Smart AI Mesh topology linking wellhead wireless sensors, MF Gateways

Engineered for hazardous, unmanned offshore environments, the MF Gateway product line combines low-power mesh networking with an on-demand GPU Field Gateway to deliver autonomous, zero-cloud LLM intelligence at the edge.

MF Gateway

ATEX Rating: Ex Zone 2: Ex d ec mc ia IIB+H2 T4 Gc
Ingress Protection: IP66 Enclosure
Power System: Solar powered + 3-day battery backup
Power Draw: 12 W Average / 15 W Peak
SCADA Storage: 200 GB Local Offline Storage
Protocols: LoRaWAN Sensor Ingestion & Mesh Routing
GPU Compute: NVIDIA Ampere Architecture (2048 CUDA Cores, 64 Tensor Cores, 275 TOPS)
CPU Compute: 12-core Arm Cortex-A78AE v8.2 64-bit CPU (3 MB L2 + 6 MB L3)
System Memory: 64 GB 256-bit LPDDR5 (204.8 GB/s Bandwidth)
Smart Power Management: 12 W dormant / 75 W peak.

On-demand GPU execution allows operation on a compact 300 W solar panel (versus 900 W required for continuous load).

Supports open-source LLMs up to 70B parameters (32B recommended for low￾latency interactive tablet chat).

Download Datasheet

Use Case: Data OptimizedPlug and Abandonment ofOffshore Unmanned Wells

Read Use Case

White Paper:Annulus Pressure Gauge Monitoring and Agentic AI Anomaly Analysis

Read White Paper

Key Architectural Features of the MF Gateway:

  • Dual Gateway Architecture:
  • MinFarm Edge Gateway: The base node provides an onboard LoRaWAN Network Server (LNS), local storage (up to ~200 GB), local anomaly detection, and mesh summary relaying across inter-gateway hops of up to 10 km.
  • MinFarm Field Gateway: Serves as the mesh’s single point of access, hosting the dashboard web app, SCADA integration APIs, and a dedicated, on-demand GPU compute unit housing the system's AI models and agent orchestrator.
  • Autonomous & Conversational AI (LLM Integration): When an anomaly is detected, the Field Gateway immediately alerts engineers and powers on its GPU to run an AI investigation. Engineers can interact with the AI assistant using plain language via ATEX-rated tablets in the field to query real-time data, inspect historical trends, or commission deep-dive investigations.
  • Energy-Efficient On-Demand Compute: To respect strict offshore power budgets, the Field Gateway keeps its GPU unit powered down in normal operation, waking it only when an anomaly is flagged or a user initiates an AI chat session.
  • Hazardous Environment Certified: Enclosures are sealed to IP66 standards and certified for ATEX Zone 2 deployment (with connected sensors certified for Zone 0/1). Gateways run autonomously on solar power with a minimum 3-day battery backup.
  • Seamless SCADA & External Integration: Exposes standard industrial interfaces (Modbus, HART, 4-20 mA support) and feeds well integrity data directly into central SCADA estates or third-party enterprise tools.

Reach 100% of Remote Wells

• Network capacity per MF Gateway: 1000
instruments, e.g. pressure gauges
• Data sampling intervals: 1 minute to 24 hours (up to 5 year
battery life)
• Transmission reliability: 99% with local edge protocol
management

Remote Well Integrity Management Dashboard

• Real-time Data Visualization
• Alerts & Notifications
• Historical Data Analysis
• Geospatial Mapping
• User Management
• Data Export
• Terminal Management
• LoRaWAN device Management