我国是水资源严重短缺的国家之一，且环境排放标准日趋严格，钢铁企业因“高耗水、高污染”备受关注，节水减排已成为一项刻不容缓的工作。水网络优化法基于过程系统工程原理，通过水资源的优化配置，从系统层面充分发掘节水减排潜力，是实现水资源利用和水污染控制的有效方法。本文以典型长流程钢铁工业园为研究对象，开展全工业园区水网络优化研究，以期为我国钢铁工业园区的节水减排提供数据和方法支持。主要研究内容和结果如下：（1）提出典型长流程钢铁工业园涉水单元多出口模型。根据涉水单元实际用水特点（双出口、多出口），对钢铁工业园水网络中典型涉水系统（软环、净环、浊环、直流用水系统、水源、水阱、废水处理系统和脱盐系统）建立统一单元模型，为充分考虑涉水单元多出口水流的回用奠定了基础。（2）工序水网络超结构设计。以涉水单元模型为基础，综合工序生产用水、排水特点，以直接回用、循环和处理后回用、处理后循环为基本策略，建立生产工序水网络超结构；（3）工业园区水网络超结构设计。在工序水网络超结构的基础上，综合考虑预处理水系统、综合废水处理及脱盐水系统，并以综合废水处理系统-多出口脱盐水系统为集成中心，且考虑工序间直接-间接综合集成模式，构成全水网络层面的园区水网络超结构。（4）建立典型钢铁工业园水网络多尺度优化模型。在以上工作基础上，建立涵盖单元、工序和园区三个尺度的园区水网络优化模型框架。该模型通过涉水单元的一级循环、工序水网络的二级循环和园区水网络的三级循环系统间的协调配合，探索钢铁工业园节水减排的优化方案。工业实际案例研究表明，仅考虑工序间间接集成优化时，综合用水成本可减少21%，而综合考虑直接和间接混合集成时，综合用水成本可减少23%。进一步的水网络关键操作参数变化对优化方案的影响研究表明，盐平衡是钢铁工业园水网络运行的关键运行因素，新水含盐量、用水单元进水含盐量控制标准以及脱盐的处理效果均对水网络优化有重要影响。 （5）提出钢铁工业园废水零排放方案。针对浓盐水为钢铁企业实现废水零排放的主要障碍的问题，本文选取了两种浓盐水零排放工艺，并对其建立单元模型，构建面向废水零排放的园区水网络超结构以及相应的优化模型。经案例分析可知，考虑了浓盐水零排放系统的水网络超结构可实现废水零排放，且以两级反渗透为基础的浓盐水零排放工艺较以多效蒸发-电渗析为基础的零排放工艺更具成本优势。（6）对不确定条件下的钢铁工业园水网络进行优化分析。钢铁园区水网络中各尺度子系统关联紧密，相互作用关系复杂，为提高水网络健壮性，本文采用两阶段随机规划法对不确定条件下的钢铁工业园水网络进行优化设计。案例中研究新水水质波动以及焦化废水溢流等不确定因素下的水网络优化方案。通过优化分析可知，考虑了不确定因素的水网络中脱盐系统、综合废水处理系统以及浓盐水处理系统的最大允许出口流量增大，表示当不确定参数实现时，系统无法通过仅调节流量使得水网络正常运行，而需在水网络设计阶段增大水处理系统规模。;With the increasingly scarcity of water source and ever-tightening environmental regulations in our country, water conservation and emission reduction for steel manufacture brooks no delay, due to the water-consuming and high-pollution features of steel manufacture which has driven more and more attention in recent years. Water network optimization method, based on process system engineering theory, via optimal allocation of water resource and exploiting potentials of water conservation and emission reduction, is efficient for water resource utilization and water pollution control. In this work, typical long process steel manufacture is used as the subject for the study of water network optimization of industrial park, expecting for data and method support for water conservation and emission reduction in steel industrial park. Main research contents and results are listed below:(1) Multiple outlets unit model in typical long process steel industrial park. According to the practical features of units, unit models for the typical water systems in the water network of steel industrial park (indirect closed circulating cooling water system, indirect open circulating cooling water system, direct open circulating cooling water system, once though water process system, water source system, water demand system, wastewater treatment system, desalination system) are developed, which lays a foundation for fully consideration of reuse of multiple outlet streams.(2) Superstructure design of plant-scale water network. Based on the unit models, integrating the features of water use and discharge in plants, the plant-scale superstructure is established via water reuse, water recycle, regeneration reuse and regeneration recycle water management strategy.(3) Superstructure design of the water network of steel industrial park. Using the plant-scale water network superstructure as the base, integrating the water pretreatment system, centralized wastewater treatment system and desalination system, the park-scale water network superstructure is established, in which centralized wastewater treatment system and multiple outlets desalination system perform as utility hub, and both direct and indirect integration strategies are proposed.(4) Development of multi-scale water network optimization mathematical model for typical steel industrial park. Based on the efforts presented above, the mathematical model framework containing unit-, plant- and park-scale models is established. By the primary recycle of streams at unit-scale, secondary recycle at plant-scale, and tertiary recycle at park-scale, as well as the coordination between different scale water systems, the model is used to explore the optimization schemes of water conservation and emission reduction in steel industrial park. An industrial example is carried out, and the results show that, when only considering indirect integration strategy between plants, the total annualized cost is reduced by 21% compared with practical values, and when considering both direct and indirect integration strategies, the total annualized cost can be reduced by 23%. Further, the analysis of how the change of operating parameters influence optimal solutions show that salt balance is significant for water network operation in steel industrial park, and the salinity in freshwater, the salinity regulations on inlet streams to water systems, as well as the performance of desalination all have significant effects on the water network optimization solutions.(5) Propose of wastewater zero liquid discharge scheme for steel industrial park. Due to that brine is the main block for steel industry to achieve wastewater zero liquid discharge, in this paper, two zero liquid discharge systems for brine are presented and compared. By developing unit model for the brine zero liquid discharge system firstly, and then integrating with the park-scale water network superstructure proposed above, the model for brine zero liquid discharge is developed. The industrial example results show that the developed superstructure and mathematical can achieve zero liquid discharge, and two stage reverse osmosis based brine zero liquid discharge system shows cost advantage over the other system.(6) Optimization of the water network in steel industrial park under uncertainty. Inside the water network of steel industrial park, the water systems are closely related, thus to improve the robustness of the water network, the two stage stochastic programming method is used to optimize the water network under uncertainty. The fluctuation of water quality of freshwater and the overflow of coking wastewater are considered as the uncertainties in the case study, and the results show that the maximum allowable flowrate in desalination，centralized wastewater treatment systems and brine-to-salt system increased under uncertainty, which means the water network cannot operate optimally by only adjusting the flowrates between units when uncertainty occurs, and also needs to increase the capacity of those water treatment units.