• Energy Research
• Closed Access close

When a number of TEMs (thermoelectric modules) are connected in a series-parallel matrix and under mismatched temperature gradients, the overall maximum output power of the thermoelectric generator (TEG) may be lowered by certain TEMs with relatively smaller temperature difference. It is possible to avoid such a performance decrease by the disconnection of these low temperature TEMs, provided that the critical power point can be accurately determined. In this paper, firstly a rigorous and universal formulation is fully detailed to mathematically determine the conceptions and conditions of the critical power point in the series and parallel TEM arrays. Secondly, experiments of a series-parallel hybrid interconnected TEG are presented to clearly quantify the theoretical analyses. Finally, the hierarchical simulation, based on the SPICE (simulation program with integrated circuit emphasis) platform, is applied to estimate the critical power point. By numerically modeling the nonlinear physical processes of the TEG, the simulation can be used as an enabling technique in any model-based controller to dynamically minimize the mismatch power loss within the TEM matrix of any configuration. In experimental and numerical results, a number of critical power points are disclosed for a 2×4 parallel-serial hybrid TEM matrix, where the hot temperature mostly ranges from 120°C to 60°C.

To maximize the energy productivity, effective in-field detection and real-time control of defective thermoelectric modules (TEMs) are critical in constituting a thermoelectric generation system (TEGS). In this paper, autonomous and distributed sensor nodes are designed to implement the wireless TEM management in terms of the measurement criteria of defective TEMs formulated for series-parallel-connected TEM arrays and the control scheme based on the TEM-oriented switches. The instrumentation of a TEGS prototype and the design of the embedded software associated with the sensor nodes are described, respectively. Defective and potentially healing conditions are dynamically monitored by a voltage sensor node and a temperature sensor node, both of which can judge the defective TEM and decide the related switching actions in a nearly independent way. The periodical wireless transmission from the nodes to a base station is no longer necessary, and with the minimized amount of communication signals, the battery lifetime of the distributed nodes can be significantly prolonged. In the experimental tests, the autonomous sensor nodes successfully disconnect and reconnect the defective TEMs, where a considerable power improvement is illustrated with the proposed measuring method and setup.

An accurate mathematical model of thermoelectric modules (TEMs) provides the basis for the analysis and design of thermoelectric conversion system. TEM models from the literature are only valid for the heat transfer of N-type and P-type thermoelectric couples without considering air around the actual thermoelectric couples of TEMs. In fact, air space imposes significant influence on the model computational accuracy, especially for a TEM with large air space inside. In this study, heat transfer analyses of air between the TEM cold and hot plates were carried out in order to propose a new mathematical model that minimises simulation errors. This model was applied to analyse characteristic parameters of two typical TEMs, and the ratio of cross-sectional area of air space to thermocouples were 48.2% and 80.0%, respectively. The average relative errors in simulation decreased from 5.2% to 2.8% and from 12.8% to 3.7%, respectively. It is noted that our new model gives result more accurate than models from the literature provided that higher temperature difference occurs between hot side and cold side of TEM. Thus, the proposed model is of theoretical significance in guiding future design of TEMs for high-power or large-temperature-difference thermoelectric conversion systems.