Neural network for short-term predictions of ambient particulate matter around thermal power plant

Abstract

A three-layer neural network model has been developed to predict suspended particulate matter (SPM) in and around a thermal power plant at Neyveli. The feed forward supervised neural network-the back propagation was used to train the network. Utilizing the meteorological data along with SPM data for all the six monitoring stations surrounding the thermal power plant, the neural gave better predictions. The aim of the present research work is to construct a forecasting model that would be suitable for the use of future planning and developments.