Abstract-This paper focus on a new methodology approach to evaluate more accurately the energy generated from Thermoelectric Generator (TEG) under the influence of its operating environmental parameters. An artificial neural network (ANN) model for predicting the energy generated by a TEG in its operating environment has been developed. The dataset generated through a validated finite volume method is trained in a supervised way and tested by a multi-layer perceptron (MLP) to predict the energy generated. However, the degree of ambiguity may vary widely across the whole range of input values therefore in this paper, a new methodological approach will be incorporated to not only predict the average value but as well as evaluating the reliability of the output value with the use of a scheme which is made up of two coupled neural network. Apart from predicting the output values, this model can perform reverse ANN to predict the input value when provided with an output value.

No abstract

No abstract