Pulsed Eddy Current Technology Deployment Using Unmanned Aerial Vehicles
Date
Jan 2022 - Jan 2024
​
Location
University of Strathclyde, Glasgow
​
Project type
PhD and Industrial Project
​
Role:
-
Lead Designer of the Sensor System
-
Hardware Assembler & System Integrator
-
3D Model & CAD Designer
-
Full Robotic System Builder
-
Aerial Dynamics Modifier for Hybrid UAV
-
Videographer & Editor for Project Documentation
​
Publication
To ensure the safety of critical infrastructure like oil and gas pipelines, accurate inspection for wall thinning is essential. While drones with cameras offer safe, remote access, they can't detect defects beneath coatings. Our solution: a compact Pulsed Eddy Current (PEC) sensor mounted on a crawler-hybrid UAV. This system enables 360° pipeline inspections, accurately measuring wall thickness—even under coatings—with errors under 4.8%. Designed for hard-to-reach environments, it offers a powerful, non-intrusive tool for reliable asset monitoring.
The UAV platform, depicted in the above figure represents an innovative, dedicated platform designed for deploying NDT sensors, in consistent proximity to the surfaces of cylindrical structures like pipes. This UAV model is equipped with six five-inch, fixed-pitch, reversible propellers. These propellers provide balanced thrust capabilities in both directions, enabling the UAV to invert its thrust output completely. The arrangement includes four propellers inclined at 25° from the vertical and two tilted 15° above the horizontal plane, allowing for a combined thrust capability of up to ±30.8 N vertically and ±28.8 N horizontally relative to the drone's frame. This thrust is suitable for the UAV, which weighs 1.91 kg with a 272 g PEC sensor system payload, to maintain contact forces exceeding 8 N in every direction around a cylindrical target, moving across its surface after landing without the need for magnetic or vacuum-based attachment methods. The UAV is equipped with two continuous rotation servos for movement, making it a hybrid between a drone and a robotic crawler. These servos, along with rigid legs that stabilize the UAV and can be adjusted to wrap around objects larger than 160 mm in diameter, ensure precise contact and navigation around the target. Running on a customized PX4 flight control software, this UAV platform uses a rigid-body interaction model for stable flight around cylindrical objects. An Inertial Measurement Unit (IMU) provides all necessary data, allowing for efficient use in industrial environments without external positioning systems
A customized modular mounting structure was engineered to secure all electronic components during the operation of the UAV around the pipe. A 3D-printed fixture was fabricated to facilitate the integration of the PEC with the UAV. The instrumentation circuit boards, and embedded computer were rigidly mounted within a 3D printed enclosure to prevent motion during flight and protect the sensitive components from the operating environment. The layout of this assembly is illustrated in the above figure.
Furthermore, the sensor is encased in a 3D-printed shell, which is further shielded by an aluminium cover, reducing noise in high frequency electromagnetic fields. This configuration isolates the probe, mitigating electromagnetic interference from the motors. Importantly, this shell structure incurs only a minimal weight increase of approximately 12 grams, aligning with our goal to improve sensing capabilities without compromising the aerial platform's efficiency.
The PEC system is designed for mobility and comprises three primary components: 1) the PEC probe, 2) the onboard PEC electronic system, and 3) the ground station. These elements work in conjunction to generate eddy currents in a test specimen for thickness measurement purposes. An excitation board within the system is responsible for sending a rectangular voltage pulse to the excitation coil. Additionally, a Raspberry Pi equipped with an Analog-to-Digital Converter (ADC) is utilized to issue control signals that activate the excitation coil and initiating the data acquisition system with data logger and signal display. The system power is supplied from the main UAV LiPo battery via a voltage regulator, dropping the input voltage from a nominal 22.2 V to 12 V and maintaining this as the battery discharges. It incorporates an excitation circuit-board equipped with a Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET) module to safeguard against potential high-current damage to the electronics. Moreover, an amplification board is integrated to boost the voltage signal captured by the receiver coils, enhancing signal clarity for processing.
​
​For the signal excitation part, a pulse with a 1 s period and 5% duty ratio is chosen. The signal was sampled at a rate of 25 per data point. The excitation circuit is composed of the MOSFET, the MOSFET driver, the control signal, and a 12 V DC Power supply. The DAQ device (in Section III.B) provides the control signal to the circuit. The 12V DC power is powered by the 6S battery going through the 12V DC/DC regulator.
​
When considering the receiver circuit for capturing the PEC waveforms, initially, the signal from the detector coil, which falls within the millivolt range, undergoes amplification through two differential instrumentation amplifiers with a total gain of 200. Subsequently, it is further processed by another amplifier with a gain of 1 to combine the differential signal for sampling by the DAQ system. The ADC samples at 40 kHz, interfacing with the Raspberry Pi Zero W 2. Finally, data is wirelessly transmitted to the ground station for recording and analysis.
(a) 2-D viewing of the tested pipe sample, (b) 3-D viewing of the tested pipe sample
The final phase of the experiment, aimed at mapping the thickness of a pipe, involved piloting a vehicle equipped with a PEC sensor system around the pipe's entire circumference while maintaining constant contact. The experiment utilized a nominal 1400.0 mm long, schedule 80, pipe composed of five sections, each with a 12.75-inch (324.0 mm) outer diameter, made of low carbon steel. The ends of the assembly had a thickness of 37.0 mm over 400.0 mm lengths. In between these ends, three sections varied in thickness—20.0 mm, 10.0 mm, and 6.0 mm, respectively, over a total length of 200.0 mm, as depicted in the above picture.
Sequential image series showing the vehicle around the pipe, covering each 45° station. Temporal progression runs from left to right, top to bottom
The radar plot, which depicts the pipe's thickness in relation to the reference pipe thickness, and the error analysis, are presented in the above figure and table, respectively. Upon reviewing the results, a relatively significant signal error was observed at approximately 180° (bottom of pipe) for the 20.0 mm pipe thickness section, with a maximum mean error of 0.412 mm. This error is likely caused by the increased downward forces of gravity and could be attributed to the increased motor speed required to support the vehicle in this position, thereby generating additional Electromagnetic Interference (EMI). Additionally, the performance of the PEC system can be influenced by varying environmental conditions, such as temperature and humidity. These factors can affect both the electronic components of the PEC system and the properties of the materials being inspected. Extreme temperatures can impact the performance of the PEC sensor and the UAV. High temperatures may cause overheating of electronic components, while low temperatures could affect battery performance and sensor accuracy.
​
Despite these issues, the prediction of thickness under three different conditions demonstrates good consistency, with the maximum standard deviation error from dynamic tests being 1.070 mm, adequately distinguishing thickness variations around the pipe. The relative errors remain below 5% across all measured conditions, indicating that the margin of error is sufficiently narrow to reliably discern variations in thickness.