Investigation of steam gasification in thermogravimetric analysis by means of evolved gas analysis and machine learning

Özveren U., Kartal F., Sezer S., Ozdogan Z. S.

ENERGY, cilt.239, 2022 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 239
  • Basım Tarihi: 2022
  • Doi Numarası: 10.1016/j.energy.2021.122232
  • Dergi Adı: ENERGY
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Aerospace Database, Agricultural & Environmental Science Database, Applied Science & Technology Source, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Communication Abstracts, Compendex, Computer & Applied Sciences, Environment Index, Geobase, INSPEC, Metadex, Pollution Abstracts, Public Affairs Index, Veterinary Science Database, Civil Engineering Abstracts
  • Anahtar Kelimeler: Gasification, Thermogravimetric analysis, Mass spectrometry, Evolved gas analysis, Machine learning, ARTIFICIAL NEURAL-NETWORKS, PYROLYSIS BEHAVIOR, COAL, BIOMASS, COCOMBUSTION, TGA, COMBUSTION, SLUDGE, ASH
  • Marmara Üniversitesi Adresli: Evet

Özet

The syngas distribution from steam gasification depends on both the feedstock and the gasification conditions. Therefore, it is of utmost importance to increase the know-how about the overall picture of steam gasification. Thermogravimetric analysis (TGA) is a commonly used method that provides valuable information about the gasification process. The TGA designed for steam gasification and its auxiliary equipment are comparatively expensive, the experiments take a long time and need a qualified operator. Therefore, the development of an easily applicable computational method for thermogravimetric behavior during steam gasification is very important. Although there are some works on predicting the pyrolysis and combustion behavior using artificial neural network (ANN), a model that predicts gasifi-cation behavior by TGA has not been studied. In this study, the gasification behavior and gas product characteristics of solid fuels were investigated by TGA coupled with mass spectrometry. Moreover, we report the first comprehensive model to estimate the thermogravimetric behavior of steam gasification using ANN as a machine learning approach. The ANN model provides a reliable estimation with an R-2 value of greater than 0.999. Moreover, MAPE values are reported to average less than 1%, while 6.5% for pyrolysis and 33.6% for extrapolated validation conditions. (c) 2020 The Author(s). This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).