Full optimal geometric exploration of M-cycle indirect evaporative dewpoint cooler using a substantial experiment-oriented AI algorithm

• Maisotsenko cooler can supply sub-wet-bulb temperature under specific geometries. • AI-based algorithm is generated for optimum geometric exploration. • All possible behaviours from geometric parameters are identified and analysed. • Some critical values were detected for some geometric parameters. • Remarkable and interesting results were observed. The most interesting feature of Maisotsenko indirect evaporative cooler (MIEC) is sub wet bulb supply temperature via water vaporization while adding zero humidity to the final product air. However, for a given thermal/fluid condition, achieving sub wet-bulb temperature while maximizing supply air mass flow is only possible within a unique range of geometric factors. Besides, the influence of each geometric factor on the performance of the cooler changes as any other geometric parameter varies. If only five values are considered for each parameter, 625 geometric combinations are possible. Artificial Intelligence seems the only helpful tool to handle such cumbersome and complex situations. Hence, in this research, first, an experimentally validated analytical model is employed to train a machine learning (ML) model considering geometric parameters of the cooler. Then, the ML model is employed to generate all possible geometric combinations of the cooler. It is noted that the inlet air thermal conditions, dry side airflow, and wet-to-dry side airflow ratio are fixed on certain values, making the absolute results valid only for these thermal/fluid conditions. For different ambient air or fluid flow conditions, the optimal geometry will vary, requiring the ML model to be re-run. However, the general trends of the curve may still serve as a guide. According to the results, the impact level of each parameter significantly depends on the values of other parameters as well. For example, while a larger channel length leads to a colder supply temperature, the extent of its impact depends on the channel gap and channel height. When the wet and dry channel gaps are at their minimum (0.5 mm), increasing the channel length from 0.05 to 0.25 m results in a supply temperature reduction from 25 °C to around 13 °C (for a channel height of 0.4 m) and from 16 °C to around 11 °C (for a channel height of 1.2 m). This represents a reduction of about 50 % in the first case and approximately 31 % in the second case. However, when the dry/wet side gap is larger (4.5 mm), increasing the channel length from 0.05 to 0.25 m causes the supply temperature to decrease from around 35 °C to about 26 °C (for a channel height of 0.4 m) and from 30 °C to around 18 °C (for a channel height of 1.2 m). This results in a reduction of approximately 25 % in the first case and about 40 % in the second case. For industrial consulting contact the corresponding author.