Faster, more comprehensive water boiling simulations

Making coffee is considerably simpler than simulating it. There are a seemingly endless number of factors, from the water temperature to the fineness of the grounds to the soil pH where the beans were grown. Simulating every factor is computationally, financially, and temporally unrealistic, and the result will often still miss the system’s nuance.

Silkie Melloul and Benjamín J. Alemán leveraged a powerful statistical technique called factorial design and analysis to simulate the effects of five factors on water boiling time. The whole study only required 32 runs — the same number needed to assess the impact of a single factor in a traditional simulation.

“When you try to understand the complexity of what goes into a simulation, at some point you have to make simplifications,” said Alemán. “Factorial analysis picks and chooses certain combinations of factors that might go into a simulation so that we can understand how the whole thing works without having to do a lot of simulations.”

The researchers began by creating a computer design of their finned water heating system, then importing it into physics software that applied the proper heat transfer equations. Then, they ran simulations with factorial analysis that determined the respective impacts of the heating system’s geometry, spacing, thickness, material, and fin configuration on the water boiling time. Finally, they created a mathematical model with the parameters their simulations suggested and validated its accuracy against experimental water heating data.

“Other researchers around the world can use this approach on simulations to save time, energy, and cost, and at the same time get a better, more powerful prediction into the system they care about,” said Alemán.

Source: “Pairing factorial design with finite element analysis to model and optimize heat transfer in finned heatsinks,” by Silkie L. Melloul and Benjamín J. Alemán, Journal of Applied Physics (2025). The article can be accessed at https://doi.org/10.1063/5.0285136 .