In Back to the Future 2, Marty McFly was able to determine the outcome of future sporting events because he had been to the future and back again. But when it comes to real world, predicting the future is a bit more complex. The most current AI technology is pushing into a whole new field of application in the manufacturing by offering to predict the future. AI has been in the manufacturing world in different ways for some time, but the newest application promises to develop an entirely new field of service.
The biggest downfall of all manufacturing processes is machine failure. When machines work well, business runs. But when machines break down, major problems can happen. Not only does work stop, but many of the components of the machine are interrelated and so one breakdown can cause component failure throughout the machinery, resulting in huge replacement costs. The key is to somehow predict the breakdown of a single component so that it can be replaced before failure happens, and therefore maintain production as well as protect the substantial costs associated with replacement. The difficulty, of course, is predicting the future in a way that doesn’t require huge amounts of man hours and production stoppage.
This is where AI steps in. Andreas Schütze at Saarland University has designed an elegant system that effectively checks the status of a multitude of components within each machine. It monitors heat, vibration, and a number of other variables in order to manage the existing components and inform the operators about how each component is interacting with its subsidiary components and if it is nearing a point of failure. This allows the operator to replace the single component in a timely fashion, stopping the breakdown, and keeping production high.
The AI component that is making this a game changer is that the sensors themselves are now able to evaluate the huge amount of data themselves and are able to detect changes in the components without the need for the external data evaluator. ‘We feed the information to the sensors, transforming them into smart devices that are able to detect these signal differences on their own,’ explains Nikolai Helwig from Schütze’s team. As smart devices the are now able to manage the individual components, test for failure, and manage and alert a single operator over a large scope of machines. Savings in salaries and risk of human error should be immediate.
The widespread business applications are clear. Customized smart sensors can soon be manufactured directly into machinery components, allowing them to be seamlessly inserted new component systems in existing manufacturing, allow for gradual shift to the system. Unquestionably, this technology will continue to be utilized across the industrial sector. It will likely also begin finding application in other business fields as well, where system failures can be averted by data monitoring. While it took a DeLorian and a flux capacitor for Marty to see the future, AI is making this a tangible reality for businesses today.