A support vector based CO2 gas emission prediction system for generation power plant

• Main Authors: Chen C.P., Tiong S.K., Albert F.Y.C., Koh S.P.
• Other Authors: 25824552100
• Format: Article
• Published: American Scientific Publishers 2023
• Institution: Universiti Tenaga Nasional

The work presents an intelligent system Support Vector Regression Emission Monitoring System (SuVEMS) developed for Tenaga Nasional Berhad (TNB) Sdn. Bhd. in Peninsular Malaysia for the prediction of harmful gas emissions from electricity generating power plants in Tuanku Jaafar Power Station (TJPS). The CO2, emissions is modelled on this work using Support Vector Regression (SVR), a statistical machine learning tool with a regression-based extension towards Support Vector Machines (SVMs). The gas is predicted using independent models and the gas prediction model is trained using feature subsets selected using the forward selection approach. The SuVEMS results are compared and measured the performance with the Continuous Emission Monitoring System, CEMS results. The SuVEMS results implemented at TJPS indicate that it has the ability for the online prediction with average prediction accuracy of 95%. � 2017 American Scientific Publishers All rights reserved.