Dottorato in PROCESSI CHIMICI PER L'INDUSTRIA E PER
L'AMBIENTE

Titolo della tesi: Study and validation of innovative heat transfer fluid to be used in power systems heat exchangers

This thesis is concerned with the investigations on chemical properties and stability of thermal oils used as cooling agents. In particular, the main target was to verify the operating conditions of heat transfer fluids to be employed in the contest of IV generation nuclear plants, where liquid sodium is the primary cooling medium. In this contest, the presented activities were carried out inside the ARDECO project, part of ASTRID Research and Development Cooperation, within a CEA-ENEA joint agreement. Thermal oils have been utilized as heat transfer fluids for several decades and are currently used in many applications including industrial facilities, power plants and solar receiver systems. Despite their large employment, very few information are available about oils behavior under thermal cycles, in particular regarding chemical and physical properties stability. Actually, almost all the data present in the scientific literature are focused on the thermal oils resistance in presence of radiochemical irradiation, but practically no results were so far reported about their performances at high temperatures. Given the applications in the nuclear field, it is necessary to select materials that present sufficient high radiochemical stability and thermal resistance. At this aim, synthetic oils were considered, namely, two aromatic based coolants, that is, Therminol 66 and Therminol SP, together with an innovative silicone based fluid, Bluesil FLD 550 HT. The formers are largely used heat transfer media and present good thermos-physical properties and an acceptable cost; the latter, while very promising due to its chemical stability, it is currently too much costly and it was investigated for research purposes, also considering the total lack of information about this material. For these reasons, a dedicated experimental and modeling campaign was planned and carried out involving three interconnected steps. Firstly, extensive laboratory scale tests were performed in order to determine the changes on chemical composition and physical features after properly conducted thermal stress treatments. Secondly, an experimental pilot scale loop was designed and constructed at ENEA Casaccia Research Center, to study under realistic conditions the actual thermal exchange capacity of the three thermal oils investigated. Finally, a fluid dynamic model was developed and validated with the obtained experimental data. Regarding the laboratory scale activities, stability tests were performed at different temperatures and periods, until 3 hours. The operating conditions were set using as starting temperature the limit value provided by the manufacturer. Hydrogen was the main decomposition product for both Therminol 66 and Therminol SP, and starts to be significantly formed at around 350°C. As far as Bluesil is concerned, also, methane was detectable along with Hydrogen, but its evolution was present only above 400°C, showing the greater chemical stability of this fluid. The condensable byproducts mainly consisted of substituted benzenes for the aromatic fluids and siloxanes cyclic compounds for Bluesil. The thermally stressed oils were then analyzed with respect to their thermo-physical features and compared with the correspondent fresh samples. Thermogravimetric analyses pointed out a polymerization process for Therminol 66 and cracking reactions for the other two oils. Viscosity measurements confirmed these assumptions; in fact, Therminol 66 indicated an increase of this parameter (up to 77 % more at 40°C) and Therminol SP exhibited a significant drop (20% lesser at 40°C). Regarding Bluesil, a slight enhancement was detected, likely related to the formation of cross-linked polymers after methane release. Specific heat and infrared spectroscopy results did not show differences between fresh and aged samples, pointing out that the variations in the oils chemical structures are anyway limited. Finally, mass spectrometric analyses were carried out for Therminol 66 and Therminol SP, and remarked the loss of low molecular weight species during degradation. Given its polymeric conformation, Bluesil was not suitable to be investigated by this technique. The pilot scale test rig was designed in order to perform thermal cycles from 130°C to 200°C and vice versa, that is, in safe conditions and very close to a real scenario. Actually, the requirements for the nuclear cooling systems concerned in this work were to identify a proper material and a suitable heat exchange geometry capable to remove thermal heat from liquid sodium. At this aim, the oil loop was configured as a single pipeline where a thermal oil is recirculated at a determined velocity, and eventually cooled back by an external plates system containing a water-glycol mixture. Each test lasted 48 hours and the system was kept under Nitrogen flow and the heat exchange section was electrically heated. In order to establish the necessary length for the test section and size the electric heater, it was necessary to preliminary simulate the heat exchanger behavior. For this purpose, the oils velocity was set at 0.9 m/s, which is the minimum velocity requested for ensuring a turbulent flow inside the pipeline and, at the same time, allowing the maximum theoretical thermal energy transfer. The model results permitted to settle the needed test section lengths for each oil tested. Regarding Therminol 66 and Therminol SP, the experimental results confirmed the calculated predictions, and it was possible to obtain the required temperature interval with both fluids. On the other hand, there was a necessity to decrease the applied power for Bluesil in order to achieve the targeted oil inlet temperature; however, despite the lower resulting temperature range (154-190 °C), Bluesil showed good thermal exchange properties. The measurements carried out on the sampled gases were in good agreements with the laboratory scale results, showing that, at the loop temperatures, neither hydrogen nor methane was produced during the test periods. The stressed oils were analyzed after the 48 hours and compared with the correspondent fresh specimens, showing only slight modifications. Moreover, the obtained results matched with the absence of hydrogen evolution, showing a substantial stability of the thermal oils at the temperature ranges used in the loop. As final experimental results, it was also possible to calculate an average heat transfer coefficient by the experimental data. In particular, this parameter was determined considering the external wall temperatures and the correspondent values for the oils taken at the center of the tube diameter. The effect of the carbon steel thickness was neglected; the results showed that, regarding Therminol 66 and Therminol SP, the actual wall temperatures were smaller than the ones assumed for the MatLab calculations. On the contrary, the experimental coefficient for Bluesil was in very good agreement with the theoretical figure. It was then possible to utilize the experimental results to implement a predictive tool developed by the ANSYS FLUENT software. The obtained model allowed highlighting a detailed temperatures and fluid velocities pattern along the heat exchange zone. To conclude, the considered heat transfer fluids are definitely promising media for the applications concerned, although the silicone based material is probably still too costly for large scale applications. It is also noteworthy that the achieved data could be feasible in other fields different from the cooling of nuclear systems, in particular, for concentrating solar power plants.