A steady state model of a thermoelectric heat pump fitted with pin fin heat exchange surfaces is developed from the equations governing the operation of a thermoelectric device and pressure drop/ heat transfer correlations for staggered tube banks in cross flow. The model computes current draw, heat pumped at the cold side, heat rejected at the hot side, hot and cold side exhaust temperatures, and hot and cold side pin base temperatures when given the following parameters: ambient temperature, thermoelectric geometry factor, number of thermocouple junctions, input voltage, pin diameter, pin height, transverse pin stagger, longitudinal pin stagger, contact resistance between the thermoelectric modules and the heat ex change surfaces, and fan performance curves for the hot and cold sides. Results of a FIDAP numerical model of a 4 row heated pin system are presented and compared to the heat transfer and pressure drop correlations chosen for use in the model. Nine tests were run with a working thermoelectric heat pump unit under various air flow and voltage input to collect data for the purpose of comparison with the model. Using empirically determined contact resistance values, the model predicted Qc (heat pumped from the cold side) values that agreed with experiment within about 10%, Qh (heat rejected from the hot side) values that agreed with experiment within about 5%, and current draw values that agreed with experiment with in about 5%. Further examination of the contact resistance values suggested the presence of an undesirable air gap at the heat sink/thermoelectric module interface. The presence of this air gap was confirmed by examining the contact patterns left on pressure sensitive films that were placed between the thermoelectric modules and the heat sinks. An optimization exercise was performed with the model in an effort to maximize Qc. Pin geometry was varied to achieve minimum values of heat sink thermal resistance for both the hot and cold side heat exchange surfaces. Using these minimized thermal resistance values, input voltage, thermo electric geometry factor, and number of thermocouple junctions were varied to achieve maximum Qc. Using the empirically determined thermal contact resistance values, a 29% increase in Qc was achieved over the stock case. The model predicted a 79% improvement over the stock case when utilizing thermal contact resistance values corresponding to the elimination of the air gaps at the thermoelectric module/heat sink interfaces.

Library of Congress Subject Headings

Heat pumps; Thermoelectric apparatus and appliances

Publication Date


Document Type


Department, Program, or Center

Mechanical Engineering (KGCOE)


Mulay, Shirish

Advisor/Committee Member

Hefner, Robert

Advisor/Committee Member

Kandlikar, Satish


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TJ262.H47 1994


RIT – Main Campus