Cooling System for Ceramic

Are you pouring this molten stream in the water? If you are than I suspect the size of your "Rocks" are the size of the stream as it hits the water. You would need to break the stream up into smaller droplets, before cooling. Maybe spread it out wider and have it pour from higher up.

Yes I'm pouring the molten material into a Hopper with water, and you are right I need to break the stream into smaller droplest, I was trying with air at high presure and also with wáter at high presure too, but don't spread out, maybe is due to the high viscosity.

Can the melt be extruded in a stream of 1/8" to 1/4" stream ?

I can't extrud the stream, due to the high flow, probably the diameter of the stream is 1/2".

I make bieg rock,.,., leetle rock but quick....

Call the GIANTSMASHER.....work cheep

I would guess that your problem is that you are not cooling fast enough to prevent small particles of ceramic from recombining after it has hit the water.

Some ideas you might try but never having done anything on melting ceramics I don't know how useful these will be.

1 Split the stream into smaller streams. When making lead shot the molten lead is poured through a brass plate with an array of small holes that splits up the single stream. The holes are sized so that instead of multiple streams the lead is split into droplets. It then drops through a great height (60 feet/20m) to allow the droplets to become spheres before hitting the water. I did a project once to use the same technology to make wax loaded with pigments into "masterbatch" for coloring plastic. Because of the different cooling characteristics the result was not spheres but 10mm diameter 2mm thick flying saucer shapes. They dispersed well when mixed with the granular plastic feed stock so the shape in that case was not critical. Given your temperatures (3812^o F/2100^o C the plate would need to be made of tungsten (melting point 3400^o C) or a similar high temperature material.

2 Increase the depth and flow rate of the cooling water. If you can achieve droplets then when they hit the water a steam bubble will form around each drop which should keep them separated. You need to transition fully from liquid to solid in the water before the droplets hit the base of the tank where if they are still soft they will recombine. We installed a submerged inclined conveyor to continuously carry our formed wax platelets out of the tank. The steam will create a floatation jacket for each drop so the water must flow fast enough to carry drops out of the path of further falling drops. Making the plate hole array into 2/3 lines of holes stretched across the direction of flow will assist.

3 Adjust the temperature of the initial stream to as low as possible. This will be a trade off between temperature and viscosity. Glass has a melt point of 1100^o C and a maximum crystallization point at 850^o C. 2100^o C seems very high as a starting point.

The separability of such a stream of molten ceramic into multiple small streams and free-fall droplet formation would take an incredibly high tower, due to the higher melt viscosity (I suspect) than for lead, steel, etc. Also surface tension and density go into such formulae.

Possibly, if the melt head pressure at the nozzle can be safely increased, that might help. I would apply cool air flow against the downward motion of the melt droplets/streamers.

If sufficient cooling has taken place at the bottom of the cooling column, then the pieces break up on impact to optimal size and do not stick together.

1 Any piece you want to put to manage the stream most be cooled with water due to the high temperature an also due to this and high viscosity any device you put in front of the stream, sticks all around the piece, I like the idea of the plate, do you have a sketch or drawing of this to share it with me?

2 I have two spray with high pressure of water hitting the stream, but is not enough to produce small shot.

3 I can't reduce the flow and the temperature because the process need it as it is.

I suggested tungsten because you can run it at 2100°C without needing to externally cool it. Contact Midwest Tungsten Service in Willowbrook, IL in USA who can both supply and machine the plate to your requirements. I did some math but you need to check my figures. Assuming a density for your material of 3.9gm/cm³ a 5mm dia x 5mm long slug of material weighs approx 0.4gm. To process 1 ton/hour you need to make 1300 slugs (drops)/second. Assume the plate holes are sized to create one drop per second you need an array of 5 rows each 320 holes long. The hole size will depend on your viscosity but the drop produced will always be bigger than the hole so let us guess at 4mm diameter. Creating droplets by passing material through a plate is a complicated science so you need expert advice but water jetting the holes in the plate allows you cut holes that taper and this should help. If you space the holes in a 25mm x 25mm array but offset each row by 5mm from the previous row then water jets directed from the side are presented with an wall of drops that vertically has gaps between each drop but horizontally is apparently solid. That should create an efficient water spread and would give you a plate approx 8m long by 250mm wide by (guess)5mm thick. Your cooling load for alumina ceramic assuming a sp ht of 880J/kg°K comes to about 160Kw to drop the temperature by 600°C which equates to 1600litres of water per hour at 100% efficiency. Your heat transfer efficiency is unlikely to be more than 20% so 8000l/hr is the minimum I would suggest. 5-8 rows of jets with a vertical separation of 300-400mm per row is one option but this would need further math to confirm the most efficient configuration.

I'm going to check all the information you gave me, and thanks for the tungsten contact I was checking the web site and I see that they work with materials for high temperature like molybdenum. thanks in advance for your help

I wonder if there is a way to heat the block of Tungsten (Wolfram), or Molybdenum independently from heat imparted by the molten ceramic, such that a positive differential temperature is maintained, thus assisting with control of flow of the melt.

I suggested that you check my math because it looked off. It was ( my bad). One tonne per hour is only 700 drops/second not 1300. Yesterday I changed a tap washer to fix a dripping tap at my daughter's house. While doing this I roughly timed the frequency of the drips as about 1Hz and as a result I have revised my estimate of a realistic production rate to 2 drops per second. So the plate would need only 350 holes in 5 rows of 70. This gives a much more practical size of 2m x 250mm.

The method for use of a shot tower is to run the stream through a screen or perforated plate of appropriate diameter and spacing. I don't know what you could use that would stand up to the temperature. The alternate is to impinge on the flow with gas jets with number of jets, diameters and pressure adjusted to maximize the size that you need.

Another approach might be to set temperature to allow you to mechanically draw off or extrude filaments of ceramic, and then mechanically break the filaments into short cylinders.

I think a monolithic block of pyrolyzed anthracite carbon, appropriately drilled as a distributor plate (provided the thickness of said plate is engineered to provide sufficient support of the melt pressure head exerted upon it) should work really well. I can see that this plate would need to be replaced periodically. Another option might be silicon carbide, and I think there are companies out there that can drill the holes in this to make it work.

It really depends on which material will stand up to atmospheric attack at those temperatures. In the fiberglass industry, do they not use platinum or palladium draw plates?

In Fact we use two SS material heads cooled with water and running at high speed to produce ceramic fiber, and I pour the stream to the Hopper with water just when I have a problem with the heads.

Now I have two jets with high pressure of water but are not enough. the jets produce small filaments of ceramic but a small quantity.

Could you describe the finished product from the melt in a bit more detail? Is it "shot" particles, or is it elongated fibers (as in rock wool insulation)?

The SS heads you mentioned? How are these arranged, are they perforated plates with water in passages between the perforations so that the ceramic and water are not supposed to be in direct contact?

What class of water purity are you using? I presume that if there is any type of metal heat exchanger (and not direct spray of water onto the ceramic draw), the water has to be devoid of minerals, or deposits would kill heat transfer.

Is it a new process you're trying to establish? Most things in the ceramics industry have been going on for many years, 1000's in some cases. Around Stoke-on-Trent every other old boy you speak to is an ex-potter!

If it's been done before maybe somebody in or from the industry can help.

Ramoncrispin,

Many industries do pilot plants to develop and test the process--I suggest you do this on a pilot scale and after getting it working then change the scale to production.

Have you considered sound at an appropriate frequency as a means of breaking the filament stream into drops, perhaps applied to the filament stream at the plate it extrudes through?

JMM

Have you any experience with sound-induced droplet control? I have seen that work, but only on molten exotic salts that still have a low melt viscosity.

These molten ceramics have a higher viscosity (as I imagine, but I could be wrong). Higher viscosity might not go hand in hand with sonic droplet method.

JS,

No experience, just an awareness of different technologies and that a standing acoustic wave may help.