India unit drives GE’s plan to build ‘digital twins’ of everything

When the HR department at General Electric’s R&D centre in Bengaluru started looking for a way to curb its already low attrition, it approached Vinay B Jammu and his team.

Jammu, the head of physical-digital analytics at GE, along with his team of 70 engineers, built ‘digital twins’ of complex machineries, such as jet engines that go into airlines, to predict wear and tear and failure before they actually happen and minimise the downtime, saving millions of dollars for customers.

“We built a model taking into account different factors of people like who their manager was, what their pay was, how many promotions and awards they had received. It wasn’t a digital twin model of one individual, but a group of 5-10 people, for which we made predictions on which was a higher risk group where you might have people leaving. Last year, we validated this and our models were pretty good,” says Jammu.

The solution they built is now getting implemented across GE globally, turning into an invaluable tool for the HR departments to curb attrition, boost productivity, and save costs.

Last year alone, GE was able to save $700 million by improving the efficiency of various processes using digital analytics to solve problems. From its factories to processes, right to reducing wastage in its cafeterias, GE is embracing its digital twin technology across the board, while using the same technology for serving customers across its business, including aviation, power, energy, and health care.

Though GE was already working on the technology, the term ‘digital twin’ was introduced by Colin Parris, currently GE’s vice-president for software and analytics research. Parris joined from IBM in 2014.

So far, GE has developed about 700,000 digital twins globally, 200,000 of which were built at the India technology centre.

Parris has set a soft target of developing digital twins of 5-10 per cent of the connected devices that are expected to hit the market by 2020. The scale of the goal can be gauged when one considers that the world is expected to have more than 50 billion such smart devices by that year. Right from coffeemakers to massive thermal power plants, everything is expected to be connected to the internet, and data analytics, machine learning and artificial intelligence will

make them even more efficient.

“The multiplicative factor is now kicking in. From having 70 people building digital twins to the aviation group, power group, renewables group, transportation group and the health care group, all building digital twins,” adds Jammu.

The scale GE is targeting puts it in direct comparison with internet giants Google and Amazon, which are investing billions in machine learning and artificial intelligence.

But given GE’s focus is largely on the industrial front, the economic impact of the prediction is high. Every prediction the company makes using digital twins for a jet engine or a thermal power plant, the decision can cost upwards of $500,000 for its customers.

The research also cuts across different streams of science including material science, physics, and metallurgy, among others. For example, to find how much impact each flight across regions such as Pacific or Atlantic may be causing a jet engine, GE uses Nasa’s database that gives a picture of different levels of particulate matter at different altitudes. Particulate matter of less than 2.5 microns is harmful for aircraft engines.

Jammu says in order to make its digital twins more accurate at predicting, it needs to get more of them out there.

After tasting initial success by proving costs can be saved, his next goal is to build software that writes its own digital twins so that more products can be infused and improved.

“These twins then become smarter as they ingest more data, they learn from other twins that are operating out there and they learn from every prediction they make, whether or not what was predicted actually happened. So, a lot of this work is happening in India, a lot of it is happening for India and is being implemented for our customers around the world,” says Munesh Makhija, CEO, GE India Technology Center CTO, GE South Asia.