A New Pairing Method for Multivariable Processes
Yongfang Cheng,Shaoyuan Li
DOI: https://doi.org/10.1021/ie901787u
2010-01-01
Industrial & Engineering Chemistry Research
Abstract:As far as the pairing of multivariable processes is concerned, with the existing method based on the relative gain array (RGA) and Niederlinski index (NI), the interactions among loops are analyzed by the open-loop gains change when all the other loops change from open states to closed states. In this method, the internal interactions among the other loops are hidden, that is, several groups of other loops with different internal interactions can result in the same effect toward the loop considered. Therefore, according to the existing method, there may be more than one feasible way of pairing. To analyze the interactions under these feasible pairings further and get one pairing way with relatively weak interactions, in this paper, a new method of pairing is proposed based on the analysis of interactions between any two loops. By means of relative gain table and relative gain graph, the interactions between two random loops of an x n system can be analyzed vividly when the number of closed loops within the (n - 1)-dimensional subsystem decreases gradually. On the definition of a new criterion to weigh the mutual interaction between loops, a new pairing method is brought forward. Finally, three examples are used to validate the effectiveness of the proposed method.