Experimental and theoretical investigation of heat transfer in metallurgical furnace sidelining cooling system

Abstract

The purpose of this work was to investigate theoretically and experimentally the heat transfer of a metallurgical furnace side-lining cooling system utilizing evaporation/boiling and condensation of a suitable working fluid. The working temperature range was specified to 400-650ºC. The heat flux was specified to 20- 30 kW/m2 under normal working conditions, under special conditions hot spots with heat fluxes up to 300 kW/m2 were to be expected. The heat flux on the cooling panel was specified as unilateral, as it would be in the side-lining of a metallurgical furnace. The number of working fluid candidates for the temperature range 400-650ºC is quite limited. After reviewing literature on working fluids, the following two candidates were selected for further investigation: • Sulfur • Potassium Sulfur was from its thermophysical properties considered as a candidate for a cooling system utilizing nucleate boiling, however, the corrosivity of boiling sulfur was a major concern. A corrosion experiment was carried out, in which samples of stainless steels AISI 316L and AISI 304 were submerged in boiling sulfur for 100 hours. The experimental results showed that the corrosion was heavy. Without any construction material known to be able to withstand boiling sulfur for years, it was decided not to proceed with the investigation of sulfur as working fluid for the cooling panel. Potassium was investigated theoretically as a working fluid candidate for a cooling panel utilizing nucleate boiling and for a cooling panel utilizing evaporation from a wick. From the theoretical study it was concluded that the nucleate boiling of potassium in a narrow panel was likely to be periodic, with relatively long waiting times followed by shock boiling. How the boiling would develop over the required lifetime of the cooling panels (years), with stops and restarts, was also a major concern. A cooling panel utilizing evaporation of potassium from a wick was therefore considered to be the best solution. With regard to construction material both AISI 304 and nickel was reported in heat pipe literature to be compatible with potassium. A novel hybrid heat pipe design was chosen for the cooling panel. The panel is a hybrid because only the evaporator surface is covered by a wick. The condenser is wickless, and the condensate is flowing back to the evaporator by gravity, like in a thermosyphon. The construction material was choosen to be Nickel 201, due to its high thermal conductivity compared to AISI 304. A theoretical and experimental wick study was undertaken in order to develop a suitable wick for the hybrid heat pipe evaporator surface. For the experimental hybrid heat pipe the wick was required to manage a lifting height of 150 mm, and be able to withstand hot spot heat flux conditions. Two types of wicks were made and analysed experimentally and theoretically in this study: • Wicks of sintered nickel powder (Inco 255). • Wicks of sintered compressed nickel foam plates. The effective pore radii and the permeabilities of the wicks were determined from rateof-rise experiments using the model fluid heptane. The porosities of the wicks were measured by use of isopropanol. The experimental results were applied to predict the performance of the wicks in the hybrid heat pipe. The results showed that the performance of the wicks made of sintered nickel powder would be very limited by the vapour static pressure limit. The vapour static pressure limit says that in a closed container (i.e. here the hybrid heat pipe) the capillary pressure produced in the wick can not be higher than the vapour pressure of the working fluid. The wicks produced of compressed nickel foam, having higher effective pore radius and higher permeability than the wicks made of sintered powder, were therefore considered better for the hybrid heat pipe. A series of 2 layer compressed nickel foam wicks was produced in order to investigate how the effective pore radius, permebility and porosity changed with the degree of compression. The experimental results were new for this type of wicks. From the maximum capillary rise of heptane in these wicks the contact angle between heptane and the nickel of the wicks was determined to be 58,4º±8,9%. The final wick for the hybrid heat pipe was made of 4 layers of compressed nickel foam sintered together. The total thickness was 1,15 mm. From the rate-of-rise experiment the effective pore radius was determined to 1,02·10-4 m (for heptane) and the permeability was 7,30·10-11 m2. The contact angle between heptane and the nickel of the wick was estimated to 60,2º from the capillary rise in this wick. The final wick was attached to the evaporator surface of the hybrid heat pipe by electron beam spot welding. The consequence of the welding spots on the performance of the wick was analysed by use of Comsol Multiphysics®, and the spots were found to cause minor (~1%) reduction of the wicks capacity. A special high vacuum rig was designed for filling potassium into the hybrid heat pipe and sealing it by use of induction welding. The high vacuum and induction welding parts were tested successfully, however, the filling process failed at other stages due to the high reactivity and stickiness of potassium. Without time and resources to solve these problems it was decided to have the potassium filled into the hybrid heat pipe at an external laboratory. A test rig for the hybrid heat pipe was built, utilizing cartridge heaters for heating of the evaporator and convective cooling by use of nitrogen gas for the condenser. In order to monitor the performance of the hybrid heat pipe it was equipped with thermocouples. A predictive thermal performance model based on existing theory was built using Excel and was utilized to produce semi-quantitative estimates of the fluid flow in the hybrid heat pipe. The experimental results were scattered but indicated that the hydrid heat pipe was able to operate at higher heat fluxes than expected. Specifically, the measured performance could occur if the permeability of the wick was higher than calculated from the rate-ofrise experiment. Increased permeability could have been caused by a small channel in the wick. Periodic instability of the temperatures in the hydrid heat pipe occurred when heat was supplied to the two lower evaporator sections only. A model of explosive boiling was found to fit the measured instability well. The thermal expansion of the hybrid heat pipe was analysed by use of Comsol Multiphysics, and showed that structure bending due to unilateral heat flux is a factor to consider even when a high thermal conductivity construction material as nickel is used. Thermal expansion has potential to cause thermal contact problems both on the outside of the cooling panel, and between the evaporator wall and the wick. Thermal expansion could have created a high permeability channel in or under the wick in the current study. A robust method for attaching the wick to the evaporator surface is required. From this study the hybrid heat pipe technology is considered as promising for a metallurgical furnace side-lining cooling system.