Effect of process parameters and synthesis method on the performance of thermoelectric ceramics: A novel simulation

Artificial intelligence (AI) approach has been successfully applied to a wide range of materials engineering problems so far. However, there is no work on incorporation of modeling algorithms based on AI to predict the performance of thermoelectric materials. This paper shows how to predict the dimensionless thermoelectric figure of merit of Sr1−xYxTiO3 ceramics as the performance symbol of these structures, using an adaptive neuro-fuzzy inference system (ANFIS). Moreover, this work investigates the effect of the process parameters including Y content, compact pressure, sintering temperature, sintering time and process temperature on the performance of Sr1−xYxTiO3 thermoelectric ceramics. Absolute fraction of variance and root mean square error of 0.99 and 0.008, in training and testing phases of the model were achieved showing the relatively high accuracy of the proposed ANFIS model. To investigate the effect of the synthesis method (synthesis type) of Sr1−xYxTiO3 ceramics on the performance of these materials, this parameter was also introduced to the model and the obtained results were shown in 3D graphs. These 3D interaction graphs between some of process variables generated by the proposed model indicated that the combustion synthesis is more effective than other ones to increase the efficiency of the Sr1−xYxTiO3 thermoelectric ceramics.