IOsE for Real Time Monitoring of Combustion Flames (RMCF) in Industrial Boilers

The innovation in the area of Internet of Smart Environment (IOsE) for the quality prediction of combustion process in industrial boilers is highlighted in this article. IOsE is quality reliant. The characteristics of flame in the combustion chamber which is used for identification and control of Carbon Dioxide (CO2) emissions. Currently, existing methods for flame analysis include non-invasive and high frequency spectral light dependent imaging techniques. Colour of the flame inside the Combustion Chamber (CCh) determines the prediction of Combustion Quality (CQ) and thereby its corresponding Combustion Quality Index (CQI). An online quality detection scheme for smart environment with colour extraction technique from the flame images to identify complete and incomplete combustion conditions is developed. For this, the intensity value of the pixels in Red-Blue-Green (RBG) plane is utilized. The coal is used as the prime fuel to heat water and convert it to steam for operating the industrial boilers. The quality prediction of combustion from flame colour and shape includes Complete Combustion (CC) and Incomplete Combustion (IC) categories. The flame images for CC and IC categories are subjected to preprocessing, attribute extraction and finally the prediction of flame quality is done using Random Forest (RF), Ada Boost (AB), Gradient Boosting (GB), Decision Tree (DT) and Multi Layer Perceptron (MLP). MLP is found to effectively categorize the CQ using the performance metrics like Sensitivity (SEY), Specificity (SPY) and Accuracy (ACY).