基于经验规则的数学规划法快速设计多杂质用水网络

摘要

A rapid design method for water using network with multiple contaminants was developed by mathematical programming based on rules of thumb.One effective strategy to solve complex or large-scale water network is that to simplify the algebraic model.However,obtaining accurate sequence of water process orders,such as water source or water demand,is the key to eliminating unnecessary stream and simplifying mathematical model of water network.Therefore,according to the analysis of limiting data and the understanding of the water process,a new sorting method of water supply and acceptance was proposed,then the necessary water pinch rules,sequences and approximation rules,and mixing rules were also set in order to remove unnecessary water network structure,simplify mathematical models and reduce the dimension of network composition.Finally,the problem with NLP (nonlinear program) model was solved with the LINGO (linear interactive and general optimizer) language.Three literature examples were used to illustrate the proposed approach.The result showed that setting the rules could effectively reduce model scale and nonlinear variables,shorten solving time,and obtain optimal design.%综合了夹点规则、排序思路等经验,提出了基于经验规则的数学规划法来快速设计多杂质用水网络.由于获取准确的用水过程顺序是剔除一些不必要流股、简化用水网络数学模型的关键,因此,基于对用水过程极限数据的分析和对用水过程的理解,提出了用水过程供水和受水的排序方法,并设定必要的水夹点规则、序列和逼近规则、混合规则以剔除不必要的水网络结构,达到简化模型、降低网络结构维数的目的,最后在LINGO软件上编写程序进行求解.通过三实例分析得知,设定经验规则后能够有效地降低模型规模、减少模型的非线性变量数、缩短模型的求解时间,并能够得到最优的设计方案.