Large numbers of visual inspections are routinely undertaken across the aerospace industry, usually involving costly and high-risk manual labour. AI domain specialist, Kiran Krishnamurthy of CFMS, looks at how artificial intelligence can be used to automate these processes and overcome the challenges faced. Manufacturers in the aerospace sector will no doubt be aware of some of the challenges manual inspectors face when they carry out an industrial inspection of an aircraft, which looks in detail at whether a structure, product, component or process meets the specification requirements.
These challenges include human error, recurring labour costs and health and safety concerns, largely associated with the need to speed up the inspection process in line with growing demand for high quality machinery that lasts.
Artificial Intelligence (AI) is beginning to fulfil its potential as the next breakthrough technology in driving productivity for high-value manufacturers. Its development has already delivered smarter and more efficient ways of working, as well as pushing down production costs.
At the forefront of the latest and most exciting advancements in this technology is AI for industrial inspection (AI4II), which is set to transform the way manufacturers carry out the visual examination process.
Firstly, AI4II technology, when sufficiently trained, can review feedback from millions of assets and process thousands of data sets and images in a significantly shorter period of time than a manual inspector could, with a much lower margin for error than a human workforce. This works particularly well for manufacturers that need to carry out fast and routine inspections of assets.
It also saves manufacturers money and reduces some of the health and safety risks associated with manual inspection. By reducing the need for an engineer, AI4II enables manufacturers to redirect the focus of their highly-skilled workforces to value creation tasks rather than troubleshooting defects and means inspectors will no longer need to examine potentially hazardous, hard-to-reach or confined spaces.
Finally, AI4II helps with spreading awareness of possible applications of emerging AI technologies in traditional manufacturing industry. Manufacturers can harvest data through the AI technology to inform future technologies and materials, therefore informing future business decisions. It is hoped this kind of automation will promote a culture of data curation within the industry, in turn leading to further adoption of AI technologies to automate business processes.
To help the aerospace industry better understand AI4II technology, CFMS has produced three demonstrators which combine computer vision and AI technologies to automate the manual inspection process and counteract some of the challenges that high value manufacturers face.
The three demonstrators have each been developed to utilise different technologies and equipment which can identify and prevent system failures and faults: the first uses a drone to capture footage, the second a smartphone camera and the third a fixed camcorder.
As part of the automated inspection process, the inspected component is photographed to produce a series of images that can be processed by the AI4II system. Footage captured by a camera is then used to train the AI to identify objects and defects, which can then be flagged to an engineer to repair or replace the component.
To enable this process to happen, AI4II uses ‘deep learning’ technology, consisting of neural networks inspired by the biological neural networks used in the human brain. Once the AI has learned how to identify defects, this data is then collected by the system to speed up processes of identification and repair for the next inspection.
It is without a doubt that AI will be a vital tool for high-value manufacturers in the future, providing an opportunity to introduce innovation and new technology to the visual industrial inspection process, as well as overcoming some of the challenges the industry currently faces.