Decoding ‘Digital Industrial’: Getting to Know Your Digital Twin
By Bhanu Shekhar, Chief Digital Officer & Head of Digital Solutions for GE Power in the Middle East & Africa
Imagine if you knew exactly when your car needed to be serviced, rather than simply taking it to the garage every 5,000km, as is generally recommended? You’d save money, not only because you were servicing the car only when it needs it, but by getting it serviced at the right time, the car would perform more efficiently and wear down more slowly.
Then, imagine if your car could give you a heads up before it breaks down and tell you how to fix whatever is wrong? You’d save money on that as well, by avoiding towing costs and your own lost time spent waiting by the side of the road.
What’s so exiting is that this isn’t science fiction.
Building such a system – which GE calls a “Digital Twin” – is already happening, for some of the biggest industrial assets in the world, like multimillion or multibillion dollar power plants or aluminum smelters. For these types of machines or plants, a Digital Twin delivers significant cost, performance, productivity, reliability and availability benefits that represent huge opportunities for operators by accelerating their digital transformation.
GE’s Digital Twin is powered by Predix, GE’s platform for the Industrial Internet of Things. It creates value because equipment or entire plants never operate exactly according to specifications or like every other piece of similar equipment. There are many variables that impact the performance of an asset.
Take a power plant, for example. Ambient temperature, humidity, variations in the fuel, wear on the hundreds or thousands of parts, interval since the last maintenance, and deviations between plant design and actual construction, are just some of the factors affecting the performance of a plant.
Especially in this region, where conditions change rapidly and are particularly hot and harsh, historically, dozens of operating settings would be developed manually to account for dozens of different operating conditions. But there is a limit to how many can be developed. Digital Twin not only dramatically increases the fine tuning of these settings, but can dynamically change parameters and settings based on current data.
Or take an aluminum smelter. Power is one of the most costly ingredients of aluminum. Facilities often have numerous potlines, or smelting cells, running, with a single plant managing more than 100 pots, each heating raw materials. Historically, all the pots would be treated the same, based on a mean regarding optimal temperature, maintenance, etc.
However, each pot is slightly different from the next, based on factors such as their own production conditions and current state of wear. By building a Digital Twin of each pot, plant operators can treat each pot uniquely, boosting efficiencies, lowering fuel and other operating costs, setting an ideal maintenance schedule, and even predicting failures or maintenance requirements before they cause the pot to malfunction.
The GE Power Digital team in Dubai led the company’s implementation of the world’s first Digital Twin project in the aluminum smelter industry at Aluminium of Greece. Learn more about this groundbreaking project here.
Using GE’s platform for the Industrial Internet of Things, Predix, Digital Twin builds a bridge between the physical and digital worlds, providing operators with a radically new and deeper understanding of each unique asset over time. Predix is the first platform specifically designed with the security, big data bandwidth, and reliability required for industrial systems.
Each twin – whether of a specific piece of equipment to an entire plant – is built by combining large volumes of sensor data with analytics, models, and material science to provide a detailed, constantly evolving picture of machines and operations.
Not just plug-and-play software
However, building such a model is not as simple as installing an app or suite of software solutions. In part, it requires an extensive benchmarking and model-building process covering the asset or system, such as a plant, that can take from a few months to more than half a year.
The process begins with a team of engineers with knowledge of each facility and asset mapping out every piece of equipment or element of the system. This includes inputting data on the production of the specific asset(s); its history of operations to date, if not newly installed; all parameters of the plant; and related third-party inputs, such as fuel consumption. These are just some of the factors incorporated into the Digital Twin.
Then, performance engineers fine-tune the model using information such as historical performance data and the expertise of reliability and maintenance engineers who have worked on these assets. Next, all the equipment and sensors are connected to cloud-based infrastructure that collects, monitors and, analyzes the data. The data and output from the Digital Twin software is then tested and validated. Once approved, the Digital Twin ready to go to work.
What’s so exciting about GE’s Digital Twin is that it allows companies to test and simulate different performance scenarios, operation optimizations and new business models virtually, without putting operations at risk. They help companies uncover deep patterns of behavior and derive the most out of each asset by integrating analytics from Digital Twins across an entire class of assets – making them smarter and smarter.
Not only can it advise when equipment is likely to break down, but it can also provide recommendations on how to fix the problem, by running simulations based on past history, context and environment, building feedback loops for continuous improvement.
For both asset and plant owners and operators, Digital Twin offers a truly revolutionary way to approach maintenance, repair and optimization. I look forward to helping more GE customers in the region adopt Digital Twin in their facilities. I invite you to reach out to me for more information.