Pectin extraction from Helianthus annuus (sunflower) heads using RSM and ANN modelling by a genetic algorithm approach

Abstract

In this work, Response Surface Methodology (RSM) and Artificial Neural Network coupled with genetic algorithm (ANN-GA) have been used to develop a model and optimise the conditions for the extraction of pectin from sunflower heads. Input parameters were extraction time (10–20 min), temperature (40–60 °C), frequency (30–60 Hz), solid/liquid ratio (S/L) (1:20–1:40 g/mL) while pectin yield (PY%) was the output. Results showed that ANN-GA had a higher prediction efficiency than RSM. Using ANN as the fitness function, a maximum pectin yield of 29.1 ± 0.07% was searched by genetic algorithm at the time of 10 min, temperature of 59.9 °C, frequency of 30 Hz, and solid liquid ratio of 1:29.9 g/mL while the experimental value was found to be 29.5 ± 0.7%. Extracted pectin was characterised by FTIR and 13C NMR. Thus, ANN coupled GA has proved to be the effective method for the optimization of process parameters for pectin extraction from sunflower heads.