Introduction.

From the mid 1970's, I have worked on improving the thermal conditioning of glass for container production. As glass thermal homogeneity in forehearths has been improved, better gobs have been produced and production figures have increased.

However, when the glass leaving the forehearth appears to be homogeneous, the gobs produced have not always been perfect. By making the glass conditioning in the channel slightly worse, better gobs could be produced. The conclusion from this was that the thermal condition of the glass deteriorated as it flowed through the spout.

Better insulation applied around the bowl and some changes to the bowl shape helped this situation, but did not cure the problem.

The nature of the glass flows in the standard bowl design had been studied by physical modelling at a very early stage in this work. A few years ago, I was asked to come up with new, improved designs for glass conditioning equipment and it was only then, that the potential of the new bowl shape became apparent.

This paper describes the development of this new bowl shape, from the initial idea, through the processes of assessment, to the development of a complete range of bowl sizes.

The problem.

Design modelling.

At this stage, physical models of the standard and the cascade bowls were constructed, in two depths, and studies were made of their flow characteristics. From this work, the benefits for glass conditioning seemed to be greater with the deeper design.

The studies showed us that, with the standard bowl , glass entering the bowl near to the surface is pulled round the bowl by the tube, and its temperature can only be affected by the bowl firing as long as it remains near the surface.

Glass entering from the bottom of the channel quickly moves away from the firing and loses heat to the generally under insulated bottom of the bowl entrance.

With the cascade bowl design , glass from all depths in the channel remains closer to the surface until well into the bowl, and will therefore be better controlled. As the glass moves around the bowl, it slowly moves downwards until it enters the lower well.

The cross section of this lower well is smaller than in the standard bowl, and so the glass falls faster. The time taken for the glass to move through the lower part of the bowl is much shorter than in the standard bowl, less heat is lost by the glass and temperatures will be more uniform.

This table summarises the changes in glass residence times for the two designs of bowl. The residence times in the lower zone (or well) of the cascade bowl are much shorter than in the standard bowl. These shorter times would reduce the conduction heat losses, and the resulting temperature differences would be smaller.

Results from factory trials.

The first cascade bowls were installed on coloured glass feeders in the container industry, and these feeders are still operating with the new bowl.

The first cascade bowl was installed in P.L.M. Redfearn, Barnsley, England in December 1991, and the next two in Saint-Gobain, Vauxrot, France in late 1992. The success of these installations has been indicated by many repeat orders for this type of bowl design.

A number of changes were observed on these first installations, the most significant of which, in chronological order, were :-

1. The curling of the gobs was reduced.

2. Cold glass near the bottom of the tube in the standard bowl was 15Deg C hotter in the cascade bowl and the differences around the bowl were reduced.

3. The lower firing level used in the bowl reduced the amount that the glass reheated as it moved around the bowl. The 24^o C reheat for the standard bowl was reduced to 12C in the cascade bowl.

4. During a very long job change on the cascade bowl, glass froze in the orifice ring and it took a long time to melt out. But after the glass was flowing, the gobs became stable much quicker.

5. The tube speed could be reduced without significantly changing the gob weights. A reduction in speed from 10 rpm to 5 rpm, caused the weights of the gobs to change by only 2g, or 1.5 %.

6. One cascade bowl was installed on half of a tandem feeder, the other half using a standard bowl, and both feeders fed a 16 section machine. Both feeders were being supplied with the same glass conditions and the short term weight variations were lower on the cascade bowl side.

7. The machine was fitted with the Heye gob weight control system which essentially records the volume of each measured gob. A comparison between these measurements was taken over 13 minutes. The standard deviation of the standard bowl variations is 2.27 times worse than the deviation on the cascade bowl.

8. The longer term variations in bowl temperatures are represented by the variations in tube height produced by the gob weight control system. Again, the variations are significantly better on the cascade bowl feeder.

9. Cold glass in the bowl entrance was eliminated, with the top-to-bottom temperature differences being reduced from 105^oC to 25^oC. The side-to-side temperature differences were also more symmetric and were reduced from 100^oC to 15^oC.

10. A significant reduction in certain bottle defects was achieved.

11. After 17 months service in P.L.M., the first cascade bowl was changed. By lowering the tube it was found that the glass flow could be stopped. The tube seat was badly worn, but the damage was very uniform around the seat, and the wear of the upper shoulder was minimal. The uniformity of the wear supports the modelling predictions that the flows under the tube would be symmetric.

Expanding the range of designs.

At the time, we did not have the facilities to carry out a mathematical analysis in 3 dimensions of the flow and temperatures within this new bowl design. We were therefore not aware of the most suitable parameters to use for the design of a whole range of bowls.

We had hoped that a small range of designs would satisfy the market needs, as the cascade bowl did appear to have a wider operating range than the standard design.

The glass container industry is renowned for the variety in equipment used and the different ways in which that equipment is operated. One factory may be able to operate a feeder outside its original specification, whereas another factory may be unsuccessful. This may be due to differences in local operating conditions, quality requirements or the production expertise.

We initially thought that, as the first designs worked well, we only had to apply scaling factors to the design to create a complete range of bowls. This approach may have worked in some cases but it was hardly based on sound technical reasoning, nor did we feel completely in control over the design process.

We first became aware of potential problems when a customer fitted a larger tube to a cascade bowl that had worked well on another line. We had not anticipated any problems with the larger tube but, on this line, the tube had to be raised so high that the improvements expected from the cascade design were not there.

The solution.

We first considered the various control volumes within the bowl. Glass has to flow through various gates on its way to the orifice, and the sum of the viscous resistances of these gates controls the glass flow rate.

With the standard bowl the major control volume is the gap between the tube and the tube seat, but with the cascade bowl there is an additional resistance within the annulus outside the tube and inside the well. The effective viscous resistances of these regions can be calculated using standard equations, and from this the relative flows can be determined.

To compare the operating conditions for a cascade bowl with a standard bowl, it is necessary to equate the relative tube heights for the same flow rates.

We first had to ensure that, to maintain the maximum pull, the tube height would not have to be raised too much.

Using a tube height of 40 mm (1.5") as a reference point, we calculated a well diameter which would ensure that the tube height would not increase by more than 10 % (or 4 mm). This choice of reference point may seem to be purely arbitrary, but if a different reference height had been chosen then a different "% increase" value would have been used.

If this predicted movement in tube height is too small, as would be the case for a very large well diameter, then the thermal and flow benefits of this design would be significantly reduced.

The upper zone depth, is normally fixed at 150 mm (6"), which matches the standard forehearth glass depth.

The upper zone diameter is chosen to ensure a shoulder width around the well of between 75 and 125 mm. In most cases this produces a diameter larger than the bowl entry width.

The throat dimensions are usually defined by the customer and are related to the diameter of the tube and the orifice ring equipment being used.

Variations on the bowl design.

All of the data presented so far has related to the design of cascade bowls which conform to our standard design of feeder. But some designs are more difficult to accommodate, in particular where the centre distance, from the channel joint to the orifice centre, is too small. We are currently considering three variations to the design which can be applied to these feeder designs :-

1. Introduce a flat face into the channel joint side of the well. Horizontal flow rates will increase as the glass passes the restriction, and the vertical glass flows will be reduced. Neither of these effects should change the glass temperatures significantly as they only occur for a fraction of the time.

2. Cut away the entry side of the well and make the channel front face complete the well. The only potential problem here is the effect of the very small dead region on either side of well entry.

3. Move the channel joint backwards, to make room for the well. This is obviously the best technical solution, but the cost could be much higher as it involves modifications to the last forehearth block and maybe the casing.

Conclusions.

In my paper, I have described the work carried out to improve the operation of equipment which is vital to the operation of a glass container production line.

This concept could, equally, be applied to the manufacture of tubing, tableware, T.V. tubes and lamp bulbs. Everybody is striving for greater consistency in glass temperature and flow rates to improve their production efficiencies. This design has shown that it has attributes which are beneficial, when applied specifically to the container industry.

I have shown how an idea, which is first rationalised on paper, is assessed using modelling techniques. A prototype design is then tested under glass, and finally a family of similar products suitable for most customer's needs, is developed.

To conclude, I would like to thank the customers who helped us in the initial assessment of this new concept. I feel that their need to improve their processes, plus the potential benefits indicated in the modelling studies, ensured that we were allowed to test the cascade bowl under glass.

Mike Stanley, B.H.-F.(Engineering) Ltd.

12th October 1994.