Unmanned aerial systems inspection of civil structures

ASSAI: Unmanned aerial systems for advanced contact inspection of civil structures

Background

Unmanned aerial systems (UAS) offer companies a potential reliable and safe solution to support their operations by accessing areas that are otherwise too difficult, without extensive manpower and support. Although rapid gains have been made in this field of technology, there remains the need to carry out advanced contact non-destructive testing (NDT) in order for UAS systems to truly become the primary method of facility inspection and monitoring.

Currently, the sensor equipment being produced for use with UAS are limited to a range of imaging equipment such as video cameras through to thermal imaging cameras, surveying and mapping technology. As a result, human inspections coupled with ultrasonic sensing systems are still required in parallel with the UAS system, especially for critical infrastructure inspections.

Objective

The ASSAI project aims to develop a functioning ultrasonic inspection unmanned aerial systems (UAS), with fully integrated UT probe capable of performing Under Bridge thickness measurements of steel support beams.

Benefits

Our project will deliver significant productivity increases for our customers, and provide exciting growth to the SME partners in the project. During the first five years (2021-2025) we will generate cumulative total revenue of £28.3M and cumulative profit of £6.2M, from the sale of 345 ASSAI systems. We estimate a very attractive financial return of 1027% (IRR). Internal-Rate-of-Return is more representative of the longer-term investment than ROI, which does takes into account depreciation/inflation

Brunel Innovation Centre's Role

BIC developed a novel Artificial Intelligence assisted Ultrasonic Signal processing method. It combines the advantages of both conventional signal processing and the deep learning assisted method, enhancing this way the accuracy and reliability of the inspection. Based on the acquired data, an algorithm decides which method can offer higher reliability and accuracy in the output data, and processes the signal accordingly. Actually, Brunel University London developed 3 algorithms:

• Corrosion detection: thanks to the combination of signal processing and machine learning, BUL was able to automate the detection but also to go beyond the accepted sensor standard sensitivity
• Quality of data assessment (standard, substandard) using machine learning. This is critical for drone inspection as it allows to know if a measurement should be repeated or no automatically and in real time without downloading and assessing the data
• Surface finish detection

Project Partners

• Air Control Entech Ltd
• TWI Ltd
• JR Dynamics Ltd
• James Fisher Testing Services Ltd
• Brunel University London