In my engineering practice, I have repeatedly observed that the quality of ductile iron castings depends not only on melt chemistry and nodularization treatment but also on the way the casting is gated, fed, and vented. A complete casting process design for ductile iron castings includes the pouring system, the feeding system, and the venting system. In addition, the designer must account for liquid metal fluidity, cooling rate, phase transformation, and the complex interaction between these factors. Relying only on theoretical equations and empirical experience is often insufficient to produce ductile iron castings without internal defects. Trial production can solve some problems, but it consumes a great deal of time and money, and it increases the final cost of the product. In recent decades, commercial software packages have appeared that are capable not only of designing the gating and riser system but also of simulating the entire casting process. Experto-ViewCast is one of these packages. It is a three-dimensional finite element simulation tool. The special pouring module and riser design module transform a purely numerical simulation into an active design assistant. In this article, I describe how I applied this software to the auxiliary design of the casting process for a valve body made of ductile iron castings. The method that I used is based on direct solidification-time calculations, defect prediction, iterative modification, and finally a full filling and solidification simulation.
Difficulties in Designing Casting Processes for Ductile Iron Castings
Ductile iron castings present a particularly interesting solidification problem. During cooling, austenite forms and carbon is rejected into the remaining liquid. Graphite nodules grow during the eutectic reaction, producing a volumetric expansion that can partly compensate for liquid and solidification shrinkage. However, this expansion can only help if the mold is sufficiently rigid and if the feeding path remains open at the right time. If the feeding path closes too early, the graphite expansion cannot reach the isolated hot spot. As a result, shrinkage porosity can form even when the riser volume appears to be mathematically sufficient. This is one of the reasons why the design of ductile iron castings cannot be reduced to simple hand calculations.
Traditional riser design for ductile iron castings relies on the modulus concept. For any solidifying body, the local thermal modulus is defined as the ratio of volume to cooling surface area. In many cases, the modulus is computed as
$$M = \frac{V}{A}$$
where V is the volume of the metal that is freezing and A is the area of the interface through which heat is lost. The solidification time can then be estimated by Chvorinov’s rule as
$$t_s = K \left( \frac{V}{A} \right)^n$$
in which K is a mold-material constant and n is usually close to 2. Although this equation is useful for simple shapes, it becomes difficult to apply to complex ductile iron castings with thick flanges, cored holes, and changing cross-sections. More importantly, the modulus method does not easily account for thermal interaction between the riser, the riser neck, the runner, and the casting. For ductile iron castings, the solidification sequence is highly nonlinear because of latent heat release and graphite expansion. Therefore, a finite element simulation is more reliable than an isolated modulus calculation.
Another difficulty is that the feeding requirement of ductile iron castings is not constant. The required liquid volume depends on the carbon equivalent, the mold rigidity, the pouring temperature, and the local cooling rate. The effective feeding volume can be written as
$$V_f = \alpha_{l} V_c + \alpha_{s} V_c – \alpha_{g} V_c$$
where Vc is the volume of the casting zone to be fed, \(\alpha_l\) is the liquid contraction, \(\alpha_s\) is the solidification contraction, and \(\alpha_g\) is the graphite expansion coefficient. For ductile iron castings, \(\alpha_g\) can be large enough to make \(V_f\) very small or even negative. This is why many ductile iron castings can be fed through the gating system or by a small feeder. However, I must emphasize that the beneficial effect of graphite expansion only appears if the solidified shell can contain the internal pressure. An early solidification of a thin wall can isolate a thick section and completely destroy the feeding path. The software allowed me to visualize this effect before making the pattern.
The table below summarizes the limitations of the traditional approach that I encountered when working with ductile iron castings.
| Traditional Limitation | Effect on Ductile Iron Castings |
|---|---|
| Modulus calculation is difficult for complex geometries | Feeding system may be undersized or incorrectly placed |
| Riser–neck–casting interaction is ignored | Neck solidifies prematurely and blocks feeding |
| Graphite expansion is not directly considered | Feeding volume is overestimated or underestimated |
| Thermal fields are not available before production | Trial iterations are needed |
| Trial production is expensive and slow | Product cost increases |
Numerical Simulation Workflow for Ductile Iron Castings
My workflow for ductile iron castings uses Experto-ViewCast as the central design environment. The software is based on the finite element method and solves the heat-transfer equation in three dimensions. For a differential control volume, the transient heat conduction equation can be expressed as
$$\rho c_p \frac{\partial T}{\partial t} = \frac{\partial}{\partial x} \left( k \frac{\partial T}{\partial x} \right) + \frac{\partial}{\partial y} \left( k \frac{\partial T}{\partial y} \right) + \frac{\partial}{\partial z} \left( k \frac{\partial T}{\partial z} \right) + \rho L \frac{\partial f_s}{\partial t}$$
where ρ is density, cp is specific heat, k is thermal conductivity, T is temperature, t is time, L is latent heat, and fs is solid fraction. The last term represents latent heat release during solidification. For ductile iron castings, the evolution of the solid fraction with temperature must be modeled correctly because the shape of the solidus and liquidus curves affects the feeding behavior. The finite element solver gives me the temperature history at every node. From this history I can obtain the solidification time, the temperature gradient, and the cooling rate at each location.
The important advantage of this software is that I do not need to calculate the modulus manually for ductile iron castings. The solidification time distribution is a direct measure of the modulus, but it is calculated by solving the full heat-transfer problem. Therefore, it automatically accounts for geometrical complexity, heat accumulation, and latent heat effects. After the first simulation run, I can examine the shrinkage defect distribution and identify the zones that require feeding. Then the riser module uses the solidification time and the volume of the casting zone to be fed in order to propose a riser neck size and the number of risers.
The complete workflow that I followed for the valve body in ductile iron castings is shown in the following table.
| Step | Action | Goal |
|---|---|---|
| 1 | Create a three-dimensional CAD model of the valve body | Represent the exact geometry |
| 2 | Mesh the model and assign thermal properties | Prepare the finite element model |
| 3 | Run solidification simulation without risers | Identify hot spots and shrinkage defects |
| 4 | Use the riser module with solidification time and volume | Determine initial riser and neck sizes |
| 5 | Add the riser system and re-simulate | Check whether defects are eliminated |
| 6 | Modify riser position if necessary | Ensure an open feeding path |
| 7 | Design a multiple-cavity gating system | Fill all cavities uniformly |
| 8 | Use the sprue as a feeder where possible | Reduce riser volume and increase yield |
| 9 | Run final filling and solidification simulation | Verify that ductile iron castings are sound |
Initial Solidification Simulation of a Valve Body
The component that I selected for this study is a valve body produced in ductile iron castings. This part has several flanges of different thicknesses, which makes it a challenging geometry for feeding system design. The thickest sections tend to become hot spots, while the thinner sections tend to solidify first and block the flow of feed metal. The image below shows the valve body geometry that I used in this work.
Before simulating the solidification of the valve body, I assigned thermal boundary conditions that are typical for ductile iron castings in sand molds. I used a pouring temperature of 1380 °C, an initial mold temperature of 25 °C, and a heat-transfer coefficient at the metal–mold interface that was updated by the software during the filling and solidification stages. The table below lists the main simulation parameters.
| Parameter | Value | Unit |
|---|---|---|
| Pouring temperature | 1380 | °C |
| Liquidus temperature | 1150 | °C |
| Solidus temperature | 1060 | °C |
| Latent heat of fusion | 220 | kJ/kg |
| Initial mold temperature | 25 | °C |
| Critical solid fraction for feeding | 0.65 | — |
| Mold material | Resin-bonded sand | — |
I first ran the solidification simulation without any riser. This was an important step because it allowed me to see where shrinkage would naturally appear in ductile iron castings when no external feed metal is supplied. The simulation predicted a large amount of shrinkage porosity in the upper part of the valve body. This was not unexpected. The upper thick section was isolated by a thin flange, and the thin flange solidified before the thick section. As a result, the contracting liquid inside the thick section could no longer be fed. The defect distribution obtained from the simulation agreed with the kind of shrinkage that I have seen in actual valve body production.
In addition to the defect distribution, the software gave me a solidification time contour plot. This plot was very useful because it showed the locations where the metal remained liquid for the longest time. For ductile iron castings, those locations are the natural centers of feeding demand. I used the solidification time distribution to understand which areas had to be connected to a riser or to an internal feeder. The combination of the defect plot and the solidification time plot is much more informative than a simple modulus calculation because it shows the entire solidification sequence in space.
To evaluate the tendency for microporosity, I also studied the local cooling conditions. One common parameter used by foundry engineers is the Niyama criterion, which is defined as
$$N = \frac{G}{\sqrt{\dot T}}$$
where G is the local temperature gradient and ˙T is the local cooling rate. A low value of N indicates that the solidification front is flat and that liquid flow through the dendrite network is difficult. For ductile iron castings, the Niyama criterion must be interpreted carefully, but it is still useful for comparing different feeding system designs. In the initial simulation, the upper hot spot had a low Niyama value, which confirmed that shrinkage porosity was likely.
Riser Design for Ductile Iron Castings
After the initial solidification simulation, I turned my attention to the riser design. The principle of riser design for ductile iron castings is to ensure that liquid metal can flow smoothly to the location that needs feeding without producing a harmful thermal disturbance in the casting. A good riser must remain liquid until the casting zone has finished shrinking. In addition, the riser neck must stay open long enough to allow that liquid to enter the casting. Finally, the riser volume must be large enough to provide the required feed metal.
In the traditional approach, the riser size is often determined from the modulus of the casting section being fed. For example, a riser modulus larger than the casting modulus is selected, and then the riser dimensions are calculated. The traditional criterion can be written as
$$M_r \geq f_d M_c$$
where Mr is the modulus of the riser, Mc is the modulus of the casting section, and fd is a safety factor that depends on the alloy. For ductile iron castings, this factor can vary because graphite expansion changes the volume demand. The Experto-ViewCast software does not force me to calculate the modulus manually. Instead, it uses the solidification time, which is the physical quantity that the modulus is intended to represent. I only need to provide the solidification time and volume of the region to be fed, together with the riser type. The software then calculates the riser neck size and the number of risers.
The design of the riser neck is as important as the design of the riser body. The neck must not solidify too early. The solidification time of the neck can be estimated from its volume and cooling surface area:
$$t_{neck} = K \left( \frac{V_{neck}}{A_{neck}} \right)^2$$
In the feeding of ductile iron castings, I require the neck solidification time to be greater than the time required to complete feeding. If the neck is too small, it freezes first and blocks the flow of liquid metal, even if the riser itself contains sufficient liquid. If the neck is too large, it behaves like part of the casting and delays solidification, which may create an undesirable hot spot. Therefore, the neck size must be a compromise between thermal condition and feeding flow.
The feeding volume of a riser should satisfy the following relation for ductile iron castings:
$$V_r \geq \frac{\beta_{eff} V_c}{\eta}$$
where Vr is the volume of the riser, βeff is the effective volumetric contraction coefficient, Vc is the volume of the casting zone to be fed, and η is the efficiency of the riser. For ductile iron castings, \(\beta_{eff}\) may be small provided that graphitic expansion is active. In my first design, the software recommended a cylindrical open riser near the suspected defect zone. The initial riser parameters are given in the table below.
| Riser Parameter | Initial Value |
|---|---|
| Riser type | Cylindrical open riser |
| Riser diameter | 58 mm |
| Riser height | 76 mm |
| Riser neck diameter | 24 mm |
| Riser neck height | 8 mm |
| Number of risers | 1 |
I placed this riser on the flange closest to the upper defect zone, and I repeated the solidification simulation. To my surprise, the simulation showed that the valve body still contained internal defects. This result demonstrated that the riser had sufficient volume but was not placed in a position that allowed the feed metal to reach the hot spot. The thinner flange that I chose as the mounting location was freezing too early. Therefore, the riser neck became solid before the riser could effectively feed the casting. This is a common issue in ductile iron castings: the feeding system is calculated correctly in terms of size, but the local geometry prevents the feeding path from working.
Optimization of Riser Position for Ductile Iron Castings
The first simulation with the riser showed that the casting was still defective. I examined the solidification time contours and the temperature distribution around the riser neck. The neck was connected to a relatively thin flange. The flange solidified quickly because its surface area was large and its volume was small. As a result, the neck lost heat to the surrounding sand and to the solidified flange. The feeding channel closed before the upper thick section had fully solidified. The local solidification time of the flange was shorter than the time required by the casting hot spot.
To solve this problem for the ductile iron castings in this study, I changed the riser position. I moved the riser from the thin flange to a thicker flange located at the end of the valve body. The thick flange had a larger volume and a larger thermal mass. This allowed the riser neck to remain molten for a longer time. The feed metal was able to flow through the thick flange and then into the former hot spot. The modified riser system was simulated again. This time, the defect distribution was clean. The shrinkage porosity in the upper section disappeared because the feeding path remained open until solidification was complete.
The comparison between the initial and optimized riser positions is shown in the following table.
| Case | Riser Position | Simulation Result | Interpretation |
|---|---|---|---|
| Initial design | Thin flange near the upper defect | Shrinkage remained | Neck froze before feeding was complete |
| Optimized design | Thick flange at the end of the casting | No shrinkage | Feeding path remained open during solidification |
This result taught me an important lesson about the feeding of ductile iron castings. The riser must not only be large enough; it must also be placed where the surrounding metal is hot enough to keep the neck open. The position of the riser has a direct influence on the temperature gradient in the casting. In the initial design, the thin flange acted as a thermal barrier. In the optimized design, the thick flange acted as a bridge between the riser and the hot spot. The finite element simulation was essential because this kind of thermal interaction is almost impossible to predict by hand.
Gating System Design for Multiple Ductile Iron Castings
After the feeding system for the valve body was optimized, I started the design of the gating system. In industrial production, it is often economical to cast several components in one mold. I therefore arranged four valve bodies in a single mold and connected them through a common runner system. The gating system design for ductile iron castings must control the velocity of the liquid metal, prevent turbulence and splashing, and ensure that all cavities are filled completely before solidification begins.
The flow of liquid metal in a gating system follows the principle of continuity. For an incompressible fluid, the continuity equation is
$$A_1 v_1 = A_2 v_2$$
where A is the cross-sectional area and v is the velocity. The velocity in the sprue can be estimated from Bernoulli’s equation. For a constant sprue height, the theoretical velocity is
$$v = \sqrt{2 g h_p}$$
where g is gravitational acceleration and hp is the effective pouring height. In real gating systems, friction losses and turbulence reduce the actual velocity. The actual flow rate through the choke area can be expressed as
$$Q = A_{choke} \mu \sqrt{2 g h_p}$$
where μ is a flow loss coefficient. The filling time of the mold can then be estimated as
$$t_f = \frac{V_{mold}}{Q}$$
For ductile iron castings, a short but controlled filling time is preferred. Excessive filling time allows the liquid metal to cool before it reaches the last cavity. Insufficient filling time can create high velocities and cause sand erosion or oxidation. I selected a closed-open gating system. In this system, the choke is located in the sprue base so that the sprue remains full of metal during pouring. The runner is then enlarged to reduce velocity, and the ingates are distributed uniformly among the four cavities. The area ratio that I used in this case was approximately
$$A_{sprue} : A_{runner} : A_{ingates} = 1 : 1.8 : 1.4$$
This ratio helped to keep the liquid metal flowing smoothly while maintaining the system full enough to prevent air aspiration. The ingates were placed so that the metal entered the casting tangentially, which reduced impact on the mold walls and core surfaces.
In addition to the geometrical design of the gating channels, I also considered the Reynolds number of the liquid metal flow:
$$Re = \frac{\rho v D}{\mu}$$
where ρ is density, v is velocity, D is the hydraulic diameter, and μ is dynamic viscosity. I tried to keep the local velocity in the ingates below the critical velocity because high velocities can produce oxide films and dross in ductile iron castings. The simulation allowed me to check the velocity field and make sure that no severe jetting occurred.
Using the Sprue as a Feeder for Ductile Iron Castings
While I was verifying the gating system, I noticed an interesting opportunity. The central sprue had to be large enough to supply four valve bodies. As a result, the diameter of the sprue became comparable to the diameter of the cylindrical riser that I had used in the feeding system design. This observation led me to consider whether the sprue itself could replace the conventional riser. If the sprue is designed correctly, it can provide feed metal through the runner and ingates to the casting cavities. This concept is sometimes called a sprue feeder or a runner feeder. For ductile iron castings, the approach can be especially effective because the graphite expansion reduces the total external feed demand.
I therefore modified the gating system so that the sprue could serve as a feeder. The runner system was not merely a transport channel; it was also a feeding path. The geometry of the sprue was designed to remain molten until all four castings had completed their solidification contraction. I used the volume criterion to check the sprue feeder:
$$V_{sprue} \geq \beta_{eff} V_{total}$$
where Vtotal is the total volume of the four valve bodies that require feeding. In this case, the large central sprue satisfied this condition. The sprue was placed at the center of the mold, which meant that it had the longest solidification time because it was surrounded by the thermal mass of the runner system. The graphitic expansion in the ductile iron castings helped to reduce the required feeding volume, so the sprue volume was sufficient without additional risers.
The next step was to simulate the complete filling and solidification process with the sprue-feeding gating system. The filling simulation showed that the liquid metal advanced smoothly through the sprue, into the runners, and then into the four valve body cavities. The temperature field during filling was stable. I did not observe any cold shut or misrun. The temperature at the end of filling remained above the liquidus in the feeding channels, so the sprue was still able to feed the casting cavities after filling was complete.
The final simulation result for the sprue-feeding system showed no shrinkage porosity in the ductile iron castings. This was a very positive result because it meant that I could eliminate the separate risers entirely. The sprue performed the double function of filling and feeding. This approach increased the yield of the mold because no external riser metal was wasted.
The following table compares the initial feeding concept with the final sprue-feeding concept for ductile iron castings.
| Design Aspect | Separate Riser System | Sprue Feeder System |
|---|---|---|
| Number of castings per mold | 4 | 4 |
| External risers | 1 per casting | 0 |
| Riser volume discarded | Large | Very small |
| Feeding source | Cylindrical open riser | Central sprue |
| Defect prediction | After optimization: no defects | No defects |
| Yield | Lower | Higher |
Filling and Solidification Verification of Ductile Iron Castings
Once the gating system and sprue feeder had been set up, I ran a full coupled filling and solidification simulation. This simulation is important because filling influences the initial temperature field of the mold. If the liquid metal enters the mold at different times, the first cavities cool more than the last cavities. This temperature difference can affect the feeding behavior and the final defect distribution. For ductile iron castings, a uniform temperature field at the end of filling is desirable because it helps the graphitic expansion to work uniformly in all sections.
The filling simulation showed that the liquid metal reached all four valve body cavities almost simultaneously. The velocity distribution in the runners was below the critical value that usually causes dross. The temperature distribution remained uniform, and the minimum temperature at the far end of the mold was still high enough to avoid cold shuts. This was particularly important for the thin flanges, which cool quickly after filling.
After filling, the solidification simulation continued until the whole mold had reached a low temperature. The final defect distribution did not show any internal porosity in the ductile iron castings. I also inspected the solidification sequence by plotting temperature contours at different times. The hottest region moved from the casting cavities toward the central sprue feeder. This is exactly the sequence that I expect for sound ductile iron castings: the thin sections solidify first, the thick sections solidify next, and the largest feeder solidifies last.
One of the main advantages of the numerical simulation is that it allowed me to verify this feeding sequence before cutting any pattern. I did not have to sacrifice experimental castings. I was able to test the entire process digitally and make changes at virtually no cost. For ductile iron castings, this is especially valuable because the cost of pattern modification and machining can be very high. The simulation gave me confidence that the first production attempt would be successful.
Comparison with the Original Conventional Design
In order to evaluate the benefit of the simulation-based approach, I compared my new process with the original casting process that was previously used for the same valve body. The original process was based on the experience of foundry engineers and on trial production. In the original design, the mold contained only two valve bodies. Each valve body had two risers. The reasons for this design were related to the fear of shrinkage and the uncertainty about the feeding path. Unfortunately, the original process still produced defects in the upper flange region. The defects had to be corrected by modifying the pattern after several experimental trials.
My computer-aided process was quite different. The mold contained four valve bodies, and the central sprue was used as the only feeder. The final simulation showed no defects. This result demonstrated that a well-designed sprue system can replace traditional risers in ductile iron castings when the solidification conditions are understood. The new process not only eliminated the shrinkage defects but also increased the number of castings per mold, saving both time and material.
The foundry yield is an important economic indicator. It can be calculated as
$$Y = \frac{W_{good}}{W_{good} + W_{runner} + W_{riser}} \times 100\%$$
where Wgood is the weight of the finished castings and the denominator includes all metal poured into the mold. Because the new design has no separate risers, the value of \(W_{riser}\) becomes zero. The yield improvement is therefore significant. The table below summarizes the comparison between the two designs.
| Comparison Item | Original Conventional Design | Computer-Aided Design |
|---|---|---|
| Number of castings per mold | 2 | 4 |
| Risers per casting | 2 | 0 |
| Feeder used | External risers | Sprue feeder |
| Defect prediction | Defects appeared in trial | No defects predicted |
| Trial production needed | Yes | No |
| Metal yield | Lower | Higher |
| Development cycle | Long | Short |
The use of Experto-ViewCast allowed me to rely on solidification time instead of manual modulus calculations. The software replaced the traditional trial-and-error method with digital prototyping. For this valve body, the result was a process that was simpler, more productive, and more reliable than the original process. I believe this is a clear demonstration of the power of numerical simulation for ductile iron castings.
Practical Conclusions from This Work
From my experience with this valve body, I can draw several practical conclusions for the design of ductile iron castings. The first conclusion is that the size of the riser is not the only important factor. The position of the riser and the solidification time of the riser neck are equally critical. In my initial design, the riser was large enough, but the thin flange near the defect zone froze too early and blocked the feeding path. Moving the riser to a thicker flange solved the problem without increasing the riser volume.
The second conclusion is that solidification time can replace the traditional modulus in practical design. The finite element solver calculates the solidification time from the actual three-dimensional geometry and thermal boundary conditions. This is much more robust than trying to calculate the modulus of an irregular hot spot by hand. For ductile iron castings, the solidification-time contour is also useful for explaining why a particular feeding system is not working.
The third conclusion is that the sprue can sometimes serve as a feeder. In this case, the large sprue that was required to fill four casting cavities was also able to feed the castings during solidification. This eliminated the need for separate risers and increased the material yield. Of course, this is not possible for every casting geometry. It depends on the volume relationship between the sprue and the total feeding demand. The simulation is the safest way to determine whether a sprue feeder can be used for ductile iron castings.
The fourth conclusion is that filling simulation is an essential part of the process design. A gating system that is theoretically capable of filling the mold can still cause problems if it produces turbulent flow, cold metal, or uneven temperature fields. The filling simulation allowed me to check the velocity and temperature during pouring and to make corrections before the mold was ever made.
I also found that the concept of a closed-open gating system works well for ductile iron castings. The sprue remains full during pouring, preventing air aspiration, while the runner and ingates are designed to reduce the velocity and avoid splashing. The numerical simulation confirmed that the liquid metal filled the cavities evenly and that the temperature distribution was suitable for feeding.
Finally, I must emphasize that the objective of casting simulation is not merely to verify a design that has been created by other means. In the workflow that I have described, the simulation was used as a design tool at every stage. For the initial solidification analysis, it identified the defect locations. For the riser module, it provided the input data and the riser dimensions. For the gating system, it enabled me to test the sprue feeder and to confirm the final process. This is the future of foundry engineering, particularly for high-quality ductile iron castings.
Broader Implications for the Production of Ductile Iron Castings
The case of this valve body is not unique. Many ductile iron castings have similar challenges: thick flanges connected by thin walls, internal cores, and feeding paths that must remain open until the end of solidification. The traditional approach is to add more risers to increase the safety factor. That approach usually increases cost and reduces yield. With numerical simulation, I can reduce the number of risers or even eliminate them completely by using the gating system as a feeder. This is possible because ductile iron castings benefit from graphitic expansion, and the simulation can determine how much external feeding is truly needed.
In addition, the software enables multi-cavity mold design. When multiple ductile iron castings are placed in one mold, the interaction between adjacent castings can be analyzed. The heat from one casting can affect the solidification of its neighbor. This is difficult to predict manually, but it is captured naturally in the finite element simulation. In my design, the four valve bodies were close enough to raise the local mold temperature, which helped to keep the feeding channels open for a longer period. This thermal interaction was beneficial for soundness.
Another useful feature of the software is the ability to test different riser geometries quickly. I can compare an open riser, a blind riser, and a sprue feeder without manufacturing any pattern. For each configuration, I can observe the solidification time, the feeding path, and the final defect distribution. This allows me to optimize not only the casting quality but also the production cost. The result for this valve body was a process with no separate risers, four castings per mold, and a high yield.
In terms of production time, the simulation-based approach reduced the development schedule significantly. A traditional trial-and-error process would require several rounds of pattern modification, casting, cutting, etching, and inspection. Each round could take weeks. The simulation allowed me to perform all of those iterations on a computer in a matter of days. For ductile iron castings, this can be the difference between losing a customer and delivering on time.
The following table summarizes the main technical benefits that I observed when using Experto-ViewCast for this valve body. In the table, I compare the traditional method and the simulation-based method across several important criteria.
| Criterion | Traditional Method | Experto-ViewCast Method |
|---|---|---|
| Thermal modulus calculation | Manual and approximate | Calculated from solidification time |
| Feeding path analysis | Not available | Visible as temperature contours |
| Shrinkage prediction | After trial casting | Before making any mold |
| Riser size | Conservative, large | Optimal, sometimes zero |
| Mold design | Two castings per mold | Four castings per mold |
| First production attempt | Usually requires correction | Likely to be acceptable |
Final Remarks on the Software and the Approach
Experto-ViewCast is not just a post-processing tool. It is a complete environment in which I can design the pouring system, the feeding system, and the venting system. The riser module in the software establishes a functional relationship between the riser, the riser neck, and the casting. This relationship is based on the fundamental requirements of feeding: liquid volume, solidification time, and feeding path. Since I am given the solidification time and the volume of the region to be fed, the software can propose a riser-neck geometry and the number of risers needed. This is more direct and more useful than the classic modulus chart method for ductile iron castings.
I should also mention the importance of the finite element mesh in the accuracy of the simulation. The valve body has curved surfaces and thin flanges that need to be resolved properly. Too coarse a mesh will overestimate the solidification time of thin sections and give incorrect defect predictions. I used a locally refined mesh around the flanges, the hot spots, and the riser neck. The results from that refined mesh were consistent with the defects observed in the original production. This gave me confidence that the simulation was reliable.
Another practical aspect is the treatment of latent heat. Ductile iron castings release a large amount of latent heat during the eutectic reaction. If this latent heat is not modeled correctly, the predicted solidification time will be wrong. The software handles latent heat through the evolution of solid fraction as a function of temperature. In my model, I used the lever rule for the austenite liquid region and a non-equilibrium model for the eutectic region. This allowed the software to reproduce the characteristic plateau in the cooling curve of ductile iron castings.
The filling module also contributed to the design of a reliable pouring system. The module calculates the free-surface flow of the liquid metal and couples it to the thermal calculation. This is important because the temperature of the liquid metal at the end of filling is not uniform. In a large sprue-feeding system, the first metal to enter the mold is colder than the metal that remains in the sprue. This temperature gradient can influence feeding. The simulation showed that the central sprue remained hot enough to feed the final stage of solidification. If I had designed the gating system using only steady-state flow calculations, this information would have been missed.
The success of this design suggests that modern foundries should integrate casting simulation into their daily engineering workflow. For ductile iron castings, the cost of simulation is low compared to the cost of experimental trial production. One rejected mold and one rejected heat treatment cycle can cost more than the license and personnel time of the simulation software. In my own work, the use of Experto-ViewCast has allowed me to design processes for ductile iron castings with far more confidence and speed than was possible with traditional methods.
I also believe that the simulation approach helps to preserve and transmit knowledge. Experienced foundry engineers sometimes have difficulty explaining why a certain riser position works or why a particular gating ratio is used. The simulation makes those explanations visible through temperature fields, velocity vectors, and defect distributions. This is especially useful for training young engineers who are entering the field of ductile iron castings. They can learn not only what works but also why it works.
In conclusion, the methods described in this article demonstrate that numerical simulation can replace expensive trial-and-error in the development of ductile iron castings. I used Experto-ViewCast to analyze solidification, design a riser, diagnose a feeding failure, move the riser to a better position, design a multi-cavity gating system, and finally replace the riser with the sprue feeder. The final process was simpler than the original process, more efficient in material usage, and more reliable in terms of soundness. I am convinced that this integrated approach will become the standard method for designing ductile iron castings in modern foundries.