fluid mechanics in steel casting

[gieter]

I’m methods engineer in a steel foundry. We make castings up to 25 tons casted weight. The metal has to be poured fast to avoid preliminar solidification and break down of the sand mould. To give you an idea, a 25 ton casting will typically be poured in 2 to 3 minutes.

The metal is poured out of a ladle with a hole in its bottom that can be opened by lifting a stopper rod. In the mould standard size ceramic tubes, bends and T’s will direct the liquid metal into the mould cavity, preferably from underneath to avoid splashing and reoxidation of the steel.

To be honest, little calculation is used to dimension this gating system. The metal velocity at the entrance of the down sprue is dependant on the level of liquid metal in the ladle, the size of the exit nozzle and the distance between the bottom of the ladle and the entrance into the mould (typically 2 to 4 m/s). At present the gating system is oversized to be sure it can “swallow” the metal delivered from the ladle, without the operator having to adjust the metal flow with the stopper as this would cause heavy turbulence and hence reoxidation and slag formation in the casting. But due to the over sizing of the gating system, this is only filled completely at the end of the pour. Especially in the beginning the metal stream is smaller than the internal diameter of the tubing. Again a problem of reoxidation occurs, as air can be sucked in with the turbulent metal stream.

A long story to come to my question: Can anyone put me in the direction of how to calculate what is happening in the early stages of the filling of the casting, when the metal stream is small compared to the size of the tubing and encounters T’s and bends while not completely filling the system. How are losses calculated? Does there exist software to simulate this behaviour. We use casting simulation software with a filling module based on the Navier-Stokes equation, but the losses due to friction are not sufficiently modelled and the software has problems handling free falling liquid streams.

Any help is much appreciated,

 

[MikeHalloran]

Now that you mention it, I can't think of a good reason why the down sprue should be exactly vertical, and you've pointed out some good reasons why it might work better if tilted say 30 degrees off vertical.

I'd be inclined to make up some small scale transparent plastic models of typical gating systems, and fill them with water in the sink. Take videos of the process, and study the behavior as you adjust the geometry. Of course, water is not molten steel, but if you can get the miniature water system to develop geometrically similar flows to what you've been able to intuit and model in steel flows, you might be able to make better guesses on what to model in math and what to try in steel.



Mike Halloran
Pembroke Pines, FL, USA

 

[BigInch]

mercury???

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[gieter]

Mike,

One reason why not to tilt the down sprue is security. The entrance to the gating system has to be vertical with the ladles we use. A change in direction would mean a bend and this should be as close as possible to the entrance for the tilt to be effective. A stopper ladle should be either closed or open and at the start of the pour you get your maximum flow as the level of the metal in the ladle is at its highest. So the beginning is quite violent and any disturbance in the flow can result in a back flow and a splash of molten metal at 1600°C with a pouring crew standing next to it.

But, I like the idea and I will look in to it to see if this can be overcome.

I also agree on your suggestion on water experiments. The problem is as ever: time and resources. But this, of coarse, is a cheap excuse.

Mik

 

[MikeHalloran]

I had another brain fart.

You could reduce reoxidation and turbulence at the leading edge of the pour by stuffing the gating system with styrofoam, say as crushed scrap, or preforms inside the ceramic pipes, or whatever. Okay, for steel castings, maybe the little extra carbon would not be a good thing.







Mike Halloran
Pembroke Pines, FL, USA

 

[metengr]

Can anyone put me in the direction of how to calculate what is happening in the early stages of the filling of the casting, when the metal stream is small compared to the size of the tubing and encounters T’s and bends while not completely filling the system. How are losses calculated? Does there exist software to simulate this behaviour. We use casting simulation software with a filling module based on the Navier-Stokes equation, but the losses due to friction are not sufficiently modelled and the software has problems handling free falling liquid streams.

There might not be a need to re-invent the wheel. In my various audits of foundries, many of the reputable foundries use casting simulation software called "MAGMASOFT" and "JMATPRO" (which determines the solidification path and all casting relevant material properties using thermodynamic equations).

 

[gieter]

"Magmasoft" is known to me and is not a solution to my problem. "JMATPRO" is new to me, so I will check that out. Casting simulations focus, as you say, on thermodynamics. The fluid dynamics are mostly simplified to reduce calculation time. Inlet conditions are defined by the programmers, so you have limited possibilities. Also most software, like Magmasoft, use a fully structured, orthagonal, finite difference mesh, resulting in the infamous zig-zag in curved surfaces, making it unpractical to study losses.