SENSITIVITY ANALYSIS VIA ARTIFICIAL NEURAL NETWORK OF BIOMASS BOILER EMISSION

According to a survey in 1999, only 76 % of the palm oil mills in Malaysia meet the regulation of Department of Environment (DOE) regarding emission. The emission is released through the chimney from the process of fuel combustion and steam generation in order to produce power to the mill. The complex and very highly non-linear process involves several variables in fuel, turbine and boiler as factors producing pollutants. Therefore, Sensitivity Analysis via Artificial Neural Networks (SAANN) and Correlation Coefficient (CC) were used to find the major and minor input variables to each pollutant from 15 input variables. The result shows the major and minor input variables for both methods are similar