Mode Physical (P) — Deliverables & Testing¶

Phase I — Design (2025–2026):

Monthly reports: due on or before the final day of each month starting in October.
Phase I Design Report: due 31 Dec 2025, 23:59 Local Time; use the naming convention “2025 Phase I Design Report”.
Optional Design Video: ≤ 5 minutes, due by 23 Jan 2026.
Final 2026 results: published June 2026. See 2026 Results.

Phase II — Implementation and testing:

Mode Physical (P): model and controls development and rig construction; Phase II test occurred May/June 2026.
Pre‑recorded team presentation: submitted ~1 week before the test; judges prepare Q&A from it.

Edge‑deployable models

Expectation: focus your effort on generating/fine-tuning data and optimizing compact multimodal models.
Inference‑only, offline, JSON outputs (e.g., Explain‑This‑Event or Rig‑State Textualizer).
Small‑model, edge‑deployable focus: teams may prototype with larger cloud‑hosted models during development, but final deliverables and performance are evaluated using locally run or on‑rig small models under limited inference hardware.
Suggested caps: ≤2 vCPU or ≤20 W GPU, ≤4–8 GB RAM (non‑LLM) and ≤8 vCPU or ≤1,150 W GPU, ≤16–64 GB RAM (LLM); ≤15 min total inference runtime. Rig compute otherwise open. All AI runs must be offline during judging; no PLC/actuator writes.
Custom task option: teams may propose a custom offline JSON task with objective ground truth and the same interface for pre‑approval (recommended deadline: 2026‑02‑01). Approved tasks use the same AI sub‑rubric and scale to +10.

Required artifacts (Magnetic surveying excerpts):

Definitive directional survey (TXT/LAS/CSV) including header info such as school name, survey date, well center (WGS‑84), TVD reference (above block), grid convergence, geomagnetic model, magnetic declination, total azimuth correction, field dip/strength references, and the error model used.
Autonomous survey process and qualification; provide raw survey data logs with timestamps, measured depth, downhole sensor values (axes + qualifier values), and survey acceptance indicators.
Telemetry dataset: fixed‑schema drilling_timeseries.csv (≥1 Hz; D‑WIS names) with semantic headers and units for apples‑to‑apples scoring.
Static logs (suggested): LAS 2.0 wellname_logs.las with required curves and metadata.

Testing notes:

Closed‑loop control using downhole data is REQUIRED; failure to integrate downhole measurements into the controller is considered failing for 2024.
Prepare a test matrix, data handling and visualization plans, and calibration/verification procedures for sensors and instrumentation.

Mandatory Rule

Interop demo is separate and read‑only/shadow; teams MUST NOT write to PLCs/actuators during checks.