Thermal Batteries are primary one-shoot high-power energy sources which, due to the molten salt used as an electrolyte within their galvanic cells, are operated at elevated temperatures (300-700 ^oC). Thanks to the use of a high-temperature molten salt (e.g., LiCl-KCl) rather than an ionic solution, large electrical currents may be drawn from these devices. However, the high operational temperatures require an intelligent management of heat transfer in the system, as electrical current cut-off will follow salt (electrolyte) solidification thereby determining the operation time in which current may be supplied. It is often due to heat transfer limitations that thermal batteries, consisting of serial-connected and thermally-insulated cells, usually function as short time power sources for different autonomous systems. Thermal design is therefore of utmost importance, involving optimization of the geometrical arrangement, cell structure and insulation materials. Meeting performance requirements, while being constrained by size and ambient atmosphere specifications, becomes a time and money consuming process using an experimental trial and error approach. On the other hand, the design process may be significantly enhanced using battery performance simulations based on a reliable mathematical model.

In this contribution, we present our in-house developed thermal simulator, TThermBat. The simulator is based on a transient mathematical model of a two-dimensional axis-symmetric geometry, to be used as a versatile and robust simulation tool for thermal battery design and analysis. This thermal model enables a portable and detailed simulation from the level of each cell sub-component (e.g., cathode disc) up to that of the entire product, considering various heat transfer phenomena, including conduction, phase-change (electrolyte salt solidification), heat of reactions and Joule-Heating; heat loss via convection and radiation to an ambient temperature is applied at the model system's external boundaries. In addition, the heat transfer model is supported by an overall (albeit simplified) mass balance, involving the electrical current drawn from the battery and an integral mass transfer resistance coefficient for each cell. Model calculations have been successfully validated through comparison with well-established analytically and numerically solved test-cases, as well as with experiments using a “dummy” (1-component inert-cell) battery.

A leading characteristic of our simulator is its versatility both in terms of geometry and in terms of material systems. For example, the virtual geometrical model is constructed by a MATLAB Graphic User Interface (GUI), allowing the user to define the battery's inner and outer structure with great flexibility; this includes defining the cell ingredients, the number of electrochemical-cells, and details of the insulation layers. The GUI also supports an extensive and upgradable material property library, in which each material is characterized by a set of temperature (and, in some cases, concentration) dependent physical properties, to be applied in each domain of the geometrical model. Model calculations are numerically performed and analyzed via the Finite Element Method (FEM). Finally, the GUI enables various post-processing options: temperature colormaps, dynamic animations and temperature-time plots with their corresponding temperature-mole fraction paths over the binary salt phase-diagram (when relevant). In addition, an approximate electrical cell analysis, based on the narrowing cross-section for electrical charge transfer due to the salt solidification is enabled.

Results of the TThermBat simulator have been achieved for both a single cell and a full scale thermal battery, uncovering a number system-level effects. Future plans involve pursuing the development of a new simulator, EThermBat, which is based on a rigorous differential mass transfer analysis for the prediction of electrical performance of thermal batteries at both the single cell and the system level.