Scholarlink

<p>This study was aimed to optimize the efficiency of silica extraction from a silica-rich, natural solid waste perlite using dissolution–precipitation method. Four different mineral acids, namely, H<sub>2</sub>SO<sub>4</sub>, HCl, HNO<sub>3</sub> and HClO<sub>4</sub> were compared in precipitation step. The optimization process was accomplished using a proposed hybrid non dominated sorting genetic algorithm-II (NSGA-II) combined with back-propagation artificial neural network (BPANN) and response surface methodology (RSM). Various experimental parameters such as stirring temperature (40–120 ℃), time (0.5–6&#xa0;h), NaOH concentration (2–8.5&#xa0;M), and pH (2–7) were optimized using central composite design (CCD). In the proposed hybrid RSM-BPANN-NSGA-II model, projected data of BPANN was used as initial score and multiple regression equations of RSM were applied to develop two fitness functions maximum % yield and minimum time. An array of best-fit solutions was attained as Pareto front, and the final optimal design point was picked using technique for order preference by similarity to ideal solution (TOPSIS analysis). The optimized results were compared against Mean Absolute Error (MAE), Mean Square Error (MSE), and Root Mean Square Error (RMSE). In the optimization comparative study, proposed hybrid RSM-BPANN-NSGA-II model outperformed the other two models. Structure, morphology, and chemical bonding of silica extracted from perlite was studied utilizing different characterization techniques. XRD, FT-IR and UV–Visible DRS analysis depict the existence of amorphous Si–O-Si network while SEM–EDX results align with experimental data confirming that maximum % yield of silica (90.02) extraction was achieved by HCl precipitation. This study showed the relative superiority of optimized process parameters with higher silica % yield and reduced time.</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

Modelling and Optimization of Silica Extraction from Perlite Using RSM, ANN, and NSGA-II Techniques

  • 期刊论文

  • 发表日期：

    2025年11月26日

查看全文
  • Deepti Goyal,
  • Monika Sharma,
  • Dipti Singh,
  • Raju Pal,
  • Sakshi Kabra Malpani

摘要

This study was aimed to optimize the efficiency of silica extraction from a silica-rich, natural solid waste perlite using dissolution–precipitation method. Four different mineral acids, namely, H2SO4, HCl, HNO3 and HClO4 were compared in precipitation step. The optimization process was accomplished using a proposed hybrid non dominated sorting genetic algorithm-II (NSGA-II) combined with back-propagation artificial neural network (BPANN) and response surface methodology (RSM). Various experimental parameters such as stirring temperature (40–120 ℃), time (0.5–6 h), NaOH concentration (2–8.5 M), and pH (2–7) were optimized using central composite design (CCD). In the proposed hybrid RSM-BPANN-NSGA-II model, projected data of BPANN was used as initial score and multiple regression equations of RSM were applied to develop two fitness functions maximum % yield and minimum time. An array of best-fit solutions was attained as Pareto front, and the final optimal design point was picked using technique for order preference by similarity to ideal solution (TOPSIS analysis). The optimized results were compared against Mean Absolute Error (MAE), Mean Square Error (MSE), and Root Mean Square Error (RMSE). In the optimization comparative study, proposed hybrid RSM-BPANN-NSGA-II model outperformed the other two models. Structure, morphology, and chemical bonding of silica extracted from perlite was studied utilizing different characterization techniques. XRD, FT-IR and UV–Visible DRS analysis depict the existence of amorphous Si–O-Si network while SEM–EDX results align with experimental data confirming that maximum % yield of silica (90.02) extraction was achieved by HCl precipitation. This study showed the relative superiority of optimized process parameters with higher silica % yield and reduced time.