Forecasting Air Quality Parameters of Kozhikode Using Artificial Neural Network

Submission Deadline-29th June May 2024
June 2024 Issue : Publication Fee: 30$ USD Submit Now
Submission Deadline: 20th June 2024
Special Issue of Education: Publication Fee: 30$ USD Submit Now

International Journal of Research and Scientific Innovation (IJRSI) | Volume V, Issue V, May 2018 | ISSN 2321–2705

Forecasting Air Quality Parameters of Kozhikode Using Artificial Neural Network

Harshan K.G#, Anjali S*

IJRISS Call for paper

#Associate Professor, Dept. of Civil Engineering, *Student M.Tech (Environmental Engineering), M DIT Ulliyeri, India

Abstract-Each day marks a proliferation in the development of a place, which definitely suggest a rich country, even with an acute mess of increasing air pollution, which is a serious threat for both environment and human beings. Unquestionably, this case need to have a breakthrough which itself is the aim of this project. Tracking two locations of Kozhikode district, the project destines at finding the best model for the prediction of air quality parameters in both the invigilated stations. Artificial Neural Network (ANN) is used along with Nonlinear Autoregressive Exogenous Input (NARX) network for predicting the best model having a minimum Mean Square Error (MSE) and Coefficient of Regression (R2) having a value nearer to 1. NARX network consist of two series of input and output representing the meteorological parameters and past values of pollutant concentrations respectively, where output is feedback to input for preparing the model. The model is prepared using pollutant like Sulphur dioxide(SO2), Nitrogen dioxide(NO2), Respirable Suspended Particulate Matter (RSPM) and Suspended Particulate Matter(SPM) using Levengerg Marquardt(L-M) algorithm. The best model for two stations Kozhikode city and Nallalam have 8 hidden layers and 4 hidden layers respectively, which is used for predicting the air pollutant concentration of future years by providing the input of the same using forecast sheet in Excel. The resulting values of both stations seems to have the concentration of SO2 and NO2 with in the limit of National Ambient Air Quality Standards (NAAQS) where as both SPM and RSPM have values greater than the NAAQS, which thus proves that there is a need for taking necessary safe measures.

Keywords-Air pollution, Artificial Neural Network, NARX, L-M algorithm, SO2, NO2, RSPM, SPM, MSE, R2