Application of Gaussian Smoothing Technique in Evaluation of Biomass Pyrolysis Kinetics in Macro-TGA. (October 2017)

• Record Type: Journal Article
• Authors: Onsree, Thossaporn; Tippayawong, Nakorn
• Is Part Of: Energy procedia. Volume 138(2017)
• Page Start: 778
• Page End: 783
• Publication Date: 2017-10
• Subjects: Biomass -- Macro-TGA -- Mathematical Smoothing -- Thermochemical Conversion -- Renewable Energy; Power resources -- Congresses; Power resources -- Periodicals; Power resources; Conference proceedings; Periodicals; 333.7905
• DOI: 10.1016/j.egypro.2017.10.059
• Languages: English
• ISSNs: 1876-6102

Abstract

Abstract: Pyrolysis of eucalyptus wood and refuse derived fuel was carried out in a continuously weighing fixed bed reactor performing as a macro thermogravimetric analyzer (macro-TGA). All samples were heated from ambient temperature to 800 ºC at different heating rates; 5, 10, and 15 ºC/min. Because of buoyancy effect and measurement uncertainties, noises were generally detected in mass change or thermogravimetric (TG) data obtained from the macro-TGA. The noises led to high degree of fluctuation in the derivative TG (DTG) data. These fluctuating data could not be used directly to analyze kinetic parameters. Smoothing was required in eliminating the noise to obtain reliable DTG data used for subsequent kinetic analysis. In this paper, Gaussian smoothing technique was applied to characterize the smoothing quality, based on mathematics and kinetics points of view. In mathematical analysis, values of smoothing parameters σ within 4-15 showed that raw-data can be improved to be sufficiently smooth to achieve DTG curves without high fluctuation by fourth-order central difference method. Smoothed TG curves did not significantly differ from the original signals. However, kinetic analysis results indicated that only Gaussian smoothing with 7 ≤ σ ≤ 10 can provide meaningful kinetic parameters ( Ea ~ 140-150 kJ/mol), which can be used to describe pyrolysis behavior in macro-TGA satisfactorily (R 2 ≥ 0.99).