The virtual engine

Aircraft engines are high-tech products in a class of their own. Back when passenger air travel was in its infancy, engineers would spend long periods hunched over the drawing board to design these complex machines. The introduction of computers simplified and speeded up the process immensely, and the innovations don’t stop there: “Collaboration among the different disciplines involved in developing an engine is something that works particularly well in the virtual world,” says Dr. Anna Wawrzinek, Digital Transformation Manager in the field of engine development at MTU Aero Engines in Munich. “It makes it easier to coordinate the various product requirements.” MTU’s primary focus here is on developing future engines such as the second generation of the geared turbofan, on which MTU is collaborating with partner Pratt & Whitney, or, in the military sector, the New Generation Fighter Engine.

Digital twin for future engine development

For some time, MTU’s propulsion engineers have been working on the virtual engine, which is made up of a series of digital twins. “A digital twin is much more than a static model that provides a snapshot of a particular stage in development. Rather, it’s a virtual representation that tracks how a given product has evolved over time,” says Dr. Martin Engber, Chief Engineer Virtual Engine at MTU. A digital twin allows developers to simulate an unlimited number of scenarios and draw conclusions about product development, manufacture, operation, and maintenance.

First to emerge is the “as-designed twin,” which describes what the ideal product should be like. This is used to plan the manufacture and assembly of the real product. With the data gathered from this physical product, the engineers can create a virtual “as-built twin.” “That means the as-built twin isn’t an exact copy of the as-designed twin, but differs from it in several respects, since it contains all deviations that occurred during manufacture and assembly,” Engber says.

Having both these digital twins available makes it possible to analyze how they differ. As a result, the engineers can determine whether or not the finished product not only fulfills stringent quality and efficiency requirements, but is also cost-effective to manufacture and maintain. The next step is for the as-built twin to go through acceptance testing and internal validation before being used in flight operations. Any wear and damage that occurs during these operations represents further deviations and is documented in what’s called the “as-used twin.”

Charting the evolution of every engine over its entire lifecycle means that each one also exists as a virtual engine and a corresponding digital thread, which documents all flows of data for that engine. “Each engine has its own traceable life story, which over time steadily moves away from its original as-designed twin,” Engber explains. The sum total of all these deviations, and all the associated data, ultimately allow the engine experts to draw conclusions about how a product will perform in the future. In turn, they can predict when the engine will require maintenance and gauge when it makes sense to replace parts or take the engine out of service.

Processing massive datasets

Predictions like these are based on vast amounts of data—and evaluating it calls for comprehensive models. Engber is aware of the sheer scale of this challenge: “To be able to access all the relevant data at all times, we have to harmonize the different data systems used for analytics, design, production, and flight operations.” The digital thread plays a key role here as the sum of all the data that links the individual phases and disciplines.

Wawrzinek has identified three factors that will be crucial to achieving success here in the future: “These are: a high level of automation in gathering, providing, and processing data; a high degree of collaboration and interdisciplinary processes within the company; and, last but not least, artificial intelligence, which can help make the forecasts and analyses more accurate.” At present, the artificial intelligence (AI) applications used in engine development tend to be for approaches—structural mechanics, for instance—designed to assess a component’s natural vibrations. But the hope is that AI will gradually start to deliver a wider range of insights and make it possible to analyze and optimize complex engines.

Customizing engines through virtualization

One of MTU’s first lighthouse projects concerns the digital twin for compressor blades. “Aerodynamics and structural mechanics have an antagonistic relationship, and there’s always a need to reconcile the two,” Engber says. “While the interests of aerodynamics are served by having blades with edges that are particularly thin and sharp, structural mechanics’ pursuit of robustness favors thicker, more rounded components,” he adds to explain the tightrope that developers must walk.

That, Engber continues, is why aerodynamics and structural mechanics have been designated as key processes in designing blades for compressors. In the medium term, MTU wants this approach to accelerate and enhance product design as well as to reduce costs. “We’ll soon be in a position to automate optimization of any target variable and develop the right product for each use case,” he says. The field of tension for optimization always consists of the technical requirements, such as efficiency and weight, as well as the manufacturing and maintenance costs. And, of course, the product must be sufficiently robust.

Following the digital thread

MTU’s main focus in this area is currently still on the as-designed twin in other words, on the development processes in engineering and on data transparency between technical departments. In the medium term, the focus will shift more toward production and operations, which will involve the as-built and as-used twins. “We want to draw a map of the entire digital thread—from development to production and operations to decommissioning,” Engber explains. To this end, at the beginning of 2024, MTU set up a dedicated team of experts to coordinate the virtual engine agenda and drive it forward in collaboration with the various technical departments.

MTU is receiving active support from research institutions. One of these is the Institute of Test and Simulation for Gas Turbines run by the German Aerospace Center (DLR), where innovations are first developed and tested on the laboratory scale before they are rolled out to the industry.

End-to-end digitalization applications like these for development, manufacturing, and maintenance are a huge opportunity for companies, but they also pose a challenge to the MTU engine experts from the various technical disciplines and areas of the company—one that will require them to rethink how they do their job. “We must draw employees’ attention to the potential that digitalization methods hold, and we have to reinforce their digital mindset and data-driven thought processes. In short: we need to take them with us on this digital journey,” Wawrzinek says. “At the end of the day, coming up with a virtual version of a product that’s as complex as an aircraft engine is definitely a marathon task involving countless tiny steps—but with each step, we inch closer to making improvements.” There’s still a long way to go before the vision of capturing and mapping an entire engine and its functions in the virtual world can be realized. “But the journey has begun,” Engber says.