Simulation of nanofluid micro-channel heat exchanger using computational fluid dynamics integrated with artificial neural network

Abstract

Waste heat utilization has been prioritized especially in various industries and sectors. Many researchers have developed heat recovery processes by designing suitable waste heat recovery units (WRU), such as heat exchangers, using water as a coolants to receive heat from the waste heat fluid in the production process. The conventional heat exchanger has limitations such as its equipment size, space for installation, and flexibility. The microchannel heat exchanger is one of many ideas for resolving these limitations. Moreover, the coolant on the cold side can be upgraded by adding nanometer-sized solid particles which is called “Nanofluid”. To reduce the high investigation cost and time, a new efficient and cost-effective simulation method was selected to use for investigating the performance of a microchannel heat exchanger with nanofluids in this study. To analyze the heat recovery at low temperature, i.e. around 100–200 °C, nanofluid property predictive models were developed using an artificial neural network (ANN). Then, the predictive models were embedded and integrated into computational fluid dynamics to design a microchannel heat exchanger. It is found that the use of nanofluids improved the heat transfer efficiency of this heat exchanger. The suitable nanofluid types and concentrations were selected based on the thermal–hydraulic​ performance. Here, the 3% weight TiO2/Water fluid with a 1.03 thermal–hydraulic​ performance ratio was found to be the most promising nanofluid for using in this condition.