# Green Fuels Production: Parameter Estimation of Fatty Acid Hydrodeoxygenation from Batch Data

## TL;DR

This paper estimates reaction parameters for converting fatty acids into green fuels using two models, finding that a simplified approach performs better.

## Contribution

The study introduces a simplified kinetic model (RS2) with linear free energy relationships that outperforms a detailed model in predicting green fuel production reactions.

## Key findings

- RS2 achieved higher accuracy (R² of 0.985) compared to RS1 (R² of 0.981) in fitting experimental data.
- Grouping reactions using linear free energy relationships provided reliable results with fewer variables.
- Kinetic-based models improved optimization of green fuel production compared to conversion-based models.

## Abstract

Replacing fossil fuels with renewable alternatives has
become increasingly
critical. The hydrodeoxygenation of fatty acids presents a viable
route for green fuel production, and understanding the kinetics of
these reactions is crucial for evaluating their economic viability.
The present paper focuses on estimating the kinetic parameters for
the hydrodeoxygenation of stearic and palmitic acids to produce n-alkanes, considering two reaction schemes: RS1, a detailed
theoretical pathway model, and RS2, a simplified kinetic representation
and categorization of similar reactions into families with linear
free energy relationships (LFERs). Aspen Plus v14 was used as the
process simulation environment to fit the batch data, employing the
generalized Langmuir–Hinshelwood–Hougen–Watson
kinetic model for RS1, while RS2 was described using a power-law kinetic
expression, both using Soave–Redlich–Kwong as the thermodynamic
model. To minimize the error between simulated and experimental data,
the sum of squared errors was used as the objective function with
an algorithm that integrates particle swarm optimization and golden
search. RS2 demonstrated superior performance in global error minimization
(R
2 of 0.985) and parameter reliability
compared to RS1 (R
2 of 0.981). Grouping
reactions into families using the LFER method successfully achieved
reliable results. This approach shows promise for estimating kinetic
parameters of systems with complex feedstocks while reducing the number
of variables. Using kinetic-based models instead of conversion-based
models enhanced the optimization of green fuel production units.