Scholarlink

<p>The production of hexamethylene-1,6-diisocyanate (HDI) via carbamate cracking avoids the use of phosgene; however, the environmental impact of this non-phosgene synthesis route has not been thoroughly studied. Conventional process simulations typically focus solely on economic indicators, which may overlook important environmental considerations. We have developed a multi-objective optimization framework that collectively accounts for total annual cost, green degree and total energy consumption for such a non-phosgene synthesis process of HDI. We used data from rigorous simulations to construct a surrogate model based on artificial neural networks. This model simplifies the numerical correlation among optimization objectives and process/operational parameters. The bat algorithm was employed for multi-objective optimization on such surrogate models with significantly reduced computational burden, resulting in a series of optimal solutions as the Pareto front. The optimal solution exhibits a 28.79% reduction in total energy consumption and a 19.72% reduction in the total annual cost compared to the real-world production data; the green degree of the optimized process is increased by 21.60%.</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

ANN assisted optimization of phosgene free hexamethylene diisocyanate production for economic and environmental benefits

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月30日

查看全文
  • Bin Liu,
  • Guoqing Zhang,
  • Suyun Hong,
  • Zhuxiu Zhang

摘要

The production of hexamethylene-1,6-diisocyanate (HDI) via carbamate cracking avoids the use of phosgene; however, the environmental impact of this non-phosgene synthesis route has not been thoroughly studied. Conventional process simulations typically focus solely on economic indicators, which may overlook important environmental considerations. We have developed a multi-objective optimization framework that collectively accounts for total annual cost, green degree and total energy consumption for such a non-phosgene synthesis process of HDI. We used data from rigorous simulations to construct a surrogate model based on artificial neural networks. This model simplifies the numerical correlation among optimization objectives and process/operational parameters. The bat algorithm was employed for multi-objective optimization on such surrogate models with significantly reduced computational burden, resulting in a series of optimal solutions as the Pareto front. The optimal solution exhibits a 28.79% reduction in total energy consumption and a 19.72% reduction in the total annual cost compared to the real-world production data; the green degree of the optimized process is increased by 21.60%.