对蒸汽网络系统进行建模与优化,可以在系统配置和操作条件满足公用工程要求的前提下,以最小产汽量为目标,实现化工厂的节能运行.本文利用在一个真实化工厂现场调研得到的实际数据,对该化工厂的蒸汽网络系统进行建模,得到LP(线性规划)和MINLP(混合整数非线性规划)模型,用线性规划算法和改进文化差分算法分别对模型进行求解,并与现场实时数据和LINGO计算结果进行对比.计算结果显示,MINLP模型更好的描述了该蒸汽网络系统,并达到了可观的优化效果,可在满足生产需求的条件下,减小锅炉产汽量和放空量,达到降低成本、提高经济效益的效果.
Modeling and optimization for steam network systems can realize the energy conservation for chemical plants, while at the same time meet the systems configuration, operating conditions, and minimum utility demands. A LP (Linear Programming) and an M1NLP (Mixed-integer nonlinear programming) model are established based on the large amount of industrial data of a real-world chemical plant. Binary variables are added to represent the operation state of the devices. The two models are solved by linear programming algorithm in LINGO software and improved cultural differential algorithm in Visual C++, respectively. After comparing with the real-time field data and the LP model's results (solved by LINGO), the computing results of the MINLP model (solved by the improved cultural differential algorithm) shows outstanding characteristics: lower steam yields and higher economic efficiency.