Performance comparison and analysis of mathematical, ANSYS and neural network model of a thermo electrical generator

Abstract

- The need to generate energy from waste energy is growing as the demand for energy rises. Heat engines can only exchange around 35% of accessible energy, leaving the excess as surplus heat. A thermoelectric generator can transform this waste heat into useful energy. The Seebeck principle is used to create a thermoelectric device, which consists of a P-Type and N-Type semiconductor coupled electrically in series and thermally in parallel. Thermoelectric generators immediately convert thermal energy to electrical energy without the use of rotating parts, and because there are no moving parts, they are possibly less prone to failure. With the rising cost of fossil fuels and their negative influence on the atmosphere, it is past time to consider renewable energy sources (RES). Thermoelectric generators (TEG) are widely utilized as autonomous RES, with applications ranging from mW to W power. In this paper mainly focusing the comparison of three different model of Thermo Electric Generator. Out of these three mathematical model give a better result compared to neural and ANSYS model. Neural network model was developed from the hardware result obtained from the experiment. In practical case the temperature difference is very less so the output voltage current and power are very less. The maximum hot side temperature of SP1848 27145 SA TEG is 50 °C and minimum temperature is 27 °C. The seebeck coefficient, voltage, current, Figure of Merit performances are also plotted.

Keywords

• Thermo electric generator
• Figure of merit
• Seebeck effect
• Neural network
• ANSYS