Quantitative evaluation of the natural convection effect on thermal conductivity measurement with transient plane source method

When measuring the thermal conductivity of liquids with transient methods, natural convection occurs inevitably and whether it can be ignored is unknown. The quantitative deviation introduced by natural convection should be revealed clearly to ensure reliable experiment. In this study, the influence of natural convection is quantitatively evaluated in measuring the thermal conductivity of liquids with transient plane source method. By numerically mimicking the transient heat transfer process within liquids, the sensor temperature responses with neglecting and considering natural convection are obtained and used to calculate the corresponding thermal conductivity. The difference between them represents the quantitative influence introduced by natural convection. The effects of sensor orientation, heating time and heating power are compared for lead-bismuth eutectic alloy, GaInSn alloy, deionized water, molten salt, glycerol and cupric sulfate. Experiments are performed to verify the simulation. It is found that for the liquids with lower Prandtl number (<10), the thermal conductivity measured by horizontally placed sensor is more precise than that by vertically placed sensor. On the contrary, the liquids with Prandtl number higher than 1000 are less affected (<1%) by natural convection. Shorter heating time and proper temperature rise magnitude can suppress the influence of natural convection to some extent.