Scholarlink

<p>Processing conditions have significant effects on mechanical oil expression from oleaginous seeds and kernels. It affects both the quantity and quality of oil, and the failure to optimise the process can result in poor yields and reduced quality, thereby limiting the potential for using oil in industries. The aim of the study was to model and optimise the impact of moisture content (MC), roasting time (RT), and heating temperature (HT) on the oil yield, extraction efficiency, and quality attributes of mechanically expressed African oil bean oil, based on soft computing technology. During oil expression using a mechanical oil expeller, MC (8–16% dry basis), HT (50–130&#xa0;°C), and RT (5–25&#xa0;min) were varied. The oil yield (Y), extraction efficiency (EE), oil moisture content (OM), acid value (AV), peroxide value (PV), pH, and iron content (IC) were analysed using normal analytical methods and compared to the FAO/WHO standards. Response Surface Methodology (RSM) and Artificial Neural Network (ANN) models were developed to reveal the relationships between processing variables and response parameters, and their predictive performances were evaluated based on the coefficient of determination (R<sup>2</sup>) and mean square error at a significance level of α = 0.05. The ANN was trained using a multilayer feed-forward network, and the Levenberg–Marquardt algorithm, along with a 50:25:25 data split, was employed to train, test, and validate the ANN. The optimal processing condition achieved was 8% MC (db), 11.7&#xa0;min RT, and 59.3&#xa0;°C HT, which gave 37% oil yield and 67.8% of extraction efficiency, with OM = 0.016, AV = 2.3&#xa0;mgKOH/g, PV = 7&#xa0;ml/g, pH = 5.8 and IC = 0.044&#xa0;mg/kg. All quality parameters were within the acceptable limits for edible oils. ANN models outperformed RSM in predictive accuracy (R<sup>2</sup> = 0.62–0.98), indicating the strength of ANN in optimising the mechanical processes of oil expression.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrated soft computing approach in modelling and optimising mechanical oil expression from pentaclethra macrophylla benth kernels

  • 期刊论文

  • 发表日期：

    2026年4月29日

查看全文
  • Clement Adesoji Ogunlade,
  • Ademola Kabir Aremu,
  • Rahman Akinoso

摘要

Processing conditions have significant effects on mechanical oil expression from oleaginous seeds and kernels. It affects both the quantity and quality of oil, and the failure to optimise the process can result in poor yields and reduced quality, thereby limiting the potential for using oil in industries. The aim of the study was to model and optimise the impact of moisture content (MC), roasting time (RT), and heating temperature (HT) on the oil yield, extraction efficiency, and quality attributes of mechanically expressed African oil bean oil, based on soft computing technology. During oil expression using a mechanical oil expeller, MC (8–16% dry basis), HT (50–130 °C), and RT (5–25 min) were varied. The oil yield (Y), extraction efficiency (EE), oil moisture content (OM), acid value (AV), peroxide value (PV), pH, and iron content (IC) were analysed using normal analytical methods and compared to the FAO/WHO standards. Response Surface Methodology (RSM) and Artificial Neural Network (ANN) models were developed to reveal the relationships between processing variables and response parameters, and their predictive performances were evaluated based on the coefficient of determination (R2) and mean square error at a significance level of α = 0.05. The ANN was trained using a multilayer feed-forward network, and the Levenberg–Marquardt algorithm, along with a 50:25:25 data split, was employed to train, test, and validate the ANN. The optimal processing condition achieved was 8% MC (db), 11.7 min RT, and 59.3 °C HT, which gave 37% oil yield and 67.8% of extraction efficiency, with OM = 0.016, AV = 2.3 mgKOH/g, PV = 7 ml/g, pH = 5.8 and IC = 0.044 mg/kg. All quality parameters were within the acceptable limits for edible oils. ANN models outperformed RSM in predictive accuracy (R2 = 0.62–0.98), indicating the strength of ANN in optimising the mechanical processes of oil expression.