The Application of Statistical Methods in the Construction of a Model for Identifying the Combustion of Waste in Heating Boilers Based on the Elemental Composition of Ashes

Emission of air pollutants constitutes one of the major hazards to human health and life. Particulate matter and harmful gases emitted by residential heating, especially, occupy a significant position among the sources of air pollution. This paper presents the research results concerning the composition of ashes obtained from the combustion of fuel samples composed of wood pellets, eco-pea coal, and coal pellets (trade name-VARMO) with various admixtures of waste materials. The study stand was equipped with a boiler having a nominal power of 18 kW. Several of the most characteristic chemical elements identified in the ash were used as the basis to classify the combustion of waste. A model based on a statistical method was designed. Within the framework of the research, a statistical multivariate technique, discriminant analysis, was applied. The statistical model was constructed for two groups of ash samples and 19 chemical elements indicating their contamination. The high prediction power of the model and the validation (fitting was 90.00% and 85.19%, respectively) confirmed the possibility of the practical application of this proprietary method. It permitted identification of the markers (chemical elements) in the ash. It confirms that the fuel is combusted with the admixture of waste materials in a given boiler. Based on the analyses performed, it was found that from among the 19 elements, five, namely K, Ti, Zn, Ca, and Rb, were selected as the markers because they are characterised by the highest discrimination ability. In addition, they are the best indicators of the contamination level of the ash samples that were examined.

Automated summary

This paper presents the research results on the composition of ashes obtained from the combustion of fuel samples with various admixtures of waste materials. A statistical model based on discriminant analysis was constructed to identify the markers in the ash samples and indicate their contamination levels. The study found that five chemical elements, namely K, Ti, Zn, Ca, and Rb, have the highest discrimination ability and are the best indicators of contamination level.