Artificial intelligence for digital twins

A pandemic, a military conflict, and even a simple hard-to-detect programming error can quickly lead to huge disruptions. Companies want to take countermeasures with digital twins, artificial intelligence, and new process mining technology.

by Ulrich Eberl

Why are programming bugs called “bugs” after beetles and other crawling insects? Anyone who deals with computers knows this story: Exactly 75 years ago, in September 1947, a moth flew into one of the many Mark II relays of the main computer at Harvard University in the US and caused the computer to crash and crash.

Small Cause, Big Effect – it’s not just the computer world: in Sweden, Japan and Israel, jellyfish have already clogged the seawater filters of power plants and caused them to shut down. The “Ever Given” is also unforgettable: a container ship that turned in the Suez Canal in March 2021 and blocked it for six days. Hundreds of ships had to wait up to two weeks in traffic, or immediately chose a much longer route around the Cape of Good Hope at the southern tip of Africa.

But compared to the situation the world is in now, those were only minor mistakes. As a result of the war in Ukraine, energy supplies to Europe are on the brink. It’s not just about electricity and heat – many companies need natural gas to produce, for example, plastics, pharmaceuticals, fertilizers, steel, glass and aluminum. With significant supply bottlenecks, almost all industries, from automotive to agriculture and the production of silicon wafers, which are essential for the production of computer chips, will have a knock-on effect.

Chaos in the supply chain

That’s not all: China still wants to get a more harmless but more contagious omicron variant of the corona virus under control with heavy blockades in Shanghai, Beijing and other megacities. Consequences: supply chains collapse, container ships and planes can no longer be loaded and unloaded as planned. In industrialized countries, microchips, magnesium, cobalt and rare earth metals are lacking. Prices are soaring and the specter of inflation and recession is rampant all over the world.

If the once well-oiled wheels of trade relations collapse and stop interlocking, an industrial fire threatens, possibly even a global economic crisis. What can be done in such a situation? How can companies react quickly and flexibly to disruptions?

Experts have high hopes for two new digitization solutions and their support by artificial intelligence (AI): “resilient digital twin” and “process mining” – automatic analysis and optimization of processes based on their various data traces.

Models, shadows and twins

Dutch IT specialist Wil van der Aalst, who has been working as Alexander von Humboldt’s professor at RWTH Aachen University since 2018, is considered one of the initiators and pioneers of these technologies. It distinguishes between digital models, digital shadows and digital twins. Models are the simplest form of digital imaging: for example, wind turbine blades are manually recreated on a computer based on their design drawings and their movements are simulated.

Digital shadow goes a step further: it adjusts as the real object changes. Sensors on the wind rotors can report how they behave in heat or storms. The computer then automatically modifies the digital shadow.

Thanks to the digital twin, communication even takes place in both directions, i.e. from the virtual world to the real world. For example, if the computer changes the angle of attack of the rotor blades, it also happens immediately with the turbine in a real wind farm.

Digital twins are especially important for complex processes: if the computer simulation shows how certain processes can be optimized, software updates can be installed on real machines and processes can be reconfigured. AI machine learning methods are very helpful in this: By using a lot of real training data, a computer algorithm learns what changes cause what – and thus also how improvements can be made.

Leave a Comment