells within a radius of the active well with their event history; a correlation view aligning drilling parameters and formation markers across neighbouring wells on a common depth axis; and a proactive alert that fires as the active bit approaches a depth or formation where nearby wells logged an incident, naming which well, what happened and at what depth.
Smallest thing that wins the room
Advance the active well's bit depth on the timeline and watch an alert fire — three offset wells within two kilometres logged mud losses entering this same formation, each named with its depth and what the crew did about it.
How crowded this one gets
A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.
Quieter than 72% of the 226 · #63 of 226 by expected field
A normal-sized field. Your idea has to be good, not miraculous.
Why: company-sponsored statements drew the smallest fields of all.
This is a guess, not a fact
Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.
The scores
The number is the shorthand. The line under it is the reason.
Acceptance potential
4/5Genuinely underrated — the requirement is precisely enumerated, the demo is immediate, and drilling operations intimidate most student teams so the field will be thin, though you must be upfront that your corpus is foreign well reports rather than Oil India's own.
Feasibility
3/5The extraction and mapping stack is standard, and public well report archives exist to work from — Equinor's Volve dataset and the Norwegian and UK public well repositories publish real completion reports — but no Indian field data is available so your corpus will be foreign wells with different formation names.
Innovation scope
4/5How you extract loosely structured operational events from decades of inconsistent report formats, and how you decide that an offset well's experience is relevant to the active well's current position, are both genuinely open problems.
Clarity
5/5The description numbers the four current gaps and then numbers the required capabilities against them, naming the exact event types to correlate — mud losses, kicks, stuck pipe, casing, cementing — so nothing about the deliverable is left to interpretation.
Effort
HeavyThe document extraction pipeline, the structured event store, the geospatial correlation view and the depth-triggered alerting are four connected pieces, with extraction from messy scanned reports absorbing most of the time.
Demo-ability
EasyA map with offset wells and an alert firing as the bit approaches a known trouble zone tells the whole story in one interaction, and the value is obvious without any drilling knowledge.
In its favour
- Green flag: Equinor's Volve dataset and the Norwegian and UK public well archives publish genuine completion and daily drilling reports, so you can build on real documents rather than invented ones
- Green flag: The alert is deterministic — a spatial query plus a depth comparison — so the headline feature works reliably without depending on model accuracy
- Green flag: The description names the exact event categories to extract, which gives your extraction schema a defensible origin
- Green flag: Institutional memory loss is a problem any judge recognises instantly, so the pitch needs no setup
Against it
- Red flag: No Oil India well data is public, so your corpus will be North Sea wells with different formations and nomenclature, and you should say so rather than implying Indian coverage
- Red flag: Historical drilling reports are inconsistent scanned documents spanning decades of formats, and extraction accuracy on the oldest ones will be poor
- Red flag: Deciding which offset well is genuinely analogous requires geological judgement — proximity alone is misleading when formations pinch out — and a drilling engineer will press exactly there
- Red flag: A false alert that cries wolf at every depth trains the crew to ignore it, so alert precision matters far more than recall here
What you will be writing
- PaddleOCR / docling for scanned report ingestion
- LayoutLM or LLM-based event extraction
- PostGIS offset-well spatial queries
- Depth-indexed time series (TimescaleDB)
- React + Mapbox well map with depth correlation panel
- Rule-based proximity alerting engine
- Drilling operations
- Document intelligence
- Geospatial decision support
Prior art to read before you start
offset well knowledge retrieval · operational event extraction from reports · depth-triggered proactive alerting
Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.