Even the most skilled inspector might have an off day. Here it can be useful to outsource mundane or mechanical tasks to ‘intelligent’ machines.
Poorly defined quality-control procedures were blamed for one of the most extensive automotive parts recalls in history, involving the airbag manufacturer Takata. The firm’s inflators, which contained the chemical ammonium nitrate, were found to be unsafe – leading 19 US carmakers to recall 69m of the products. Similar recalls were issued in Japan, China and Oceania. Such procedures are known for being particularly costly in the automotive sector.
With such high stakes, it’s little wonder that major carmakers are hoping to leverage artificial intelligence (AI) to improve their quality-control processes. They’re pairing high-resolution smart cameras with image-recognition technology to cut the costs and improve the accuracy of inspections. In 2018, Audi began using small cameras in the press machines at its Ingolstadt production facility in Germany to assess the quality of components. Software based on an artificial neural network was then used to detect and mark the location of fine cracks in sheet metal.
AI checks components
Similarly, BMW has piloted an AI application to evaluate hundreds of components during its own vehicle production process. In a matter of milliseconds, the technology can identify deviations from the standard and check whether parts are mounted correctly. To ‘train’ its neural network, the company assembles a database of 100 or so images of a given component, with and without flaws, and runs them through a high-performance server. Following a test run and possible tweaks, BMW says its neural network reaches 100% reliability.
The Capgemini Research Institute, the consultancy’s in-house think tank, identified three foremost use cases for AI in manufacturing in a 2019 report. Product quality control was among them, in addition to demand planning and intelligent maintenance. According to Capgemini’s researchers, 29% of AI implementations in manufacturing were for maintaining critical machinery – while 27% were for quality control.
Research by Deloitte found that unplanned downtime costs industrial manufacturers $50bn a year. While many firms have perfected the art of preventive maintenance, in which maintenance is performed before an equipment failure occurs, there remains room for improvement. With the help of machine learning, it’s possible to carry out predictive maintenance, in which sensor input and algorithms gauge the health of equipment – encouraging the just-in-time replacement of components.
Although there have been successful pilot schemes to introduce AI to factory floors in various roles, there is still a long way to go before these applications are ubiquitous. One 2018 PwC survey of manufacturing executives in 26 countries found that only 9% had implemented AI in their processes to improve decision-making. However, in a post-Covid world, the idea of reducing costs through better planning is likely to have universal appeal. The picture could soon change.
For decades, large, centralised production hubs have been the norm in industrial manufacturing. Now that companies have been reminded of how fragile global supply chains can be, they might seek to open up smaller production facilities in a greater number of locations.
Agility, and the ability to respond to shocks in supply and demand, will be paramount. While AI can’t literally predict the future, it can provide helpful maintenance forecasts and spot errors in production before they become a problem.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.