Making the virtual world work for reality
In the early days of space travel, NASA came up against a perplexing question. How do you work out how to maintain, monitor and modify systems that, to say the least, can’t be easily accessed? The space agency began experimenting with digital mirroring technology that created virtual replicas of physical systems and equipment. These were the precursor to digital twins, but it wasn’t until 2002 that Michael Grieves first used the term in an academic paper to describe a virtual recreation of a process, product or service. Now, 15 years later, the expansion of the Internet of Things has led to what could be the cusp of a digital twin explosion.
By combining artificial intelligence, machine learning, software analytics and data, digital twins provide live, self updating simulation models. They gather data via sensors, sharing it on a cloud based system. This is analysed against a variety of other sources including human expert input, information from similar machines, and the wider environment. The aim is to better understand how the process, product or service could be improved. This can involve troubleshooting, real time monitoring and prognostics (predicting future outcomes).
The beauty of digital twins is that they can be explored in a way that doesn’t affect the original. Problems can be predicted ahead of time, reducing downtime and repair costs while enhancing operational efficiency and customer experience. By 2020, IDC forecasts that 30 per cent of Global 2000 companies will use the technology, improving product innovation and organisational efficiency by up to 25 per cent. Digital twins have applications within a range of different industries, particularly those with complex or inaccessible elements. Wind turbines, aircraft engines, large structures, and any complicated piece of machinery are ideal candidates for virtual replicas. As machine learning and cloud connectivity expands, so will the potential of digital twin technology.