Optimization of parameters affecting surface roughness and experimental investigation of rough microchannel flow

In the study, rough microchannel heat sinks were produced and the effect of production parameters on roughness was investigated. It provides optimization of the effects of changes in processing repetition, roughness, hydraulic diameter and aspect ratio on microchannel heat sinks on issues such as wing structure, flow characteristics and heat transfer. The experimental findings have been compared with traditional correlations and create an infrastructure that will shed light on researchers who will encounter similar problems and may serve as a reference to the branches of industry that will work on the subject. To optimize the roughness parameter affecting the system performance in a microchannel heat sink. Microchannel heat sinks were produced and tested at Sakarya University Development and Application Center (SARGEM) Microchannel Heat Transfer Laboratory. The heat sinks used in the tests were made of aluminum and copper materials. There are 15 parallel channels in each heat sink, and the channels are manufactured by the wire erosion method. Channels with different roughness properties were produced by keeping the processing and feed speeds constant and changing the processing repetition. The channel dimensions of the heat sinks were determined by scanning electron microscopy (SEM, Jeol JSM 6060 LV). Average values and hydraulic diameter were calculated by measuring the channel width and length in at least 3 different channels. Channel roughness measurements were made with a KLA Tencor P6 standard profilometer device. When the roughness values (Ra) were examined, the highest roughness value was measured as 7.35 µm, the lowest roughness value was measured as 0.5 µm, and it was generally seen that the data was below 3 µm. The average roughness value of all experiments was 3.71 µm. A total of 18 different cooler samples were prepared, each with external dimensions (20 mm x 20 mm), different aspect ratios and roughness values of the channels. Hydraulic diameters used in calculations were obtained from SEM images. Taguchi experimental design method was used in this study. 18 heat sinks with hydraulic diameters ranging from 400 to 500 µm were produced from Al and Cu materials. These thermocouples have parallel microchannels with rectangular cross-sections. The average roughness values (Ra) of these micro channels vary between 0.5 and 7.5 micrometers, and their aspect ratios vary between 0.4 and 2.5. The roughness parameter affecting the system performance in a microchannel heat sink was optimized with the Taguchi method. In optimization, optimum signal-to-noise ratios for material type, design and processing factor were used. A formulation that will be a reference in finding the roughness value for microchannel studies with similar hydraulic diameters was obtained. It was observed that the pressure and temperature values were consistent with the traditional theoretical calculations and the experimental results, but there were differences in the friction factor. In the study, 18 experiments were conducted as required by the experimental design and a total of 18 observation was taken. Surface roughness measurement in terms of Ra value was repeated 3 times for each sample and the average was taken. The collected data were analyzed according to the "smallest is best" criterion in the Taguchi analysis section of the Minitab program and the necessary data and graphics were obtained. In the application of Taguchi experimental design method, the most ideal parameter for black image ratios is Copper as the material type and 3 cuts in the A type Processing repetition in the Design. The most influential parameter in the Anova method is Design with %49.20, %27.4 and Processing repetition with 12.87%, respectively, material type parameters. The reliability of our Anova model was %89.62. The fitness value, which is generally accepted in the literature, is above %85. For this study, we can reach the conclusion that the model can be applied. In polynomial regression analysis, it has been shown that if you experiment with products of different designs with one of 2 different materials, the roughness values and regression equation can be reached without experimenting with the given formula with different processing repetitions. Surface Roughness (µm) = - 0.1154 + 1.070 Estimated Surface Roughness (µm) - 0.00876 Estimated Surface Roughness (µm)2 The main elements of the experimental setup are water tank, pump, filter, chiller experimental section, heat exchanger, water bath, valves, power supply, measurement and data collection element. A glass plate is used on top to obtain an image of the flow and is fixed to the stainless steel plate with stud nut sets to ensure tightness. Deionized water was used as the working fluid in flow experiments. The heat sinks were tested under constant heat flux conditions. In single-phase experiments, the inlet temperature of the fluid was set at 20°C with a constant temperature bath. Inlet and outlet fluid temperatures and wall temperatures were measured with K type thermocouples. Pressure drop was measured with a high-precision Keller pressure transmitter in the range of 0-1 bar placed at the inlet. Cartridge heaters were used, where the total heating power could reach 123 W, and the heating power was regularly checked with a power supply and digital watt meter at full input power. Heat losses were determined by a method used in the literature. To determine heat losses, the power supply was set to a constant thermal power value and temperatures were measured. Temperatures were measured again at different thermal power settings. The heat loss calibration curve was obtained through measured temperatures. In flow experiments, the tests were repeated at least three times under the same conditions and were instantly recorded to the computer via the data collector. The experiments carried out under single-phase flow conditions were in the range of 155