Numerical investigation on the scale-down approach of Harwell furnace

Conducting experimental studies with prototypes is both costly and time-intensive. Scaled-down models can be used for initial experiments. However, scaling requires high precision and clarity. The present study aims to evaluate and compare various scaling approaches for predicting prototype results and to relate the temperature distribution of the scaled model to that of the prototype. Steady-state simulations were carried out using the k–ε turbulence closure model and the probability density function approach. Temperature profiles and stream traces from different scaling methods were compared to identify the optimum scale-down method. Simulation results showed a change in the position of the combustion core; the radial temperature profiles for scale-down models with 50% heat input showed a significant deviation of about 85% at an axial position of 0.1 m, while for the rest of the axial positions, the difference was less than 7%. The results reveal that geometrically larger models align more closely with experimental data, particularly when using the Constant Residence Time (CRT) method, compared to smaller-scale models. To account for the dependency of temperature on energy input and energy release at specific locations within the combustor, a non-dimensional temperature variable, θ was introduced. A comparison of results indicates that the CRT method effectively scales down the combustor, showing strong agreement with experimental data from the literature. The variation in θ for scaled-down models aligns closely with prototype and experimental results, with maximum deviations of 17%. This suggests that θ is a novel and effective variable for establishing reliable connections between scaled-down models and prototypes.