Online source tracing of waste paper by smoke based on laser-induced breakdown spectroscopy

Abstract

Paper is a widely used material and common recyclable household waste in waste disposal, which gets more attention nowadays for the misclassification of recyclable waste. In this work, an online source tracing system combined with machine learning algorithms to identify and classify the smoke of waste paper incineration based on laser-induced breakdown spectroscopy (LIBS) was established. Four types of waste paper, including tissue, corrugated paper, printing paper, and newspaper, were taken as examples. The smoke of four different waste papers was detected by LIBS and then further analyzed. The detected spectra with C, N, O, Mg, Al, and Ca could hardly be distinguished artificially. The random forest algorithm and the linear discriminant analysis were introduced to classify the smoke, and its accuracy reached 95.83%. The results indicate that source tracing of waste paper can be realized by identifying and classifying the smoke via the developed system. This could provide some reference for helping us to monitor the effectiveness of waste classification and incineration and monitor the atmosphere pollution.