Latest Branch News

News from the Institute Branches – May 2019 

| South Wales Branch
Ultrasonic data analysis using AI


Colin Macfarlane reports

On 21 February 2019, the South Wales Branch gathered at The Village Hotel, Cardiff, to welcome Neil Harrap, Principal Project Leader at TWI, who presented a talk titled: ‘Ultrasonic data analysis using artificial intelligence’.

Neil began with a short description of his background since 1976, which included 15 years in the Royal Air Force (RAF) as an Engineering Technician, 12 of which involved non-destructive testing (NDT). This was followed by 14 years as an RAF NDT Technician and Team Manager and then two years as a Ministry of Defence (MOD) NDT Engineering Project Manager. After working at TWI as a Principal NDT Lecturer, Advanced Ultrasonics, for four years, Neil then spent two years at both OEM Industrial NDT and Ashtead Technology before returning to TWI Ltd in 2015.

Neil explained artificial intelligence (AI), noting that it can be used to automatically control complex tasks/systems with the capability of ‘learning’ from experience. However, the capacity for human intervention is also an important element. It was noted that we encounter AI often during our daily lives. ‘Smart motorways’ was one example given, where AI is used to control variable speed limits by reacting to the volume of traffic and incidents to ensure optimum traffic
flow.

The presentation then posed the question: Why do we need AI for data analysis? Using phased array ultrasonic testing (PAUT) as an example, it was identified that the volume of complex data is constantly increasing and the number of competent personnel who can analyse the data is not. There is an ever-increasing need for consistent high-quality data analysis to provide accurate defect quantification. Inspection personnel are under pressure to analyse and provide reports as quickly as possible, resulting, on occasion, in rushed data analysis being undertaken ‘back at the hotel’ or at the office at the end of a taxing work day in the field, increasing the likelihood of human error.
The solution identified by Neil was a powerful, scalable, cloud-based analysis software with the use of 3D AI algorithms, which has future potential to achieve automatic defect recognition (ADR) and sentencing against code, standard or specification. Data is acquired by the operator, which is sent to cloud-based servers. AI takes all of the data from the inspection to create a true 3D image. The data is analysed in terms of quality (lack of data, lack of couplant, etc) and an SMS or email is sent to the operator by the supervisor to confirm whether the data is acceptable or if repeat scans are required. The data is analysed by AI and a full report is generated. The report includes an assessment of data quality, defect detection and measurement and location and characterisation of geometrical indications.

The presentation listed applications for the AI software, including audits, both external and internal, checking the quality of work and as a follow-up tool to check previous inspections or the evolution of indications. It was further termed as an efficiency tool due to the ability for more analyses to take place per hour, combined with higher-quality analysis.

Neil then described one case study of a third-party cross-check, involving the fabrication of a high-pressure, high-temperature system. There were quality issues with the contractor performing critical PAUT and it was suspected that PAUT files were being duplicated. Using conventional OEM 2D software, many hours were spent attempting to identify copied files, but this proved inconclusive. However, when a random selection of 650 data files were uploaded to the cloud, the AI 3D software took only a few minutes to analyse the data and identified that 40% of the files were duplicated.
Neil concluded by highlighting that AI can be used to analyse not only PAUT but any digital data, such as eddy current and radiographic data.

For further information, contact Neil Harrap on: neil.harrap@twi.co.uk

The South Wales Branch would like to express its gratitude to Neil for a very interesting and enjoyable evening and presentation. 




| South Wales Branch
Innovations in digital radiography and computed tomography


Steve O’Brien reports

On 21 March 2019, the South Wales Branch gathered at The Village Hotel, Cardiff, to meet Guy Tolley from North Star Imaging (NSI), who presented a talk titled: ‘Digital radiography (DR) and X-ray computed tomography (CT)’.

Guy began with an overview of NSI, which is part of the Illinois Tool Works (ITW) Group. ITW is a global diversified manufacturer of advanced industrial technology with 50,000 employees worldwide and includes well-known brands such as Buehler and Magnaflux, among others. NSI’s core business is systems sales and manufacturing, equipment software training and maintenance, technical support of X-ray equipment and inspection services consultation.

Guy then gave an explanation of the technology, starting with the basics of DR and how very small focal spot sizes can be used to create geometric magnification with improved image sharpness and to increase spatial resolution by spreading a given indication over a greater number of pixels.

Guy then moved on to the basics of CT, explaining how the 3D image is, in effect, the combination of thousands of 2D images and the smallest distinguishable element is a 3D pixel called a voxel.
The presentation continued with slides illustrating some industrial CT system designs and the various methods of image generation, for example a cone X-ray beam, where a flat-panel detector is used, and a fan X-ray beam, where a line detector would be used. In addition, recent innovations such as helical CT scanning and detector shift/motion were compared with standard CT images.

Guy then showed us a series of examples of applications in which the technology is used, ranging from castings, medical and electronics to plastics and composites, military/defence and explosives. Of particular interest was its use in the relatively new field of inspecting components made using additive manufacturing, particularly given the challenges these present to more conventional NDT methods.

It was noted that one of the more ‘unusual’ uses of CT systems is in the field of reverse engineering, in which the images taken can even be used (with the appropriate software) to generate 3D models.
Guy concluded by giving us an insight into the latest innovation, 4D CT radiography. This is where items are CT scanned while, for example, liquids are flowing through galleries, motors/pumps are running or material failure modes are being observed during fatigue testing.

For further information, contact Guy Tolley on: gtolley@4nsi.com or visit: 4nsi.com

The South Wales Branch would like to express its gratitude to Guy Tolley for a very interesting and informative presentation.