Publication: Gas Emission Prediction for Environmental Sustainability via Heterogeneous Data Sources Correlation with Support Vector Regression
No Thumbnail Available
Date
2018
Authors
Chen C.P.
Tiong S.K.
Koh S.P.
Chooi Albert F.Y.
Mohd Yapandi M.F.K.
Journal Title
Journal ISSN
Volume Title
Publisher
Institute of Electrical and Electronics Engineers Inc.
Abstract
With the emerging of industrial revolution 4.0, artificial intelligence (AI) together with big data analytics will be playing an important role in environmental sustainability by improving system efficiency and intelligent environment monitoring. The increasing of electricity demand and urbanization process have caused more power plants to be built from time to time, which may cause environmental issue for its surrounding. Hence, necessary measures need to be taken to ensure environmental sustainability. This paper is to investigate the ability of a regression based artificial intelligent algorithm, namely Support Vector Regression (SVR), correlating with multiple sources of big data sets to predict the Sulfur Dioxide (SO2) emission level at atmosphere surrounding a Combined Cycle Gas Turbine (CCGT) power plant. The heterogeneous data sources that have been used to train and establish the knowledge of SVR are meteorological data, terrain and land use data, historical emission data and power plant parameters particularly related to the point source emitter. With the correlation of multiple big data sources, SVR was then trained for the prediction of emission rate at the chimney and certain targeted areas such as residential area surrounding the power plant, which are classified as air sensitive receptors (ASR). Although there are a number of gasses emitted from power plant, SO2is selected as the key emission in this paper due to inhaling of sulfur dioxide will cause respiratory symptoms and diseases for living things. The developed predictive model is incorporated into an online monitoring tool namely Integrated Support Vector Regression Emission Monitoring System (i-SuVEMS). The predicted SO2gas emission result by i-Su VEMS was compared with the actual emissions results from the CEMS. The predicted values from i-SuVEMS shows good accuracy with RMSE less than 0.02 as compared to the actual measured emission values. This prediction performance result indicates that i-Su VEMS is able to meet the requirement of US EPA 40 CFR Part 60 in predicting the quantity of SO2gas emission into the atmosphere and consequently can be used as a tool for environmental sustainability monitoring. � 2018 University of Split, FESB.
Description
Air pollution control equipment; Artificial intelligence; Combined cycle power plants; Environmental Protection Agency; Forecasting; Gas turbines; Land use; Meteorology; Monitoring; Regression analysis; Sulfur dioxide; Sustainable development; Vectors; Combined cycle gas turbine; Emission monitoring system; Environmental sustainability; Heterogeneous data sources; Intelligent environment; Power plant parameters; Prediction performance; Support vector regression (SVR); Big data