Neutron-Assisted Modelling for Thermochemical Heat Storage

Abstract

In the attempt to solve the problem of ever-increasing energy demand, thermochemical heat storage appears as a promising technology because it allows balance supply and demand of renewable energy power systems and recover low-grade waste heat. To construct a performant heat storage system, great attention should be given to the design of the thermal reactor, the core of the system, since it determines to a great extent its energy and power density, i.e. how much heat can be stored per unit volume and how fast this heat can be recovered. This has set the motivation to develop and validate a three-dimensional accurate numerical model of the reactor, coupling the kinetics of the chemical reaction with heat and mass transfer. In the present PhD project, such a model was developed for a reversible reaction between solid strontium chloride (SrCl2) and gaseous ammonia (NH3). This working pair allows storing large amount of heat at material level, 1579 kWh/m3, at relatively low cost, 3.9-4.9 $/kWh. Moreover, SrCl2 is environmentally benign and abundant. As the first step of the work, the intrinsic kinetic parameters, constituents of the reaction rate equations, were experimentally determined for the sorption reactions between NH3 and Sr(NH3)zCl2, where z varies between 1 and 8. The absorption and desorption kinetic curves were collected on a barometric Sieverts type apparatus using isothermal heating programs over a wide pressure-temperature range. To ensure that during the experiments the reaction kinetics was not limited by the heat and mass transfer, we used an optimal mass of solid SrCl2 inserted into a heat conductive matrix of expanded natural graphite. As a result, the obtained rate equations predicted the experimental data with a good accuracy. Moreover, the results from numerical simulations performed using these equations were found to be in a good agreement with other experimental datasets, which had not been used for the determination of the kinetic parameters. For comparison, numerical simulations with the apparent kinetic equations from literature were found to deviate largely from the same datasets, underlying the importance of using intrinsic kinetic parameters for reactor modelling. During the next step of this work, we developed and validated numerical model simulating NH3 absorption into Sr(NH3)Cl2 and desorption from Sr(NH3)8Cl2. The model was built in COMSOL Multiphysics environment. It was validated against experimental data obtained from neutron radiography, and the results from the simulations were found to be in a very good agreement with the data. The originality of this study is that neutron radiography has been used for the purpose of model validation for the first time. Neutron radiography allowed following the spatio-temporal evolution of NH3 within the reactive bed at each point of the reactor, which could not be achieved by other conventional methods of model validation, like temperature and flowmeter readings. The results from the absorption model were successfully crosschecked using the conventional methods. A distinctive feature of our model is that it is valid not only for the designed SrCl2-NH3 but also for other working pairs and different reactor configurations, as it utilizes intrinsic rather than apparent material properties. The final step of the project focused on the optimization of the reactor by determining the main factors limiting its efficiency. A sensitivity study was performed on the thermal properties of the reactive bed and heat exchanger (thermal conductivity and specific heat capacity) as well as on the heat transfer coefficient between the reactive bed and heat exchanger. Based on the results of this study, practical recommendations for the improvement of the reactor design were made. To confirm and complement these recommendations, a set of sorption experiments, combined with neutron radiography measurements, were carried out using materials with superior and inferior thermal properties. The results revealed that all the studied thermal properties, except for the specific heat capacities of the reactive bed and heat exchanger, have an impact on the desorption rate and have to be considered when designing a reactor. For the absorption rate, the heat transfer coefficient was demonstrated to be the most influencing parameter among the heat transfer properties. In addition, it was also shown that absorption is a mass-driven process, and the parameter having the greatest impact on its rate is the permeability of the reactive bed.