In this example Flowservice was asked to optimise a die already in production, which had significant problems causing large amounts of scrap. The picture shows the geometry of the part.
Problem found during production
The main problem of this part appeared on the cast product as several surfaces with poor surface quality. The main cause of this kind of problem is usually the variation of metal temperature during filling.
The main factors related to this phenomenon are:
1. Part thickness
If not correctly designed, thicknesses unbalance the thermal map during the various filling stages.
2. Casting shape
Castings with particular shapes, for example with restrictions or islands inside the part, increase the problems related to temperature differences.
3. Filling time
If the gates are well designed, the shorter it is, the better the surface finish. As an indication, given the average thickness of the part and the surface requirements, the filling time should not exceed 0.035 – 0.040 seconds, calculated with the NADCA formula. This time, however, depends on the runner geometry, which should be designed better than the current one.
4. Runner design
It must not have restrictions, hot spots or section increases. The section at the gates should not be larger than in the adjacent part of the runner, as instead happens in this case.
5. Gate velocity
It must be high, while staying within the limits related to die soldering: 45–50 m/sec must not be exceeded. This parameter is directly related to the desired filling time.
6. Alloy temperature
The higher it is, the better the surface finish; care must be taken with oxidation, hydrogen absorption and soldering at excessive temperatures. The recommended range is between 680 °C and 700 °C.
7. Die temperature
The higher the temperature, the lower the risk of these defects (an optimal range for the temperature at ejection is between 250 °C and 330 °C, with the higher values of the range for thick parts such as this one).
Auxiliary die coatings can also be considered to delay cooling in critical areas.
Evaluation of the current process
Before simulating the casting process for this part, some working conditions used in the first productions were checked. The most important finding was that the shot sleeve fill ratio for this part is far too low. Keeping all the other parameters unchanged, this may worsen the final quality of the part: such a low fill ratio does not guarantee correct filling of the sleeve, since it causes excessive temperature drops already at this stage of the cycle, and it increases the problems related to gas entrapment during filling, with more air in the die.
With a 100 mm plunger and a machine with 580 mm of useful sleeve stroke, the fill ratio is no higher than 39%. This value is too low: it is advisable to work in a sleeve fill range between 50% and 75%, preferably at the higher end for parts with special requirements. We therefore recommend using an 80 mm plunger, if available, or changing machine to shorten the sleeve. Bear in mind that a machine with a relatively high clamp force is needed to cast this part at high filling velocities (short time), to reduce the importance of temperature drops during the process.
Parameters used in the simulation with the original geometry
- Materials: Die = H13; Metal = Al-EN-AC 46100 AlSi11Cu2(Fe).
- Heat transfer coefficient between metal and die = varying from 2500 W/m^2/K to 7500 W/m^2/K in the 350-700 °C range, with a linear profile.
- Metal temperature at the biscuit surface = 660 °C.
- Average first phase velocity = 0.25 m/sec.
- Second phase velocity = 3 m/sec (related to the size of the runner cone).
- Average die temperature = 260 °C.
- Plunger diameter = 100 mm.
- Gate area = 606 mm^2.
- Smallest flow section in the runner = 400 mm^2.
- Total sleeve length = about 580 mm
Simulation findings on the original geometry
Simulation provided several indications on the geometric configuration of the part.
The geometric evaluations can be divided into two parts: the first concerning the runner and the second the part itself. In the runner layout a couple of weaknesses affecting the quality of the final part can be identified. Starting with the runner next to the sleeve, the picture above shows that it has a constant section along its length, which causes air entrapment already in the first filling phase (grey area in the picture). A runner geometry with a decreasing section would be advisable, making this section tapered instead of straight.
A second remark on the runner design is that the following tapered section does not favour good die filling: its section even increases, rather than staying constant as in the previous case. As a result the highest velocities occur at the restriction in this tapered section and not at the gates, as they normally should. This is unfavourable for cavity filling.
The previous picture, showing the velocities, makes clear that they actually decrease along the tapered section, causing unwanted air entrapment due to the non-linear velocity pattern of the metal. This section should therefore be reconsidered and designed more appropriately.
Several considerations can be made on the part geometry. First of all, the thicknesses are not properly distributed: many areas are thicker at the end of filling than in the areas that fill first.
As a first remark, the thickness of the wall next to the runner (the vertical circular wall) should be increased overall: the current thicknesses do not give the metal enough room to fill without excessive temperature drops. We therefore recommend increasing this thickness.
Also in the first filling section, the flow changes direction too abruptly: the existing fillet radii are too small, so larger radii are recommended to let the metal flow better.
The previous picture also shows that the holes and bosses of the part greatly influence filling: the areas shown in red cause a considerable drop in metal velocity next to them. Preferential high-velocity paths clearly form on the vertical wall of the part. Thicker flow channels could therefore be considered in the areas where lower velocities concentrate, to reduce this negative effect, which delays the filling of the vertical walls beyond these areas.
In short, the overall thickness of the vertical wall could be increased to about 3 mm, at the same time creating rectangular flow channels with a further 1 mm of extra thickness.
Finally, the horizontal walls (the 4 side walls) have too many thickness differences, with the thickness increasing at the end of filling. We recommend making these thicknesses as uniform as possible.
The previous pictures also show slightly preferential areas for the metal flow. These create air pockets that close inside the part. They can be found along the whole upper side wall, at the corners opposite this wall and in the central area of the two vertical side walls.
Based on these considerations, before testing the recommended changes on the die, a modified geometry was created and several simulations were run on it.
Simulation findings on the modified geometry, without overflows
The modified geometry included the following changes (the overflows on the die were not considered at this stage):
- Increased wall thickness of the circular front surface, with a consequent increase of the gate thickness.
- Increased wall thickness in the first section of the side walls.
- Larger fillet radii between the parts of the casting.
- Dimensional change of the first runner section, with a decreasing section.
- Larger circular section at the head of the tapered runner section.
- Three thicker bands on the circular front surface: towards the two edges of the part near the shot sleeve and towards the centre of the side surface opposite the sleeve.
These geometric changes led to the results shown in the following pictures.
The simulation used the same parameters as before, except for the plunger filling velocities: since the flow sections had been increased, an average first phase velocity of 0.3 m/sec and a second phase velocity of 4 m/sec were used.
As the picture shows, the changes to the first runner section were beneficial: during the first filling phase no air is trapped in the initial runner, as happened before.
A second positive effect of the new geometry is shown in the next picture. The overall metal temperatures are slightly higher than in the pictures of the original geometry under the same conditions. This was essentially the main result sought, and shows that the changes are positive, although not decisive.
Although the first goal was reached, some unacceptable situations remain during filling: as the pictures show, the flow front is still not correctly balanced in the final filling phase.
In addition, as the next picture shows, there are adjacent areas on the same surface with excessive velocity differences. This increases the imbalance during filling and creates turbulence in the metal, with consequent gas entrapment.
The picture shows the areas with stagnant metal (low velocities), marked by the coloured arrows. Red arrows indicate problems caused by the core pins forming the holes in the part, yellow arrows problems caused by the two openings in the casting. Removing the pins would greatly benefit the final quality of the part. For the two openings marked by the yellow arrows, bridges could be added to let the metal flow, reducing the opening section, or, if possible, the openings could be closed completely.
Simulation findings on the further modified geometry
Following the previous considerations, further geometric changes were made to improve the situation even more:
- Thicker bands towards the edges made even thicker.
- Two more thicker bands on the circular surface towards the other two edges of the part (narrower).
- Removal of the core pins on the front surface.
- Two material bridges in the two hollow areas of the part highlighted in the previous results.
The results are shown below.
The pictures show that removing the pins greatly improved the velocity distribution on the front surface. This is certainly an improvement with respect to the problem found on the part with the original geometry.
Temperatures also improved, although some imbalance remains in the final filling section, as shown in the following pictures.
The next picture shows the side surface opposite the sleeve, where air is trapped because the flow front closes on itself when it meets the thicker area at the end of the part. To overcome this, overflows were used to draw the metal in the two side areas into two specially positioned overflows, giving the central part time to fill.
The same consideration for the vertical side surfaces led to using overflows to slow down the flow shown in the next picture, which reaches one side of the surface first, so that the other side can also fill correctly; an overflow will be placed there anyway to collect impurities, since the air remains in that area.
For the last side surface, as shown in the final picture, two overflows were added to collect the air flowing into that surface, especially at its sides.
The picture below shows the behaviour of the air inside the die at a given instant, which helps to understand the positions chosen for the overflows.
Based on these considerations, a geometric layout with overflows was created, as shown in the picture.
A further simulation was then run to check how the overflows work, with good results.