Machine learning-based regression and multi-objective optimization of thermo-hydraulic performance in a tubular heat exchanger with co-and counter-arranged V-winglet twisted tape inserts
This study develops a data-driven framework integrating machine learning models, explainable AI, and multi-objective optimization to predict and optimize the thermo-hydraulic performance of tubular heat exchangers equipped with co- and counter-swirl V-winglet double perforated twisted tape inserts.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the modern world, the demand for energy is growing faster than ever, pushing engineers to find smarter ways to move heat from one place to another. At the heart of this challenge are heat exchangers, the unsung workhorses of industry that transfer thermal energy between fluids without letting them mix. To make these devices more efficient, scientists often use passive techniques, which involve adding physical structures inside the pipes to stir the fluid. One common method is inserting a twisted tape, a strip of metal that spirals down the length of the tube. As the fluid flows over this spiral, it is forced to swirl, which breaks up the thin layer of stagnant fluid hugging the pipe wall and allows heat to escape much faster. However, this swirling action creates friction, which acts like a brake on the flow, requiring more energy to pump the fluid through. The goal for engineers is always to find the perfect balance: getting the most heat transfer possible while using the least amount of energy to overcome that friction.
A team of researchers at Dehradun Institute of Technology and Sankalchand Patel University has tackled this balancing act by combining physical experiments with advanced computer learning. They focused on a specific type of insert called a double perforated twisted tape with V-winglets. Imagine a metal ribbon twisted into a spiral, but this one has two layers and small holes punched through it to let some fluid pass through, reducing the drag. Along the edges of this ribbon, they added small, V-shaped fins that stick out like tiny wings. These wings were arranged in two different ways: either both spiraling in the same direction or spiraling in opposite directions. The researchers wanted to see how these specific shapes performed under different flow speeds and how they could be optimized to save energy.
To understand the behavior of these inserts, the team built a long, clear test tube and ran air through it at various speeds, simulating conditions found in real industrial systems. They tested the air flow at speeds corresponding to a specific range of turbulence, from roughly 10,400 to 22,800 on a scale used to measure flow intensity. They also varied the tightness of the twist in the tape, testing ratios where the tape made a full turn every 2, 4, or 6 units of length. For every test, they measured how well the heat moved, how much the air slowed down due to friction, and how much energy was wasted as disorder, known as entropy. They then fed all this data into a computer system that used several different machine learning models to learn the patterns. These models acted like highly trained students, studying the experimental results to predict how the system would behave in situations they had never seen before.
The results showed that the specific design of the insert mattered immensely. The double twisted tape with the V-shaped wings, especially when the two spirals twisted in opposite directions, produced the best heat transfer. This configuration created strong swirling motions that mixed the air thoroughly, pulling heat away from the tube walls much more effectively than plain tubes or simpler inserts. However, this high performance came with a cost: the friction was also higher, meaning the air faced more resistance. The researchers found that by adjusting the twist ratio, they could influence this trade-off. A tighter twist generally improved heat transfer but increased friction, while a looser twist reduced the friction but offered less heat exchange. The study identified that the version with the V-winglets and counter-swirling orientation achieved the highest overall efficiency, improving the system's performance by about 16 percent compared to other setups when the flow speed was increased.
To make sense of these complex relationships, the researchers used a technique called explainable artificial intelligence. Instead of just letting the computer give a number, they asked the model to explain which factors were most important. The analysis confirmed that the speed of the air and the specific shape of the tape were the dominant forces driving the results. The computer models, particularly one called the Least Squares Support Vector Machine and another called XGBoost, proved to be incredibly accurate, predicting the heat transfer and friction values with a reliability of nearly 99 percent. This means that engineers can now use these digital models to predict how a new heat exchanger will perform without having to build and test every single physical prototype, saving significant time and resources.
Finally, the team used a sophisticated optimization tool to find the ideal operating conditions. Since improving heat transfer often worsens friction, there is no single "perfect" setting that maximizes one while minimizing the other. Instead, the computer generated a list of the best possible compromises, known as a Pareto front. These solutions offer engineers a menu of choices: one might prioritize maximum heat transfer for a compact system, while another might prioritize lower energy consumption for a system where pumping costs are high. The study concluded that by using these data-driven tools, it is possible to design heat exchangers that are not only more effective at moving heat but also more efficient in their use of energy, addressing the critical need for better thermal management in a world that consumes increasing amounts of power.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.