.png){kind=icon}
.png){kind=icon}
Return
A novel node-based non-structural model for mass exchanger network synthesis using a stochastic algorithm
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jclepro.2022.134227.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The use of mass exchange networks is an effective means of reducing both pollutant emissions and investment costs in chemical industrial processes. Here, a new simultaneous model based on a mixed integer non-linear program called the node-based non-structural model is proposed for the synthesis of mass exchange networks. This model uses the concept of nodes in both rich and lean streams to represent the possible generation positions of mass exchange units, with no fixed matching pattern. The presented model is more random and flexible in forming structures compared to other structural models with fixed matching relationships. These characteristics, in addition to the adoption of non-iso-compositional mixing, resulted in the search region of the proposed model being significantly enlarged. The established model was also modified to incorporate networks with regenera-tion. However, the high nonlinearity and nonconvexity of studied problems make it difficult to obtain a global optimal network design. To overcome this, a random walk algorithm with compulsive evolution, a kind of stochastic algorithm first used in solving heat exchange network problems, was applied for the solution. The algorithm's strength is its feature of independent evolutionary strategy among individuals, which can optimize both the integer and continuous variables simultaneously. The effectiveness of the proposed method was demonstrated by solving four different examples from the literature. For each of the four examples, a lower total annual cost result which satisfied all kinds of constraints was obtained compared to previous studies.
Keywords:
Mass exchanger networks
Process integration
Simultaneous method
Non-structural model
Stochastic algorithm
Journal
IF:
10
Papers:
4.6W
Citations:
36.8W
Organization
No organization information available