Prediction And Analysis Of Physical Andchemical Parameters Of Biomass Andpyrolysis Carbon

In recent years,China’s demand for energy has increased year by year with the development of economy,and the country has strongly supported the development of new energy.Biomass energy is widely used in many new energy sources with the advantages of low cost and sustainability.However,most of the testing processes of biomass physical and chemical properties are complex and involve expensive testing equipment,so it is necessary to find appropriate methods to quickly predict its performance parameters.In this paper,through mining and analyzing the relevant data,the relationship between biomass industry analysis and biomass high calorific value,element composition,biochar yield and biochar pH value is obtained,and mathematical regression and neural network methods are respectively used to establish the prediction model of the relationship between parameters,and the corresponding functional prediction software is developed.The main research contents of this paper are as follows:A total of more than 2000 relevant data samples were collected and sorted out for mathematical analysis.The analysis results show that the high calorific value and the content of main components of biomass increase with the increase of volatile content in biomass,and decrease with the increase of ash content.In addition,the yield of biochar decreases with the increase of pyrolysis temperature and pyrolysis time,and the pH value increases with the decrease of the yield of biochar.According to the analysis results of the data,the corresponding prediction models are constructed by using the least square method,genetic algorithm and BP neural network,and the error analysis of the established model is conducted by using the average absolute error and the average absolute percentage error.The prediction ability of the selected model is verified by using the adjustment decision coefficient.The results show that the nonlinear model can show good prediction performance in parameter prediction under normal circumstances,But in the prediction model of oxygen content,the first-order linear model has better prediction performance.In addition,the prediction performance of BP neural network in the prediction model of hydrogen content and biochar yield is better than that of traditional mathematical regression model.