Abstract: The optimization of water-using network in a complete batch process system with multiple contaminants was studied in this paper. A superstructure of water-using network and the NLP（non-linear programming）mathematical model are constructed，and the quantity of flow in each water-using process is limited between the minimum and maximum in order to accelerate convergence. In this study，storage tanks are used to bridge the time gap. First，the key contaminant in each water-using process was ascertained，and then the superstructure was simplified by eliminating the reuse and recycle of the water resource with maximum concentration of the key contaminant. Second，the model is solved by GAMS（General Algebra Model System），and then the minimum fresh water and waste water discharge are gained as well as the number and volume of central storage tanks. Third，the optimized water-using network is obtained. Two examples in literature are solved to illustrate the validity and advantages of the proposed approach，which requires fewer cycle periods of calculation and smaller number of central storage tanks than before.

Key words: completed batch processe, water-using network, multiple contaminants, optimization

摘要： 对多杂质体系完全间歇过程用水网络优化进行了研究，建立了用水网络超结构和非线性规划数学模型，利用中间储罐跨越时间约束，设置用水过程流量上下限以加速收敛。首先确定关键杂质，并将关键杂质浓度最大的水源直接排往废水，获得简化的用水网络超结构；其次，运用GAMS求解，得到最小新鲜水量、废水量、中间储罐数目及优化的用水网络结构。实例研究证明本文提出的方法可行，并能以较少周期获得优化的网络结构，同时可以减少中间储罐数目。

关键词: 完全间歇过程, 用水网络, 多杂质, 优化