Momentum-preserved node for fusion magnet cryogenic network analysis and its implications. (October 2018)

• Record Type: Journal Article
• Authors: Oh, Sangjun; Lee, Hyun-Jung; Oh, Dong Keun; Jung, Laurent
• Is Part Of: Cryogenics. Volume 95(2018)
• Page Start: 110
• Page End: 115
• Publication Date: 2018-10
• Subjects: Fusion magnet -- Cryogenic network analysis -- Thermo-hydraulic analysis -- AC-loss estimation; Low temperature engineering -- Periodicals; Low temperature research -- Periodicals; 536.56
• DOI: 10.1016/j.cryogenics.2018.09.007
• Languages: English
• ISSNs: 0011-2275

Abstract

Highlights: 0D node for fusion magnet cryogenic analysis cannot be regarded as a reservoir. 0D node (e.g. volume node in Supermagnet code) is a frictionless pass-through. Novel concepts on 0D nodes having momentum are developed. Frictional force cannot hinder backflow when there is AC loss. Empirical AC loss estimation error can be quite substantial, as high as over 10%. Abstract: Cryogen supply is one of decisive factors for the analysis of fusion magnet. In order to analyze cryogenic networks, numerical node joining each component together is required. In SUPERMAGNET code, such joint called as 'volume' node, behaves like a small reservoir, cryogen completely loses its momentum. However, when there is severe AC loss, especially at inlet, transient massive backward flow can occur and the assumption on total momentum loss is a bit dubious. Here, we discuss plausible momentum-preserved node concepts suitable for thermo-hydraulic analysis of fusion magnet. Based on requirements of cryogenic network solver, two types of node, namely 'plenum' and 'manifold' are proposed and compared. Plenum node is an extension of volume node including velocity and manifold is a quasi-0D node calculating velocity similar to finite volume method (FVM). Interestingly, the 'volume' node does not behave like a reservoir. Instead, it is more like a frictionless pass-through. Therefore, we argue that we need to introduce additional pressure drop at both inlets and outlets, somewhat like bending loss. Furthermore, if there is a severe backflow at inlet then it is also possible heat generated by AC loss can detour to another path, to other magnet. A case study has been carried out and it is shown that AC loss estimation error can be quite substantial as high as over 10%.