Entropy Generation Minimization and Heat Transfer Enhancement for CSP Systems

Abstract

Advancements in clean, inexhaustible energy technologies remove humanity’s reliance on fossil fuels which are an unsustainable resource. Solar energy, unlike other forms of renewables, has a consistent source of energy, the sun. Solar panels and concentrated solar powered systems are implemented to harness sunlight energy. However, energy is wasted when there is an overproduction of solar energy during the day. This energy can be stored to be used from sunset until dawn when there is no sunlight. Solar panels require batteries for energy storage and provide additional initial and maintenance costs to plant owners and operators. Concentrated solar power systems have an advantage since they may provide power to the grid directly or store energy in the form of thermal energy. Thus, this research will focus on concentrated solar power systems. The aim of this research will be to maximize the amount of available energy by minimizing entropy generation and reducing thermal stress of the receiver as well as increasing the thermal efficiency of the receiver. In previous work, which will be referenced, Blasius and Dittus-Boelter correlations were used to determine the minimum entropy generation rate for a smooth tube. This study will involve determining the minimum entropy generation rate for the molten salt system NaCl/KCl/MgCl2 using the Petukhov and Gnielinski correlations which are proven to describe friction factor and heat transfer data more accurately. In addition, the entropy generation rate of twisted-tape inserts was investigated. The entropy generation of the twisted-tape insert was found to be less than that of the smooth tube for flow velocities less than 4 m/s. Also, the ternary salt, NaCl-KCl-MgCl2 entropy generation rates were slightly lower than the binary salt, KCl-MgCl2. Furthermore, a general method for determining the minimum entropy generation rate for both smooth and enhanced tubes whose friction factor and heat transfer correlations are known is presented; assuming the heat rate, heat flux, inlet temperature, and outlet temperature are fixed. Additionally, friction factor and heat transfer correlations were developed for two novel 3D printed tube designs, a single-head helically finned tube and multi-head helically finned tube, that were previously tested via CFD methods. These correlations tested Reynolds numbers from 3,900-61,300 and Prandtl numbers from 4.26 to 6.26. The single-head helically finned tube had 2.3-2.8 times the heat transfer enhancement of the smooth tube while the multi-head helically finned tube had 1.6-2.3 times the heat transfer enhancement of the smooth tube for Reynolds numbers above 10,000. The significant heat transfer enhancement of the single-head helically finned tube is accompanied by significant frictional losses of 52-61 times that of the smooth tube for Reynolds numbers above 10,000. The multi-head helically finned tube showed much lower friction losses of 4.8-8.1 times that of the smooth tube for Reynolds numbers above 10,000. The experimental results were also compared to the CFD simulations. The CFD simulations agreed well for the experimental heat transfer results of both tubes and the friction factor experimental results of the multi-head helically finned tube. However, the CFD prediction of the frictional losses of the single-head helically finned tube were 37-46 times that of the smooth tube. The difference between the CFD simulation and experimental results may possibly be attributed to greater surface roughness of the single-head helically finned tube. These same tubes are investigated in the laminar regime via CFD methods. The Moody friction factor, average Nusselt number, and entropy generation rate were determined. The entropy generation rate results were then compared to other enhanced tubes that were tested in the laminar regime. The heat transfer enhancement for the single-head and multi-head helically finned tube were 3.46 to 7.24 and 1.01 to 1.17 relative to the smooth tube, respectively. The frictional losses of the single-head helically finned tube were 16.3 and 40.5 relative to the smooth tube which were significantly larger than the multi-head helically finned tube frictional losses of 1.23-1.48. The entropy generation of the single-head helically finned tube was lower compared to the multi-head helically finned tube despite having much larger frictional losses. Lastly, a derivation of the optimum Reynolds number and normalized entropy generation rate per unit length for a thermally developing flow were presented in Chapter 5.