|
Itiri, H. U.
Department of Chemical Engineering, Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria.
Ugwuodo, C. B.
Department of Chemical Engineering, Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria.
Kalu, C. U.
Department of Chemical Engineering, Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria.
Emmanuel, I. N.
Nazarbayev University, Astana, Kazakstan
ABSTRACT
This research utilized neuro-fuzzy modeling to
predict bio-oil yield from the pyrolysis of luffa cylindrica fiber. The luffa
cylindrica fiber was washed, oven-dried, ground using a mill, and screened to
obtain various particle sizes before final volatilization. The luffa cylindrica
fiber was characterized through proximate and ultimate analysis, scanning
electron microscopy (SEM), and thermogravimetric analysis (TGA). This
characterization indicates its potential for bio-oil production. Sensitivity
analysis employing the Adaptive Neuro-Fuzzy Inference System (ANFIS) exhaustive
search indicated that temperature, particle size diameter, and inert gas flow
rate significantly influenced the bio-oil yield from luffa cylindrica fiber
pyrolysis. Parametric analysis using ANFIS surface plots demonstrated that
higher temperature and inert gas flow rate enhance bio-oil yield. The
characterization results from FT-IR (Fourier Transform Infrared Spectroscopy)
and GC-MS (Gas Chromatography-Mass Spectrometry) confirm that the bio-oil meets
ASTM specifications. The ANFIS optimal results indicated a bio-oil yield of 12%
at a temperature of 648 °C, a particle diameter of 5 mm, and an inert gas flow
rate of 1 L/min. The Gbell membership function exhibited the highest
coefficient of determination (R² = 0.9999) and the lowest Average Testing Error
(ATE = 0.00045162) compared to other membership functions, highlighting the
model's significant predictive capability. The study's observations demonstrate
that the ANFIS technique is an effective method for predicting the pyrolysis
process of luffa cylindrica fibre.
Keywords: Adaptive neuro-fuzzy inference system, characterization, coefficient of determination, L luffa cylindrica, pyrolysis
https://doi.org/10.33922/j.ujet_v11i1_6
|
View: 39 | Download: 2
Published
Monday, February 03, 2025
Issue
Vol. 11 No. 1, June 2025
Article Section
GENERAL
The contents of the articles are the sole opinion of the author(s) and not of UJET.
|
