Low-Pressure Die Casting and Casting Defect Analysis of GIS Shells

I have investigated the low-pressure die casting process for a large thin-wall aluminum alloy shell used in gas insulated switchgear. The shell is a pressure-tight component that must contain SF6 gas, provide electrical insulation, and withstand mechanical loads during operation and testing. The alloy selected for this work is ZL114A, an Al-Si-Mg casting alloy. The main objective of my study is to understand the formation mechanisms of casting defects in the shell and to propose practical solutions that improve the qualified rate and gas-tightness of the final product. I developed a process parameter management system, performed CAD/CAE-based design and stress analysis, simulated the filling and solidification processes, optimized the process parameters by orthogonal experiments, and verified the optimized process in actual production. The results show that the optimized low-pressure die casting process significantly reduces casting defects and achieves a gas-tightness qualified rate above 98%.

Background and Significance

In recent years the power transmission industry has required increasingly reliable and compact high-voltage equipment. Gas Insulated Switchgear (GIS) has become a preferred solution because it occupies a small area, is insensitive to external environmental conditions, and offers high operational reliability. The aluminum alloy shell is one of the most important components of GIS. It must be leak-tight, strong enough to withstand internal pressure, and free from internal casting defects that could lead to gas leakage or mechanical failure. Traditionally, GIS shells were produced by sand casting. However, sand casting often results in low yield, poor surface finish, and a high incidence of casting defects such as shrinkage porosity, gas porosity, and slag inclusions. Low-pressure die casting (LPDC) has therefore become the main manufacturing route for GIS shells. In LPDC, the molten metal is pushed upward into the mold cavity under controlled gas pressure, and solidification occurs under a sustained holding pressure. This promotes directional solidification and feeding, which helps suppress casting defects.

Despite the advantages of LPDC, the production of large thin-wall shells still faces significant challenges. The wall thickness is about 12 mm, while the overall height is about 832 mm and the casting mass is about 61.7 kg. The shell has several thick flanges and local bosses that create hot spots and uneven wall thickness. These features are prone to casting defects such as shrinkage cavities, porosity, and cold shuts. In addition, the shell must pass strict non-destructive testing and gas-tightness tests. Any unacceptable casting defect can cause leakage and lead to scrap. Therefore, I focused on the following aspects: (1) standardizing the low-pressure die casting process design through a parameter management system; (2) designing and comparing two gating systems using CAD/CAE tools; (3) predicting casting defects by numerical simulation; (4) optimizing process parameters by orthogonal experiments; and (5) analyzing actual production casting defects and implementing corrective measures.

Materials and Experimental Methods

The material used in this study is ZL114A aluminum alloy. Its chemical composition is given in Table 1. The alloy contains silicon, magnesium, beryllium, and titanium, with aluminum as the balance. The magnesium and silicon form Mg2Si precipitates during heat treatment, which improves strength. Beryllium reduces oxidation and improves fluidity. Titanium refines the grain structure. The mechanical properties of ZL114A after T6 heat treatment are listed in Table 2. The thermal and physical properties used in the numerical simulation are listed in Table 3.

Table 1. Chemical composition of ZL114A alloy (wt.%)
Alloy Si Be Mg Zn Mn Ti Al
ZL114A 6.5–7.5 0.04–0.07 0.45–0.60 ≤0.10 ≤0.10 0.10–0.20 Balance
Table 2. Mechanical properties of ZL114A after T6 heat treatment
Alloy Casting method Condition Tensile strength σb (MPa) Elongation δ (%) Brinell hardness (HB)
ZL114A Metal mold, modified T6 ≥310 ≥3 ≥90
Table 3. Thermophysical properties of ZL114A used in simulation
Property Density (kg·m−3) Specific heat (kJ·kg−1·K−1) Latent heat (kJ·kg−1) Thermal conductivity (W·m−1·K−1) Solidus (K) Liquidus (K)
Value 2.7×103 0.964 29.3–1294 92.8–184 829 889

The experimental procedure included alloy melting, refining, modification, mold preheating, coating, low-pressure casting, heat treatment, mechanical testing, metallographic analysis, non-destructive testing, and gas-tightness testing. The melting process used argon gas for degassing and a universal flux for refining and modification. The mold was coated with a low-pressure casting release agent and a refractory coating. The sand cores were coated with an alcohol-based coating to reduce gas evolution and improve surface finish. The T6 heat treatment consisted of solution treatment at 535±5°C for 8 h, quenching in water at 60–80°C, aging at 155±5°C for 6 h, and air cooling. Tensile tests were performed on a universal testing machine. Hardness was measured with a Brinell hardness tester. Microstructures were observed with an optical microscope after polishing and etching with 0.5% hydrofluoric acid solution.

Process Parameter Management System

To standardize the low-pressure die casting process design and reduce the reliance on individual experience, I developed a process parameter management system using Visual Basic 6.0 and an Access database. The system consists of eight main modules: file management, alloy batching, gating system design, mold design, low-pressure casting process parameters, connection between the lifting tube and mold, heat treatment process, and casting defect analysis. The system enables rapid query, calculation, and optimization of process parameters. It also stores typical process cases and generates process cards in Excel format. This reduces design time and improves the consistency of process design.

The gating system design module includes calculations for the lifting tube outlet area, ingate area, and runner area. For a low-pressure die casting, the filling time and the metal flow rate must be controlled to avoid turbulence and gas entrapment, which are common causes of casting defects. The lifting tube outlet area can be estimated by:

$$A_{\text{lift}} = \frac{m}{\rho \, h \, t}$$

where \(m\) is the mass of the casting, \(\rho\) is the density of the molten alloy, \(h\) is the height of the metal lift, and \(t\) is the filling time. For an open gating system, the following relationship should be satisfied to ensure proper feeding and avoid aspiration:

$$A_{\text{ingate}} \ge A_{\text{runner}} \ge A_{\text{lift}}$$

The filling pressure can be calculated from the static pressure required to lift the metal to the top of the mold cavity and the additional pressure needed to overcome flow resistance:

$$P_{\text{fill}} = \rho g h + \Delta P_{\text{loss}}$$

where \(g\) is the gravitational acceleration and \(\Delta P_{\text{loss}}\) is the pressure loss in the gating system. The holding pressure is typically higher than the filling pressure to promote feeding during solidification. The system provides a pressure curve module that generates the complete low-pressure casting pressure profile, including lift, filling, boost, and holding stages. By using this system, I could quickly determine the initial process parameters for the shell casting and then refine them by numerical simulation.

CAD/CAE-Based Process Design and Stress Analysis

The isolating switch shell has a complex geometry with a four-way cylindrical body, four circular flanges, and several external bosses. The average wall thickness is about 12 mm, and the overall height is about 832 mm. The casting mass is about 61.7 kg. The main technical requirements are: no cracks, cold shuts, incomplete filling, gas porosity, sand inclusions, shrinkage cavities, or porosity; shot blasting on the surface; gas leakage rate less than 0.5% per year; working pressure 0.5 MPa; design pressure 3.35 MPa; and mechanical properties σb ≥ 250 MPa, δ ≥ 1.4%, HB ≥ 90. Because the shell is a pressure vessel, the stress distribution under internal pressure is important for identifying weak regions that should be free from casting defects. I performed a finite element analysis of the shell under uniform internal pressure. The material properties used are given in Table 4. The boundary conditions fixed the two ends of the shell, and a uniform internal pressure of 3.35 MPa was applied.

Table 4. Material properties used in stress analysis
Material Density (g·cm−3) Elastic modulus (MPa) Poisson’s ratio
ZL114A 2.7 70000 0.33

The first principal stress, second principal stress, and third principal stress were computed. The results show that the stress is relatively low over most of the shell, but stress concentrations occur at the transition fillets between the thin-wall shell and the thick flanges. These regions are also the last to solidify and are prone to casting defects such as shrinkage porosity. The maximum stress occurs at the fillets, which indicates that these areas should be strictly controlled during casting to avoid any casting defect that could reduce the pressure-bearing capacity. The stress concentration factor can be approximated by:

$$K_t = \frac{\sigma_{\max}}{\sigma_{\text{nom}}}$$

where \(\sigma_{\max}\) is the maximum local stress and \(\sigma_{\text{nom}}\) is the nominal stress. The analysis confirmed that the transition fillets are the most critical regions. Therefore, the gating system should be designed to ensure good feeding to these regions and to avoid local overheating. I also found that the thick flanges and bosses create hot spots, which can lead to casting defects if the cooling rate is not properly controlled.

Based on the stress analysis and the process requirements, I designed two gating system schemes. Scheme 1 is an “arborization” gating system with a single lifting tube and multiple ingates. The main sprue is placed inside the sand core, and the ingates are arranged to feed the thick flanges and bosses. Scheme 2 is a circular crevice step gating system with two lifting tubes. The step gating system fills the bottom flange first, then the side flanges, and finally the top flange. A blind riser is used at the top for feeding and venting. Both schemes were modeled and simulated to compare their filling and solidification behavior.

Numerical Simulation of Filling and Solidification

I used a commercial casting simulation software based on the finite element method to simulate the filling and solidification processes. The finite element mesh was created with tetrahedral elements. The material properties, interfacial heat transfer coefficients, and boundary conditions were assigned. The filling process was simulated with a pressure curve that included lift, filling, boost, and holding stages. The filling time was set to 30 s for Scheme 1 and 24 s for Scheme 2. The initial mold temperature was 200–300°C, and the pouring temperature was about 700°C.

For Scheme 1, the filling simulation shows that the molten metal rises smoothly through the lifting tube, enters the sprue, and then flows through the ingates into the mold cavity. The filling front is stable, and no splashing or turbulence is observed. The metal fills the bottom flange first, then the cylindrical body, and finally the top flange and bosses. The total filling time is about 30 s. For Scheme 2, the filling is also smooth, but the two lifting tubes require careful balancing to avoid asynchronous filling. The step gating system fills the bottom flange, then the side flanges, and finally the top. The filling time is about 24 s. In both schemes, no premature solidification is observed during filling, which indicates that the filling pressure and pouring temperature are adequate.

The solidification simulation shows the evolution of the solid fraction. For Scheme 1, the thin-wall shell solidifies first, followed by the flanges and bosses. The arborization gating system provides multiple feeding paths, so the thick sections can be fed effectively. The last solidification regions are the ingates and the sprue, which is the desired sequence. For Scheme 2, the flanges and bosses also solidify last, but the bosses on the shell surface cannot be fed directly by the ingates. As a result, shrinkage porosity is predicted in those regions. The simulation predicts casting defects such as shrinkage cavities and porosity in Scheme 2, while Scheme 1 shows a much lower risk of casting defects. The defect prediction results are consistent with the solidification sequence.

The mold temperature distribution was also analyzed. The highest mold temperature occurs at the transition fillets between the thin wall and the thick flanges. This is consistent with the stress analysis and with the observed casting defects in production. To reduce the risk of casting defects, the coating thickness at the fillets should be adjusted to increase the cooling rate, and the ingate positions should be optimized to avoid local overheating.

Optimization of Process Parameters

To optimize the low-pressure die casting process parameters, I designed an orthogonal experiment with three factors and three levels. The factors are pouring temperature, mold temperature, and filling pressure. The levels are given in Table 5. The orthogonal array L9(34) was used, and the experimental layout is shown in Table 6. The response variables included filling completeness, solidification sequence, and predicted casting defects. The best combination was selected based on the simulation results.

Table 5. Factors and levels of the orthogonal experiment
Level Pouring temperature (°C) Mold temperature (°C) Filling pressure (MPa)
1 680 220 0.050
2 700 260 0.055
3 720 300 0.060
Table 6. Orthogonal experiment layout and results
Run Pouring temperature (°C) Mold temperature (°C) Filling pressure (MPa) Predicted casting defects
1 680 220 0.050 Cold shuts, incomplete filling
2 680 260 0.055 Minor shrinkage porosity
3 680 300 0.060 Gas porosity
4 700 220 0.060 Shrinkage porosity at bosses
5 700 260 0.050 Acceptable, minor defects
6 700 300 0.055 Acceptable, no significant defects
7 720 220 0.055 Gas porosity, oxidation
8 720 260 0.060 Shrinkage porosity
9 720 300 0.050 Cold shuts, gas porosity

The orthogonal experiment shows that the optimal combination is pouring temperature 700°C, mold temperature 260°C, and filling pressure 0.055 MPa. However, the simulation results indicate that a mold temperature range of 200–300°C is acceptable when the pouring temperature is 700°C and the filling pressure is 0.055 MPa. The final optimized process parameters are listed in Table 7. The pressure curve for the optimized process includes a fast lift stage, a slow filling stage, a rapid boost stage, and a holding stage. The holding time is 120 s, which ensures sufficient feeding to compensate for solidification shrinkage and to reduce casting defects.

Table 7. Optimized low-pressure die casting process parameters
Parameter Value
Pouring temperature 700°C
Mold temperature 200–300°C
Lift pressure 0.04 MPa
Lift time 15 s
Filling pressure 0.055 MPa
Filling time 30 s
Boost pressure 0.08 MPa
Boost time 15 s
Holding time 120 s

Casting Defect Analysis and Remedies

During production, I tracked the actual casting process and analyzed the casting defects that occurred. The main casting defects observed were gas porosity, shrinkage cavities and porosity, cold shuts, slag inclusions, and cracks. Table 8 summarizes the types of casting defects, their characteristics, causes, and the corrective measures I applied. Each casting defect was analyzed in terms of its formation mechanism and its relationship to the process parameters and mold design.

Table 8. Casting defect analysis and corrective measures
Casting defect Characteristics Causes Corrective measures
Gas porosity Pebble-shaped pores after machining; rain-drop pores on surface; pinholes inside the wall Poor venting of mold cavity; wet coating; gas from resin sand core; insufficient degassing of melt Improve mold venting; ensure coating is dry; control sand core gas evolution; use argon degassing and measure hydrogen content; shorten melting time
Shrinkage cavity and porosity Internal porosity at thick flanges; porosity at wall-thickness transitions; macro shrinkage in heavy sections Insufficient feeding pressure; poor directional solidification; local overheating; improper ingate position Increase holding pressure; optimize ingate position; adjust coating thickness for balanced cooling; ensure sequential solidification; add chills if necessary
Cold shut Lack of fusion on the top wall; visible line or seam Low pouring temperature; low mold temperature; slow filling speed; poor venting Increase pouring temperature and mold temperature; optimize filling speed; improve mold venting; ensure smooth filling
Slag inclusion Sand sticking on inner wall; rough irregular inclusions visible after machining Low sand core strength; insufficient melt cleaning; oxide film entrapment Use ceramic filter; improve sand core strength; clean mold cavity with air; control filling speed to avoid sand core erosion; shot blast the lifting tube
Crack Cold crack at parting line; hot crack at boss fillet Early mold opening; thermal stress during cooling; poor collapsibility of sand core Optimize holding time; adjust ingate position to reduce local overheating; use hollow sand core with good collapsibility; control coating thickness

For gas porosity, I found that the most important factors are the venting of the mold cavity and the gas evolution from the resin sand core. The mold was equipped with vent plugs, vent pins, and V-shaped vent grooves. The sand core was provided with venting ropes. The coating on the mold and core was thoroughly dried before casting. The melt was degassed with argon and the hydrogen content was measured before pouring. These measures significantly reduced gas porosity casting defects. For shrinkage porosity, I improved the feeding by increasing the holding pressure and optimizing the ingate positions. The transition fillets were found to be critical because they are both stress concentrators and hot spots. By adjusting the coating thickness and adding local cooling, I achieved a more uniform temperature distribution and reduced shrinkage porosity casting defects. Cold shuts were eliminated by increasing the pouring temperature and mold temperature and by optimizing the filling speed. Slag inclusions were controlled by using a ceramic filter and by improving the sand core strength. Cracks were prevented by optimizing the holding time and by improving the collapsibility of the sand core.

The analysis of casting defects showed that most defects are related to the filling and solidification processes. Turbulent filling can entrap gas and oxide films, leading to gas porosity and slag inclusions. Improper feeding during solidification leads to shrinkage cavities and porosity. Uneven cooling leads to thermal stresses and cracks. Therefore, the key to reducing casting defects is to achieve smooth filling and directional solidification. The optimized process parameters and the improved mold design achieved this goal. In particular, the arborization gating system with multiple ingates provided good feeding to the thick sections, and the holding pressure of 0.08 MPa ensured that the casting solidified under pressure, which suppressed shrinkage porosity casting defects.

Mechanical Properties and Quality Verification

After heat treatment, I analyzed the microstructure and mechanical properties of the shell. The T6 heat treatment refined the eutectic silicon and made it more spherical and uniformly distributed. The α-Al solid solution became more uniform, and the dendritic structure was largely eliminated. The microstructure of the as-cast and T6-treated samples is shown in the micrographs. The mechanical properties were measured on specimens cut from the thick sections of the shell. The results are given in Table 9. The tensile strength reached 308–311 MPa, the elongation was 2.4–2.6%, and the Brinell hardness was 96–99 HB. These values exceed the required minimum values of σb ≥ 250 MPa, δ ≥ 1.4%, and HB ≥ 90. The mechanical properties confirm that the optimized low-pressure die casting process produces a sound casting with a low density of casting defects.

Table 9. Mechanical properties of the shell after T6 heat treatment
Property Required value Measured values
Tensile strength (MPa) ≥250 309, 311, 308
Elongation (%) ≥1.4 2.5, 2.6, 2.4
Brinell hardness (HB) ≥90 98, 99, 96

Non-destructive testing was performed using X-ray radiography. The results showed no significant internal casting defects in the shell. The gas-tightness of the shell was tested with helium leak detection. The shell was filled with helium at a certain pressure, and the helium concentration around the shell was measured with a mass spectrometer. The gas-tightness qualified rate reached 98% or higher. The pressure test was performed by filling the shell with water and increasing the pressure to 1.0 MPa, then to 3.35 MPa and holding for 10 minutes. No leakage or deformation was observed. These tests confirm that the shell meets the requirements of GIS equipment. The combination of optimized process parameters, improved mold design, and careful casting defect control resulted in a high-quality product with a low rejection rate.

Discussion

The results of this study demonstrate that casting defects in low-pressure die cast GIS shells can be significantly reduced by a systematic approach that integrates process design, numerical simulation, and production verification. The process parameter management system provided a standardized method for calculating the gating system dimensions and pressure curve. The CAD/CAE-based stress analysis identified the critical regions that are susceptible to casting defects. The numerical simulation of filling and solidification allowed me to compare two gating systems and predict the locations of casting defects. The orthogonal experiment optimized the pouring temperature, mold temperature, and filling pressure. The production trial confirmed that the optimized parameters produce a sound casting with mechanical properties exceeding the requirements.

The casting defects observed in production were mainly gas porosity, shrinkage porosity, cold shuts, slag inclusions, and cracks. Each casting defect has a specific formation mechanism. Gas porosity is caused by gas entrapment during filling or by gas evolution from the mold and core. Shrinkage porosity is caused by insufficient feeding during solidification. Cold shuts occur when the molten metal streams meet but do not fuse because of low temperature. Slag inclusions are caused by oxide films or sand particles entrained in the melt. Cracks are caused by thermal stresses during cooling or by early mold opening. By addressing each casting defect with specific corrective measures, I was able to improve the qualified rate and the reliability of the shells.

The optimized process parameters are pouring temperature 700°C, mold temperature 200–300°C, lift pressure 0.04 MPa, lift time 15 s, filling pressure 0.055 MPa, filling time 30 s, boost pressure 0.08 MPa, boost time 15 s, and holding time 120 s. These parameters produce a smooth filling process and a directional solidification sequence. The arborization gating system with multiple ingates provides good feeding to the thick flanges and bosses. The mold design includes proper venting and cooling measures to reduce casting defects. The T6 heat treatment improves the microstructure and mechanical properties. The combination of these measures results in a shell with a gas-tightness qualified rate above 98%.

Conclusions

I have studied the low-pressure die casting process for a large thin-wall aluminum alloy shell used in GIS. The following conclusions can be drawn:

1. A process parameter management system was developed to standardize the low-pressure die casting process design. The system includes modules for alloy batching, gating system design, mold design, process parameters, heat treatment, and casting defect analysis.

2. Two gating systems were designed and compared. The arborization gating system with multiple ingates produced a more uniform filling and better feeding than the circular crevice step gating system. Numerical simulation showed that the arborization gating system resulted in a lower risk of casting defects.

3. The optimized process parameters are pouring temperature 700°C, mold temperature 200–300°C, lift pressure 0.04 MPa, lift time 15 s, filling pressure 0.055 MPa, filling time 30 s, boost pressure 0.08 MPa, boost time 15 s, and holding time 120 s.

4. The main casting defects observed in production were gas porosity, shrinkage cavities and porosity, cold shuts, slag inclusions, and cracks. Each casting defect was analyzed and corrective measures were implemented, including improved venting, optimized ingate positions, adjusted coating thickness, better melt degassing, and improved sand core collapsibility.

5. After T6 heat treatment, the shell exhibited a dense microstructure and mechanical properties exceeding the requirements: tensile strength ≥250 MPa, elongation ≥1.4%, and Brinell hardness ≥90 HB. Non-destructive testing and gas-tightness testing showed no significant casting defects, and the gas-tightness qualified rate reached 98% or higher.

6. The systematic approach of combining process parameter management, CAD/CAE analysis, numerical simulation, orthogonal optimization, and production verification effectively reduced casting defects and improved the quality and reliability of GIS shells. This approach can be applied to other similar thin-wall pressure-tight castings.

The findings of this work provide a practical route for reducing casting defects in low-pressure die casting of GIS shells. By controlling the filling and solidification processes and by applying the appropriate corrective measures for each type of casting defect, it is possible to achieve a high qualified rate and to meet the stringent requirements of high-voltage switchgear. Future work may focus on further optimizing the cooling system and on applying advanced sensors for real-time monitoring of the casting process to prevent casting defects before they occur.

Scroll to Top