A simplified proof of concept is provided to demonstrate the potential capabilities of TEPSS. Simulation of the simple system shown below is performed by the first prototype of the TEPSS software.
The simulated system consists of three components (a blower, a combustor, and a thermoelectric power unit) and five domain state nodes coupling the components as shown. The first prototype only uses fluid nodes, although electrical nodes could be used in the future for the case where the blower might be directly driven by the thermoelectric power unit. Ambient air enters the blower which drives the flow through the system. The air then passes through the cold side of the power unit before entering a combustor where the air is heated. The heated air then flows through the hot side of the thermoelectric power unit before exiting to ambient. This system is simple and demonstrates electric power recovery from waste heat. For this system there might be interest in maximizing net output power or efficiency or minimizing material cost for some desired net power. Here the net power is defined as the difference between power generated by thermoelectrics and power consumed by the blower.
Component models were developed independent of the other component models to ensure that they are reusable in other systems. Component model descriptions follow:
Power Unit Model
The power unit is divided into zones, each with the possibility of any number of modules configured thermally in series and parallel as shown below.
The model assumes 1D heat transfer within each zone and is based on a model developed by Smith . The basic thermal resistance network for each zone is shown below. The Rcomb’s are the thermal resistances between the fluid and the surface of the thermoelectric modules. This resistance is calculated based on the fin geometry, fin material and the flow conditions.
The model developed for the thermoelectric modules is the standard 1D model. The model can either accept module level parameters such a Seebeck coefficient and electrical and thermal resistances or module geometry (leg length, cross sectional area, insulating plate thickness, etc.) and material property data which can be temperature dependent. For the current case, it was assumed both sides of a leg couple had identical properties and dimensions.
The model also calculates the pressure drop across both sides of the power unit which restricts the system flow. TEPSS allows the user, and eventually the optimization shell, to quickly vary user identified design parameters and run multiple simulations. Parameters of interest for this particular case are the TE leg and fin lengths.
Blower and Combustor Unit Models
Both the blower and combustor models for the proof of concept are basic. The blower assumes a user defined efficiency in converting electrical power to fluid power. The temperature increase across the blower is assumed to be negligible. The input power is currently assumed to be an input design parameter. For the combustor, the outlet temperature is assumed to be dependent on the mass flow rate to ensure an adequate air supply for combustion of methane. Also, the combustor outlet temperature is constrained to be at or below 250°C.
Parameters are provided by the user to the simulator which determines the dimensions of the power unit, geometry of the fins, and dimensions and operating parameters of the thermoelectric modules. Standard geometric and material property values were used for this simulation.
The module leg length and fin length are varied over the design space of 0.5 to 10 mm and 5 to 80 mm, respectively. There are several trade-offs when varying these parameters. Reducing the thermoelectric leg length increases power by reducing overall electrical resistance but decreases the temperature difference across the module, thereby reducing module efficiency due to contact resistance. The fin length impacts the pressure drop hence flow rate and thermal resistance between the fluid and the surface of thermoelectric modules. The figures below show the total net power and efficiency per unit of thermoelectric material. As can be seen there is clearly an optimal leg length for this particular system as well as an optimal fin length when considering efficiency per unit of thermoelectric material.