GOALI: Long-term Planning of Plant-wide Recovery and Treatment Options

研究机构：伊利诺伊大学芝加哥分校提案号：0090370多用途批量生产场所的废物减少和污染预防规划涉及到原材料、质量分离剂和能源等资源的合理利用，以支持化学品的生产。这是一个动态的开放式问题，因为战略应该预测未来商业运作的新趋势，超越当前制造实践的新技术的可用性或环境问题。不确定性源于连续的生产活动序列，每个活动持续数周至数月。整个工厂的废物管理通常建立在现有的现场基础设施上，如溶剂回收厂、焚化炉、化学处理技术和生物处理。存在不确定性的改造情况与大宗商品业务中使用的专用工厂中占主导地位的确定性基层设计方法形成鲜明对比。本研究包括在固定的规划范围内为全厂废物管理开发最佳设计方法。由于问题的多面性，计划进行协同学术-产业合作（GOALI项目）。pi计划开发一种他们称之为“组合过程合成”的新技术，该技术使用两相解决方法。在第一阶段，上层建筑生成，基于知识的推理机制搜索技术上可行的恢复和治疗方案的上层建筑。采用非线性平衡方程和本构方程来保证状态任务网络的一致性。在第二步中，上层结构优化确定工厂范围内的废物管理策略，该策略受工厂特定物流、容量和排放限制的约束。最优战略还建议投资决策，如新的回收技术或处理方案，以同时提高经济和生态绩效或解决更严格的环境标准。组合过程综合方法的主要优点在于它能够通过信息搜索和数学规划的结合来提供开放式问题的算法解决方案。

ABSTRACTPI: Andreas Linninger and Richard Colberg Institution: University of Illinois at chicagoProposal Number: 0090370Planning of waste reduction and pollution prevention efforts for multi-purpose batch manufacturing sites is concerned with the mindful use of resources such as raw materials, mass separating agents and energy to support the production of chemicals. It is a dynamic open-ended problem, since strategies should anticipate emerging trends in future business operations, availability of new technologies beyond current manufacturing practices or environmental concerns. Uncertainty stems from a continuous sequence of production campaigns, each lasting from weeks to months. Plant-wide waste management typically builds on existing site infrastructure such as solvent recovery plants, incinerators, chemical treatment technologies, and biological treatment. The retrofit situation in the presence of uncertainty contrasts the deterministic grass-roots design approach predominant in dedicated plants used in the bulk commodity business. This research consists of developing optimal design methodology for plant-wide waste management for a fixed planning horizon. Due to the multi-faceted nature of the problem, a synergistic academic-industrial collaboration (GOALI project) is planned.The PIs plan to develop a new technique they call "combinatorial process synthesis," which uses a two-phase solution approach. In phase one, superstructure generation, a knowledge-based reasoning mechanism searches for a superstructure of technically feasible recovery and treatment options. Non-linear balance and constitutive equations are enforced to safeguard the consistency of the evolving state task network. In step two, superstructure optimization identifies plant-wide waste management strategies subject to plant-specific logistic, capacity and emission constraints. Optimal strategies also suggest investment decisions such as new recycle technologies or treatment options to simultaneously increase economic and ecological performance or to address more stringent environmental standards. The major advantage of the combinatorial process synthesis method lies in its ability to render an algorithmic solution to open-ended problems via a combination of informed search and mathematical programming.