DFG Project PF 394/2–1

Numerical and experimental investigation of stationary and transient flows in tundishes

Sin­ce inves­ti­ga­ti­ons of the real melt flow in the steel­works are not fea­si­ble due to the extre­me boun­da­ry con­di­ti­ons and the lack of opti­cal acces­si­bi­li­ty, the­se inves­ti­ga­ti­ons were car­ri­ed out in com­pli­ance with the cha­rac­te­ristic simi­la­ri­ties on a redu­ced water model dis­tri­bu­tor on a sca­le of 1:1.7. Laser Dopp­ler ane­mo­me­try and par­tic­le image velo­ci­me­try were used as mea­su­ring methods. The nume­ri­cal cal­cu­la­ti­ons were per­for­med with the com­mer­cial flow sol­ver FLUENT based on the RANS equations.

The nume­ri­cal and phy­si­cal simu­la­ti­on of the sta­tio­na­ry cas­ting show­ed that very hete­ro­ge­neous are­as with strong vor­ti­ces are pre­sent in the tun­dish. The domi­nant vor­tex struc­tu­re is a hor­se­shoe vor­tex, which lies around the shadow tube beam and is pre­ser­ved up to the out­let area, whe­re it rests on the tun­dish base. With regard to flow inten­si­ties, the mani­fold can be divi­ded into three zones, the inlet zone to x/L1 0.3, the cen­tral zone 0.3 x/L1 0.6 and the out­let zone x/L1 > 0.6. The inlet zone is cha­rac­te­ri­zed by a high inter­nal recir­cu­la­ti­on, which cha­rac­te­ri­zes it as “well-mixed”; in the x‑direction, the inter­nal recir­cu­la­ti­on decreases.

Tran­si­ent cas­ting is inves­ti­ga­ted using the lad­le chan­ge as an exam­p­le. The results show that with the omis­si­on of the input impul­se, the vor­tex inten­si­ty in the dis­tri­bu­tor rapidly decrea­ses. Back­flow is com­ple­te­ly sup­pres­sed so that a pis­ton flow occurs in the mani­fold. After cas­ting on the new lad­le, it takes approx. Δθ = 0.12 until the struc­tures known from sta­tio­na­ry cas­ting are res­to­red. The velo­ci­ties and thus also the back­flows in the mani­fold are con­sider­a­b­ly hig­her when cas­ting the new lad­le than when sta­tio­na­ry cas­ting, but this still has to be quantified.

The com­pa­ri­son bet­ween the expe­ri­men­tal and nume­ri­cal inves­ti­ga­ti­ons car­ri­ed out for both sta­tio­na­ry and tran­si­ent cas­ting shows that with FLUENT it is pos­si­ble to repro­du­ce the flow in the con­ti­nuous cas­ting dis­tri­bu­tor in a reasonable way. Howe­ver, it is important to use rea­li­stic boun­da­ry con­di­ti­ons. In tran­si­ent cas­ting, this means that even pro­grammed boun­da­ry con­di­ti­ons are inte­gra­ted into the model.

Logo DFGThe pro­ject was fun­ded by the Deut­sche For­schungs­ge­mein­schaft (DFG) under the refe­rence num­ber PF 394/2–1.