There have already been steam turbines which have lasted for over 82 years before they were decommissioned and with rapid advancements, machine immortality seems not too far away.
In any industry, there has always been a notion of models for a given asset, whether a jet engine or a compressor. What if there was a living, learning model which could be constantly fed with data and updated, so that it allowed engineers to predict any problems and prevent it, throughout the entire lifecycle of the asset? The vice president of GE Software Research, Colin Parris, calls this model, developed and maintained digitally, a digital twin.
“We started from a value perspective to arrive at a digital twin. There are long-cycle assets like jet engines which last 40 years and cost hundreds of millions of dollars. When I sell an asset, I don’t make much money,” says Parris, explaining what necessitated a digital twin. “The bulk of the money and the customer value is usually made when these assets are deployed. The customer reaps the value over its lifetime. If I can find a way to increase the value that a customer gets over this period and reduce its cost to me, it becomes a win-win situation for both of us. The customer gets more value and I get more profit by servicing the asset over its lifetime.”
There are three ways in which a customer gets more value—early warning on a failure of an asset, receiving accurate predictions about an asset’s health, optimising an asset for better utilisation. According to Parris, customers start to trust digital twins only when the values translate onto their profit and loss statements. For instance, a digital twin can bring about process transformations in a business. This could result in a customer reducing unplanned maintenance or damages to an asset in a given year as compared to the previous years. At scale, the customer saves money and sees a hike in revenue which may not have occurred in the absence of a digital twin and subsequent lack of transformation in his business processes.
This journey leads to the understanding that the return on investment for any business using a digital twin over the lifecycle of its assets would be exponential. And for a manufacturer and service provider like GE, the insights from digital twins could mean better products and services to offer its customers. Parris says that until 2017, GE has had over 1.2 million digital twins and has saved over $518 million worth of value for its customers.
One example of GE’s digital twins in India is in its collaboration with Reliance Industries to work on their pipelines. This is a 4,000-km pipeline processing 1.3 million barrels of crude each day. On this, GE has installed 30 sensors to monitor a 2.5-km stretch at the refinery. “This system lets us predict leakages of a litre with 80-90% accuracy and prevent them. If the leakage is going to increase, then the reliability of predicting it also increases close to 100%. The value here is in terms of saving inspection cost over thousands of kilometres and avoiding massive downtimes,” says Vinay Jammu, technology leader, Physical-Digital Analytics and Software, GE Global Research. What’s interesting here is that the implementation of digital twins comes down to the marriage of physical and digital—the domain knowledge of engineers who deal with the physical infrastructure and the data science involved at the backend.
Parris says that there are three challenges in the implementation of digital twins. “Everybody believes they have sufficient usable data but they don’t. It costs money and needs a big clean-up which delays things,” he says. Any actionable insight from digital twins requires integration between business processes and financial system of a client so that he can see the results on the P&L statement. Parris says that customers who don’t realise this are the ones who don’t act quickly to integrate the two aspects of their business. Lastly, the lack of learning culture and entrepreneurial bent is a significant problem which leads to people being afraid of losing jobs to digital transformation rather than getting excited about the new doors it can throw open.
If digital twins is a journey in itself in the era of Industry 4.0, overcoming these challenges along the way will lead to immortal machines. The digital twins can constantly adapt. The brain of a machine is constantly changing and learning from new experiences. “When that change happens to the brain, what if I remove a damaged part and replace it with another made from new materials using new manufacturing technologies which would eventually let the machine live longer? What if I can change every part like that? The machine will adapt to any new environment and continue to live,” explains Parris.
There have already been steam turbines which have lasted for over 82 years before they were decommissioned and with rapid advancements, machine immortality seems not too far away.
---
Autor(en)/Author(s): Srinath Srinivasan
Quelle/Source: The Financial Express, 04.04.2019
Bitte besuchen Sie/Please visit:
