Emission of organic aerosols from wood combustion: Bridging the gap between aerosol science and wood pyrolysis

Fawaz, Mariam M

Permalink

https://hdl.handle.net/2142/115325 Copy

Description

Title

Emission of organic aerosols from wood combustion: Bridging the gap between aerosol science and wood pyrolysis

Author(s)

Fawaz, Mariam M

Issue Date

2022-01-18

Director of Research (if dissertation) or Advisor (if thesis)

Bond, Tami C

Doctoral Committee Chair(s)

Bond, Tami C

Committee Member(s)

• Riemer, Nicole
• Brewster, Quinn
• Verma, Vishal

Department of Study

Civil & Environmental Eng

Discipline

Environ Engr in Civil Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Wood Pyrolysis
• Particle Emission

Abstract

Organic aerosols (OA), from biomass burning, are a large contributor to primary organic aerosols in the atmosphere. One of the most significant sources of biomass burning is wood combustion due to wildland fires and solid fuel combustion for residential use. The emission of OA from biomass burning is not well constrained, making the determination of emissions difficult. However, high certainty in OA emission from biomass burning is essential because it is used in emission inventories, air quality models, and health impact studies. Wood combustion includes gas phase processes above the wood and solid phase processes inside the wood. The thermochemical degradation of wood, called wood pyrolysis, occurs in the solid phase during combustion and produces OA precursors that may partition to the aerosol phase. This dissertation studies the fundamental processes responsible for OA emission during wood combustion and evaluates the relationship between emission and transport phenomena to establish an improved representation of the emission system. It spans three objectives: (1) using pyrolysis modeling to understand the underlying processes to OA emission; (2) connecting pyrolysis to OA emission; and (3) predicting OA emission from pyrolysis. This work follows an integrated modeling and experimentation approach to study pyrolysis processes in the absence of gas phase processes. First, Gpyro, an open-source transport phenomena modeling software, was used to predict the change in temperature and mass in the wood and the emission of classes of gaseous products throughout the process. Second, experiments were performed to verify the pyrolysis model from Gpyro and to measure emissions at different conditions. A reactor was built to control pyrolysis conditions throughout the experiment and to collect data in real time. The reactor was made up of a cylindrical heater connected to a temperature controller. Mass and temperature changes of the wood samples were collected to compare and verify the model output. Real-time emission measurements of particle and gas concentrations were performed at conditions that changed the temperature, wood size, wood moisture content, and wood type. Third, using both the experimental and model results, a machine learning model was trained to predict the particle emission rate under different input conditions. The main results, from this dissertation, show that heat transfer controls the pyrolysis process and product formation. In particular, the external heat transfer influenced the initial phase of pyrolysis, and the remaining period was influenced by the heat transfer resistance between the surface and center of the wood. Further, the emission results revealed that particle emission is a deterministic process that depends on the experimental conditions; pyrolysis is a repeatable process with a maximum variation of 20% and that the OA emission rate followed the mass loss rate of the wood. The change of OA emission in response to the change in the experimental conditions was governed by the heat flux the wood receives and the thermophysical properties of the wood. The trained machine learning regression model was evaluated using experimental emission data with a 0.9 goodness of fit (R2); the regression model was then used to predict emission using the Gpyro model output instead of the experimental data. The machine learning method predicted particle emission rate with a 35% error using inputs from the model. In conclusion, the work in this dissertation establishes the link between modeling the transport phenomena and emission measurements from wood pyrolysis. This dissertation is the first step towards predicting OA emission and the changes the particles undergo during combustion.

Graduation Semester

2022-05

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Mariam Fawaz

Owning Collections