In our first story, we explored how AUTOASSESS partner University of Zurich uses drones for real-time mapping, obstacle avoidance, and wind compensation to navigate complex ship environments. In this second instalment, we look at another important challenge: how drones can physically contact steel structures to carry out Non-Destructive Testing (NDT) in dark, confined, and highly corrosive areas of a vessel.
Visual inspections can identify problematic areas of a vessel affected by corrosion, damaged coatings, or suspected cracks. However, they cannot always determine how much steel has been lost due to corrosion or whether a structure remains within its required thickness limits.
To assess the condition of a vessel’s structure, inspectors need reliable measurements of the remaining steel thickness. This often means working inside difficult-to-access spaces, such as ballast water tanks and cargo holds, where conditions can be hazardous.
In this instalment, we highlight the work led by the University of Twente, in collaboration with NDT specialist Sensima Inspection and drone platform partners Flyability and ScoutDI.
The University of Twente leads Work Package 2 (UAS Hardware Adaptation) and Work Package 8 (Contact-based Thickness Inspection), adapting a fully actuated UAS (AeroX) and developing a portable contact-based NDT module to enable controlled physical contact with steel surfaces for ultrasonic thickness measurements in confined ship spaces.
The Operational Challenge: Measuring Steel Thickness Without Damaging Protective Coatings
Vessels such as bulk carriers, container ships, and oil tankers require regular inspections to assess the condition of their steel structures. When a visual inspection identifies significant corrosion or a suspected defect, further testing may be needed to determine whether the steel has lost thickness and whether repairs are required.
Ultrasonic Testing (UT) is a widely used method for measuring steel thickness. It uses sound waves to measure the distance through the steel and typically requires a suitable contact surface for the probe. Depending on the equipment and coating condition, paint, heavy rust, or other surface layers may need to be removed before reliable measurements can be obtained.
Removing these layers takes time and can damage the vessel’s protective coating. The affected area may then need to be repaired and recoated to prevent further corrosion. For drone-based inspections to become a practical alternative, two main challenges must be addressed:
- Measuring steel thickness through coatings: The drone needs to carry an inspection sensor that can assess the remaining steel thickness through non-metallic coatings, paint, and rust scale, without requiring the surface to be ground back to bare metal.
- Maintaining stable contact with the steel: The drone must be able to press its sensor against vertical walls and inclined or horizontal overhead surfaces with a controlled, steady force. It must do this without slipping, bouncing off the surface, or losing flight stability.
Inside the Technology: The University of Twente’s Solutions
1. Pulsed Eddy Current (PEC): Measuring Steel Thickness Through Coatings
To speed up inspections and reduce the need for surface preparation, the University of Twente worked with Sensima Inspection to adapt Pulsed Eddy Current (PEC) sensors for use on drones. PEC is an inspection method that uses electromagnetic fields to assess the remaining thickness of ferromagnetic steel. It can measure through non-metallic coatings, including paint, and can tolerate some distance between the sensor and the steel surface. This distance is known as lift-off.
Unlike conventional ultrasonic measurements, PEC screening can often be carried out without removing the vessel’s protective paint coating. This helps preserve the corrosion protection and reduces the time needed to prepare inspection points.
What did the laboratory tests show?
Sensima tested PEC technology on steel plates exposed to salt-spray corrosion. Results showed its potential to measure steel loss through corroded surfaces and help distinguish metal defects from cracks in protective paint. This could reduce surface preparation and unnecessary follow-up inspections.
2. Adapting Different Drone Platforms for Inspection
The University of Twente coordinated hardware design and software integration to deliver multiple drone configurations for configurations for contact-based inspection:
- ScoutDI – Scout-137 (TIUAS):Equipped with the UT-based NDT payload, the tethered inspector UAS is suitable for inspection of large cargo holds, as it receives power and data through the tether.
- Flyability – Elios 3 (IUAS): Equipped with the UT-based NDT payload, the inspector UAS is suitable for inspection of confined ballast water tanks, thanks to the protective cage.
- University of Twente Fully Actuated Platform (AeroX / miniThex): Customised with dedicated multi-rotor thrust vectors to apply contact forces against inclined surfaces (walls, bulkheads, and ceilings).
In addition, the collaboration between Sensima and the drone producers has led to the currently under-development:
- ScoutDI – Scout-137 (TIUAS): Integrated with USB-based PEC acquisition electronics.
- Flyability – Elios 3 (IUAS): Fitted with an Ethernet-based integration architecture using a Single Board Processing Computer (SBPC) to connect the PEC payload to the drone’s onboard systems.
These will bring the advancements in PEC-based NDT inspection to the inspector drones.
Video 1: The fully actuated platform can control its translation and orientation independently, which allows the platform to slide on a contacting surface without changing its orientation, ultimately improving the stability of the contact. At the same time, the platform can be controlled as an underactuated platform, which is more energy-efficient. The University of Twente developed the controller to perform such a switch smoothly during the flight, which is shown in the demonstration.
3. Controlled Contact: Stable Measurements on Steel Surfaces
Drones can become unstable when flying close to walls because of changes in propeller airflow. Besides causing an excessive expense of energy, applying too much force can also affect flight stability and measurement quality. To address this, the University of Twente developed a physical interaction control system, based on indirect force control, that allows the drone to be compliant to the contacting surface, thus maintaining steady pressure against the steel surface.
The automated inspection process begins with the drone navigating to the selected inspection points. The system then prepares for contact by removing any biases in the force estimation due to unmodeled dynamics and eventually gently but stably presses the sensor against the steel, following a filtered version of a trajectory reaching a target point at a predefined depth inside the inspected surface. Once the measurement quality is confirmed, the drone detaches and moves to the next inspection point.
This approach allows drones collect reliable thickness measurements while reducing the risk of slipping, bouncing, or losing stability.
Video 2: Field validation showing a caged inspection drone making stable contact with a bulkhead inside a ballast water tank.
Video 3: A 60-second test showing the drone contacting a steel plate, maintaining stable pressure, recording a 12 mm thickness measurement, and safely detaching.
What This Means for the Maritime Industry
The work led by the University of Twente could help make internal ship inspections faster, safer, and less disruptive. The main potential benefits for shipowners, marine surveyors, and classification societies include:
- Less surface preparation and fewer coating repairs: PEC screening can measure steel thickness through suitable protective coatings, reducing the need to grind away paint at every inspection point. This can save preparation time and help preserve the vessel’s corrosion protection.
- More inspection points in less time: Drones can reach multiple inspection locations during a single deployment, reducing the time spent moving equipment and setting up access.
- Reduced inspection time and potential off-hire costs: Faster inspections could reduce the time a vessel spends waiting for structural surveys or undergoing inspection during a maintenance period.
- Improved safety for inspection teams: Using drones to carry out initial screening and selected measurements can reduce the need for surveyors to enter these spaces or work at height on scaffolding. This helps limit exposure to hazardous environments while allowing inspection teams to focus their time on interpreting results and planning any necessary follow-up work.
What’s Next in the Series?
With the ability to navigate confined ship spaces and make controlled contact with steel structures, the next challenge is turning inspection data into a clearer picture of a vessel’s structural condition. In upcoming instalments of our Operational Deployment Stories series, we will explore:
- Semantic 3D Mapping & Digital Twins (TUM & Cognite): Merging point clouds, defect annotations, and NDT measurements into a browser-based Digital Twin.
- AI-Powered Defect Detection (DTU): Domain-adapted deep learning models capable of identifying >95% of corrosion, cracks, and coating breakdowns across vessels of varying ages.
- Full-Scale Industrial Trials: Field reports as the complete AUTOASSESS solution is tested on bulk carriers, container ships, and oil tankers.
Stay tuned as AUTOASSESS continues to advance autonomous vessel inspection!
