AI correction algorithms that dramatically improve the precision of subsea power cable surveys — enabling faster inspections, lower costs, and better maintenance decisions.
Subsea power cables are critical infrastructure — transporting electricity between wind farms and the grid, across hundreds of kilometres of seabed. But locating them accurately is surprisingly difficult.
Existing magnetometer-based survey systems contain systematic scale-factor errors that distort measurements. During a campaign, deviations of 50 to 150 centimetres are common — sometimes more.
These inaccuracies have serious consequences:
Cable owners want confidence that they know exactly where their infrastructure is. Survey companies want a tool that reduces re-survey costs and improves their competitive position.
The gap between measured and actual cable position — caused by scale-factor errors in magnetometer readings.
Tarucca, in partnership with Seekable B.V. (Amsterdam), is developing an AI-driven correction model for subsea cable survey data. Seekable conducts subsea cable surveys using autonomous underwater vehicles (AUVs) equipped with magnetometers. Tarucca analyses the electromagnetic field data and builds machine learning models that identify and correct the scale-factor errors that cause systematic measurement deviations.
The correction model accounts for cable type, burial depth, orientation, and environmental factors — learning to distinguish the cable's magnetic signature from background interference. The result is dramatically improved location accuracy, enabling faster surveys, reduced repeat-measurement costs, and better-informed maintenance decisions.
Models trained on the physics of electromagnetic fields around cables — accounting for cable type, burial depth, and orientation. Goes beyond pure data-driven approaches for better generalisability.
Automated identification and correction of systematic scale-factor errors in magnetometer measurements. Reduces position deviations from 50–150 cm to sub-10 cm in target conditions.
The correction model integrates directly with Seekable's survey platform — outputting corrected position data in real time during survey campaigns, with uncertainty quantification.
| Project name | ZeeCAIbel |
| Lead partner | Seekable B.V. (Amsterdam) — survey hardware & operations |
| R&D partner | Tarucca Technology B.V. — AI & correction algorithm |
| Funding | MIT R&D Samenwerkingsproject 2025 |
| Primary application | Offshore wind export & inter-array cables |
| Secondary applications | Electricity grid cables, telecom cables, general cable monitoring |
If you manage subsea cable assets or conduct subsea cable surveys, ZeeCAIbel offers a direct path to better data quality and lower costs.
We are also interested in conversations with parties who face similar localisation challenges in other infrastructure domains — electricity grid cables, telecom cables, and pipeline monitoring.