Stochastic type-2 fuzzy modelling on GHG emission prediction for gas-fired power plant

Abstract

This paper demonstrates the implementation of type-2 fuzzy logic model in prediction of greenhouse gas emission for gas fuelled power plant. The data collection and analysis on the GHG emission had been conducted with emphasis on the industrial scale with many uncertainty factors. The developed model will be used as a prediction tools to evaluate potential GHG emission for future study and mitigation planning. In this work, 100 sample sizes had been used as training dataset in the model framework which serve as the input information while 14 sample sizes are used for model validation to achieve model generalization. A novel Karnik-Mendel (KM) algorithm had been proposed with the Genetic Algorithm (GA) model optimization to achieve high performance and accuracy for the respective predictive model. In conclusion, the result showed that the model able to give an outstanding accuracy. On the validation analysis, the results showed developed model able to give satisfactory prediction with R2 of 0.978 and 0.84968 for training and validation, respectively.