Predictive Modeling Coupled with Multiple Optimization Techniques for Assessing the Effect of Various Process Parameters on Oil and Pectin Extraction from Watermelon Rind
This study combines artificial neural network (ANN) with a novel metaheuristic technique, satin bowerbird optimizer (SBO) for predicting oil, and pectin yield from watermelon rind. The experimental design was based on two-level factors, drying temperature (°C) and heating time (hours) for oil yield...
Gespeichert in:
Veröffentlicht in: | Process integration and optimization for sustainability 2022-09, Vol.6 (3), p.765-779 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | This study combines artificial neural network (ANN) with a novel metaheuristic technique, satin bowerbird optimizer (SBO) for predicting oil, and pectin yield from watermelon rind. The experimental design was based on two-level factors, drying temperature (°C) and heating time (hours) for oil yield (%) and three-level factors, drying temperature (°C), extraction time (min), and pH for pectin yield (%) using response surface methodology (RSM). The RSM and SBO optimization resulted oil yield of 17.7–26.6% at conditions of drying temperature (90–100°C), and heating time (8.7–9 h), while pectin yield is between 21 and 37.6% at extraction temperature (80–100), extraction time (46.3–60 min), and pH (1–3). However, ANN predicted more accurately than the RSM model, with a lower percentage relative error. It was observed that pH and extraction time are pertinent process parameters for predicting pectin yield. Similarly, drying temperature is significant for oil extraction from watermelon.
Graphical abstract |
---|---|
ISSN: | 2509-4238 2509-4246 |
DOI: | 10.1007/s41660-022-00248-0 |