MODELLING OF CO-TORREFACTION PROCESS OF LOW RANK COAL AND BIOMASS

• Main Author: ALFARISI (NIM: 23016040), SALMAN
• Format: Theses
• Language: Indonesia
• Subjects: Teknik kimia
• Online Access: https://digilib.itb.ac.id/gdl/view/30778
• Institution: Institut Teknologi Bandung

<p align="justify">Indonesia is a country which has quite large amount of coal resources with production in 2016 is 128 billion tons. Based on heating value produced, coal available in Indonesia mostly at medium (21%) and low rank (45%). This caused the coal cannot be used for purposes of the processes that require high heat. On the other hand, there is another energy that potentially considerable, namely biomass. Biomass in Indonesia is very abundant. In 2016 utilization of biomass as energy resource reach 20,06% from total energy usage in Indonesia. To be used as a good fuel biomass has to be processed first. One of the process that is commonly done is torrefaction. The purposes of torrefaction are removing water content, increasing heating value, removing volatile compounds, increasing grindability and hydrophobicity. Torrefaction of low rank coal and biomass blends be able to produce hybrid coal which is has higher heating value than low rank coal. The presence of biomass in the mixture makes non-neutral CO2 emissions from combustion of hybrid coal become less. <br /> <br /> This research was conducted to obtain a mathematical model that corresponds to the process of co-torrefaction low rank coal and biomass blends that are useful for predicting final mass of solid produced and temperature distribution inside single particle in the process that produces optimum hybrid coal. In this research, model was built based on Thermo Gravimetric Analysis (TGA) data. Kinetics parameters which is obtained used for modelling process of co-torrefaction macro particle. On macro particle, temperature will be distributed at any position inside particle and then final mass of hybrid coal produced can be predicted. The result difference between models and experimental data yield is between 2,5% to 7%. While the temperature distribution inside single particle between model and data can fit well. With the result that, this model can correlate quite well with the experiment.<p align="justify">