Numerical Validation of Injection Models for Hollow-Cone Nozzles using Euler-Lagrange Simulations

In the quest to make air travel cleaner and more efficient, engineers are looking beyond traditional jet engines to concepts that might seem almost counterintuitive: adding water to the mix. One such idea is the "Water Enhanced Turbofan," a system that recycles water from the engine's exhaust, turns it into steam, and injects it back into the combustion chamber. This process cools the engine, allowing it to run hotter and more efficiently, which ultimately reduces fuel consumption and harmful emissions. However, for this system to work, the water must be introduced into the engine in a very specific way. It cannot simply be a stream of liquid; it must be broken down into a fine mist of tiny droplets that spread out evenly. If the water is not atomized correctly, the engine could suffer, or the efficiency gains could vanish. To design these systems, scientists rely on computer simulations to predict how water sprays behave, but these digital models are only as good as the rules they use to describe how the water enters the air.

Researchers at the University of Stuttgart set out to test how well different computer rules could predict the behavior of a hollow-cone nozzle, a common device used to create these precise water sprays. In the real world, when water shoots out of such a nozzle, it forms a cone shape with a hollow center, much like the spray from a garden hose nozzle set to a wide fan, but with a distinct empty space in the middle. The team used a powerful simulation software to recreate this spray, pitting two different digital strategies against each other to see which one could best match real-world measurements. They compared a simplified approach, which treats the entire spray as a single, uniform cone of droplets, against a more complex method that places thousands of tiny, individual injection points to mimic the exact conditions measured in a laboratory.

The study began by building a digital twin of the experimental setup used by their colleagues. In the lab, a nozzle sprayed water into a quiet room, and high-speed sensors measured the size and speed of the droplets at two specific distances: 100 millimeters and 300 millimeters away from the nozzle tip. The researchers then ran their computer simulations, first using a "solid cone" model. This method is like drawing a single, perfect cone in the air and filling it with droplets that follow a general rulebook for size and speed. It is computationally cheap and fast, making it attractive for large-scale engineering projects. The results showed that this simple model could capture the general shape of the spray and the average size of the droplets reasonably well. However, it struggled to get the details right. The simulation tended to overestimate the speed of the droplets in the center of the spray and underestimated the size of the droplets there, creating a mismatch with the physical measurements. The researchers found that this error stemmed partly from how the model handled the air around the spray; the computer air was moving too fast in the center, pushing the droplets along more than they should have been.

To fix this, the team refined the conditions of the air in their simulation, essentially telling the computer to slow down the airflow in the center of the spray. This adjustment improved the results, bringing the simulated droplet speeds closer to reality. They also discovered that the angle at which the spray opens in the computer needed to be wider than the actual angle measured in the lab. In the physical world, friction against the air slows the water down and spreads it out, but the simplified computer model didn't naturally account for this friction, so the researchers had to manually widen the digital cone to compensate. Despite these tweaks, the solid cone model remained a broad approximation. It worked well enough for general engineering estimates, but it could not reproduce the intricate details of how droplet size and speed were linked in the real spray.

Seeking higher precision, the team then tested a "point array" injection model. Instead of one big cone, this method used a grid of hundreds of tiny injection points, each programmed with specific data taken directly from the laboratory measurements. Crucially, this approach allowed the researchers to link the size of a droplet to its speed at the exact moment it entered the simulation. In the real world, larger droplets often move differently than smaller ones, and this complex relationship was lost in the simpler model. By using the point array method, the simulation could reproduce the spray's behavior with striking accuracy. The digital droplets spread out and slowed down in a pattern that matched the physical measurements almost perfectly, capturing the subtle variations in droplet size across the spray.

However, this high level of accuracy came with a significant cost. The point array model required vastly more computing power and time to run, and it depended entirely on having detailed experimental data to set up the injection points. If an engineer did not have those specific measurements for a new nozzle, they could not use this method. The researchers concluded that the choice of model depends on the goal. For broad studies where the exact details of a specific nozzle are less important, the simpler solid cone model is a practical and efficient tool, provided the user adjusts the spray angle and air conditions carefully. But for detailed investigations where understanding the precise behavior of every droplet is critical, the point array model is superior, offering a faithful reproduction of reality at the price of computational effort.

Ultimately, the study provided a clear roadmap for how to simulate water sprays in future engine designs. It demonstrated that while simple models can get you close, they miss the nuanced connection between droplet size and speed that defines a real spray. The researchers showed that by carefully tuning the boundary conditions—how the air and water enter the simulation—engineers can trust their digital predictions to guide the development of the next generation of cleaner, more efficient aircraft engines. The work confirms that with the right level of detail, computers can indeed replicate the complex dance of water and air, offering a reliable way to test new technologies before a single physical engine is built.