In modern foundry practice, the quality of castings is affected by a wide variety of defects. Among them, blowholes are undoubtedly one of the most frequent and troublesome sand foundry defects. They not only reduce the effective cross-section of castings but also degrade mechanical properties and increase scrap and repair costs. A thorough understanding of the formation mechanisms and influencing factors is essential for implementing effective countermeasures. In this work, we focus on the blowhole defects occurring in a diesel engine cylinder block produced on an automated static-pressure molding line. The defects are predominantly categorized as gas-hole defects, and the analysis covers molding sand, core sand, melting, pouring, and mold design. Through systematic experiments and process optimization, the rejection rate due to blowholes was significantly reduced. The purpose of this paper is to share the technical analysis and practical measures used to control such sand foundry defects.
1. Introduction
Casting is one of the oldest manufacturing processes, and modern foundry technology has become a critical branch of materials engineering. In the production of cast iron components, gas porosity remains a major challenge. According to statistical surveys, blowholes account for about 40% of total casting rejections in many foundries. The formation of blowholes can be attributed to complex interactions among molten metal, mold sand, cores, coatings, gases, and process parameters. The American Foundry Society classifies gas defects into several types: pinholes, blowholes, gas porosity, and shrinkage-related cavities. In general, gas defects can be divided into invasive, entrapped, precipitated, and reaction-type gas holes. Each type has its own characteristic appearance and formation mechanism.
The key to solving gas-related sand foundry defects lies in understanding the source of gas, the pressure relationship, and the solidification conditions. The gas in castings may exist as solid solution, compounds, or pores. Gas sources include moisture in molding sand, binders, additives, core resins, melting raw materials, and the surrounding atmosphere. When the internal gas pressure exceeds the resistance of the molten metal, gas bubbles can form and become trapped during solidification, producing blowholes.
In this thesis, we investigate the blowhole defects that occur on the water-jacket face of a diesel engine cylinder block produced by a high-pressure molding process. The study includes a detailed characterization of the defects, theoretical analysis of the formation mechanisms, and systematic optimization of molding sand, core making, melting, and pouring processes. By applying scientific methods and controlled experiments, we successfully reduced the blowhole rejection rate from the original high level to below 1.2%, as well as decreasing the welding repair rate. The results provide practical references for controlling similar sand foundry defects in complex thin-wall cast iron components.
2. Analysis of Blowhole Defects in the Cylinder Block
2.1 Production Process Overview
The cylinder block under study is an important component of a four-cylinder diesel engine. It has a maximum external dimension of approximately 560 mm × 360 mm × 250 mm and weighs about 62 kg. The material is high-strength gray cast iron, for example HT250, with a minimum wall thickness of only 4.5 mm. The casting is produced on a static-pressure molding line with vertically divided flasks. The molding process uses high-pressure squeeze with multiple feet, achieving a mold hardness of 90–95 (above scale). The casting is made in a two-part mold, with one casting per mold. The core package consists of 11 cores, including a water-jacket core, cylinder head bolt core, push-rod chamber core, oil pump core, and four small blocking cores. All cores are manufactured by the hot-box process using furan resin sand. After assembling the cores, the mold is closed and poured. The pouring system is a semi-closed bottom-gating arrangement.
2.2 Current Defect Status
Blowhole defects have been present since the initial commissioning of the molding line. During the first phase, the total rejection rate was extremely high, sometimes exceeding 20%, with blowholes accounting for 30–50% of all rejections. After implementing early countermeasures such as adding extra vent pins and improving the pouring system, the total rejection rate decreased to about 8–10%, but blowholes still represented a large proportion. Table 1 shows the monthly blowhole statistics for a typical year.
| Month | Blowhole share of total rejection (%) | Blowhole rejection rate (%) |
|---|---|---|
| Jan | 28.4 | 1.92 |
| Feb | 31.7 | 2.35 |
| Mar | 26.5 | 1.68 |
| Apr | 24.9 | 1.57 |
| May | 33.6 | 2.28 |
| Jun | 29.1 | 1.88 |
| Jul | 22.7 | 1.36 |
| Aug | 27.4 | 1.79 |
| Sep | 24.2 | 1.51 |
| Oct | 30.1 | 2.01 |
| Nov | 26.8 | 1.74 |
| Dec | 32.5 | 2.16 |
From the trend, the blowhole defect rate was unstable, indicating that the process control was not yet robust. Therefore, a deeper and more systematic analysis was necessary.
2.3 Theoretical Analysis and Feature Classification
Through long-period observation, more than 90% of the blowholes were found on the water-jacket face of the cylinder block, and most of them became visible only after machining. According to their visual characteristics, the defects can be divided into two categories:
- Type I: Larger blowholes, circular, flat, or pear-shaped, with smooth walls showing metallic luster or oxidized color. The pear-shaped holes have their narrow end pointing toward the water-jacket core, indicating gas invasion from that core. This type accounts for about 30–40% of the total blowholes.
- Type II: Deep irregular holes located under or near the vent pins on the water-jacket face. These holes are 8–10 mm in size, with rough walls, and the narrow end often connects with the water-jacket core. They are formed by gas from the core invading the molten metal at a late stage of solidification. This type accounts for about 60–70%.
The image below illustrates typical blowhole defects observed in such castings.
Based on the visual features, both types are classified as invasive gas pores. In Type I, gas produced by the water-jacket core before the metal surface layer solidified enters the molten metal and cannot float out because the metal has high viscosity and the casting surface has solidified quickly due to the many cooling fins on the water-jacket face. In Type II, the thick vent pins create a local hot spot on the core, allowing gas to invade the still-liquid metal while the adjacent areas have already formed a solid shell. The gas bubble growth is constrained by the dendritic network, resulting in irregular shaped pores.
2.4 Formation Mechanism
Invasive gas pores form when the gas pressure at the mold/core surface exceeds the total opposing pressure from the molten metal and the cavity. The condition can be expressed as:
$$P_g > P_l + P_c + P_r$$
where \(P_g\) is the gas pressure at the mold/core surface, \(P_l\) is the static pressure of the molten metal (\(P_l = \rho g h\)), \(P_c\) is the cavity pressure, and \(P_r\) is the resistance of the metal due to surface tension and viscosity. The static-pressure molding line produces very hard molds with high compaction, which reduces the ability of gases to escape through the sand. When the mold or core has a poor venting path, the residual gas pressure increases rapidly during pouring. If the metal is insufficient to oppose this pressure, gas bubbles penetrate into the liquid metal and become trapped.
The main influencing factors can be divided into three categories:
- Permeability of molding sand and core sand
- Gas evolution of molding sand and core sand
- Fluidity and static pressure of the molten metal
These factors were analyzed systematically through cause-and-effect diagrams, and each area was optimized in the following chapters.
3. Optimization of Molding Sand and Molding Process
3.1 Molding Line and Sand System
The static-pressure molding line uses pre-compaction by air flow, high-pressure multi-piston compaction, and counter-pressing. The mold hardness reaches 90 or more, which is favorable for dimensional accuracy but unfavorable for gas permeability. There are 32 pairs of flasks, and the sand system is a conventional loop with a water-cooled double-disc cooler, a sand mixer, and an automatic moisture control system. The sand is a mixture of Hunan sand and local sand with a fineness of 50/140 or 50/100 mesh. The sand is 100% reclaimed, and the added powder consists of sodium-activated bentonite and coal powder (or a substitute).
3.2 Control of Sand Properties
The gas evolution and permeability of molding sand are decisive factors for invasive blowhole formation. During the early production period, the total clay content of the sand was high (about 14–16%), and the wet permeability was low (around 70–80). This caused a high rate of mold explosion (“bumping”) and a high blowhole rejection rate. The wet compression strength was only about 0.08 MPa, while the moisture content was 3.5–4.2%. To improve the situation, we focused on three aspects: moisture control, bentonite quality, and the use of a coal powder substitute.
3.2.1 Moisture Control
Moisture in molding sand is the main source of gas. The first step is to stabilize the moisture content of reclaimed sand. The automatic moisture control of the old sand cooler was tuned so that the moisture of the reclaimed sand stayed around 2.0–2.5%. The controller was calibrated for ten temperature ranges; the theoretical humidity values are given in Table 2.
| Temperature (°C) | Humidity (%) |
|---|---|
| ≤50 | 2.5 |
| 50–60 | 2.4 |
| 60–70 | 2.2 |
| 70–80 | 2.0 |
| 80–90 | 1.8 |
| 90–100 | 1.6 |
| 100–110 | 1.4 |
| 110–120 | 1.2 |
| 120–130 | 1.0 |
| 130–150 | 0.8 |
In addition, the mixing process was changed from dry mixing to a wet mixing process. In the original dry process, dry powder entered the mixer first, leading to powder agglomeration and non-uniform dispersion. We tested three wet mixing schemes and compared their performance. Table 3 summarizes the measured sand properties.
| Scheme | Moisture (%) | Compactability (%) | Wet compressive strength (MPa) | Shear strength (kPa) | Permeability | Appearance |
|---|---|---|---|---|---|---|
| Scheme 1 | 3.4 | 42 | 0.085 | 18 | 110 | Some clumps, rough |
| Scheme 2 | 3.3 | 40 | 0.088 | 19 | 120 | Slight clumps |
| Scheme 3 | 3.2 | 38 | 0.095 | 21 | 135 | Fine, smooth, moderate |
| Original dry mix | 3.8 | 46 | 0.078 | 16 | 85 | Many clumps, dry |
The optimum wet mixing scheme required pre-water addition of more than two-thirds of the total amount, and the mixing time after coarse water addition should be at least 40 seconds. The final adjusted mixing cycle is shown in Table 4 for a typical batch.
| Total water (L) | Pre-water (L) | Coarse water (L) | Fine water (L) | Moisture (%) | Compactability (%) | Wet compressive strength (MPa) | Shear strength (kPa) | Permeability |
|---|---|---|---|---|---|---|---|---|
| 100 | 70 | 20 | 10 | 3.2 | 38 | 0.095 | 21 | 135 |
| 100 | 60 | 25 | 15 | 3.4 | 41 | 0.088 | 19 | 118 |
| 100 | 50 | 30 | 20 | 3.6 | 44 | 0.082 | 17 | 95 |
After optimizing the moisture and mixing process, the bentonite addition per batch was reduced from 45–50 kg to 30–35 kg, and the moisture was reduced by about 0.4%. Consequently, the gas evolution tendency decreased and the permeability increased.
3.2.2 Use of High-Quality Local Bentonite
The original bentonite was an activated sodium bentonite from Sichuan, which had a wet compressive strength of about 0.08 MPa and a methylene blue index of about 0.25 mmol/g. Due to its low bonding capacity, an excessive amount was needed, which raised the total clay content. We tested a locally produced Yunnan bentonite with better properties. The comparison is shown in Table 5.
| Property | Sichuan bentonite | Yunnan bentonite |
|---|---|---|
| Wet compressive strength of standard specimen (MPa) | 0.078–0.088 | 0.095–0.105 |
| Methylene blue index (mmol/g) | 0.24–0.27 | 0.30–0.34 |
After replacing the bentonite, the addition amount decreased from 45–50 kg per batch to 30–35 kg per batch, and the performance of both the reclaimed sand and the mixed molding sand improved. Table 6 and Table 7 show the measured values.
| Property | With Sichuan bentonite | With Yunnan bentonite |
|---|---|---|
| Clay content (%) | 15.2 | 12.4 |
| Effective bentonite (%) | 8.5 | 7.8 |
| Gas evolution (ml/g) | 28.5 | 25.1 |
| Property | Sichuan bentonite | Yunnan bentonite |
|---|---|---|
| Compactability (%) | 42 | 38 |
| Moisture (%) | 3.7 | 3.2 |
| Wet compressive strength (MPa) | 0.085 | 0.095 |
| Permeability | 95 | 135 |
3.2.3 Use of a Coal Powder Substitute
The high clay content remained a problem even with the improved bentonite. To further reduce the clay content and the gas evolution of the molding sand, we introduced a coal powder substitute, referred to as “P powder”. This material consists of coal powder, refined pitch, and starch. It provides better lustrous carbon formation, increases permeability, reduces moisture sensitivity, and improves compaction flexibility. The trial began with an addition rate of 5 kg per batch, which later was adjusted to 4 kg per batch to avoid sand burn-on. After one month of use, the molding sand properties changed as follows:
- Bentonite addition was reduced to 25 kg per batch.
- Wet permeability increased from 95–110 to 120–140.
- Mold explosion rate for the cylinder block decreased from 5–8% to less than 1%.
- Mold breakage rate decreased from 1.5% to 0.8%.
- Moisture content decreased from 3.5% to 2.9%.
Table 8 compares the key sand properties when using coal powder versus P powder.
| Parameter | Coal powder | P powder |
|---|---|---|
| Addition rate (kg/batch) | 6 | 4 |
| Wet compressive strength (MPa) | 0.088 | 0.095 |
| Shear strength (kPa) | 18 | 22 |
| Moisture (%) | 3.5 | 2.9 |
As a result, the blowhole rejection rate for the cylinder block dropped to around 1.0%, compared with 1.5–2.0% before the change. This confirms that reducing the gas evolution source is an effective measure for controlling sand foundry defects.
3.3 Improvement of Molding Process and Venting System
Although lowering the compaction pressure was initially considered to improve permeability, it caused sand expansion defects. Therefore, we focused on enhancing the venting capability of the mold. The following measures were implemented.
3.3.1 Rearrangement of Vent Pins
On the water-jacket face, three vent pins near the core print areas were removed, and three smaller vent fins were added. Additionally, four small vent pins of 6 mm diameter were added on the cylinder head top face through a bridge connection. This reduced the contact hot spots caused by thick vent pins and improved gas evacuation. As a result, Type II blowholes decreased from about 60% of all blowholes to below 20%.
3.3.2 Adding Blind Vent Holes on the Top Mold
To compensate for the low permeability of the highly compacted mold, we drilled 24 blind vent holes of 10 mm diameter from the back side of the cope. The holes did not penetrate through the mold wall; the remaining distance to the cavity surface was 5–7 mm. This distance was chosen to prevent molten metal penetration while ensuring that the holes connect to the moisture condensation zone, which otherwise acts as a gas barrier. After this modification, the mold explosion rate was further reduced, and the blowhole rejection rate improved by about 0.3%.
3.3.3 Improvement of the Water-Jacket Core Print Sealing
Liquid metal sometimes entered the venting channel of the water-jacket core print, obstructing the gas path. We added a semicircular chill groove (stopper) around the core print on both the cope and drag patterns. This significantly reduced the occurrence of metal penetration, from about 15% to less than 3%, ensuring a clear venting route for the core gases.
4. Optimization of Core Sand and Core Making Process
The cylinder block requires 11 cores, all made by the hot-box process using furan resin sand. The gas evolution from cores and the venting of core cavities are critical for controlling blowholes. The typical core sand formulation was:
- New sand: 100 parts by weight
- Furan resin: 1.0–1.2% of sand weight
- Sulfonic acid hardener: 35–45% of resin weight
- Silane coupling agent: 0.3% of resin weight
The mixing cycle included dry mixing for 40 seconds, then adding resin and mixing for 40 seconds, then adding hardener and mixing for 40 seconds, followed by discharge.
4.1 Effect of Resin Content on Gas Evolution
The gas evolution of cores increases with resin content. We tested resin additions from 1.0% to 1.4% and measured the gas evolution and the core breakage rate upon stripping. The results are given in Table 9.
| Resin content (%) | Gas evolution (ml/g) | Core breakage rate on stripping (%) |
|---|---|---|
| 1.0 | 18 | 12 |
| 1.1 | 20 | 6 |
| 1.2 | 22 | 3 |
| 1.3 | 24 | 2 |
| 1.4 | 26 | 2 |
To minimize gas evolution while maintaining acceptable core strength, we reduced the resin content from 1.2%–1.4% to 1.0%–1.1%, and increased the hardener slightly to ensure adequate surface curing. Additional measures included using weighed resin instead of volumetric measurement, and handling cores more carefully. The core scrap rate remained around 2%.
4.2 Optimization of Curing Time
The curing time and temperature determine the degree of curing of the core surface and the interior. Insufficient curing leaves unreacted resin that will generate gas during pouring. For the large water-jacket core, we varied the curing time and measured the gas evolution of the core at 5 mm below the surface. The results are shown in Table 10.
| Curing time (s) | Gas evolution (ml/g) |
|---|---|
| 80 | 32 |
| 70 | 28 |
| 60 | 25 |
| 50 | 24 |
| 40 | 26 |
| 30 | 36 |
The optimum curing time for the two-station core shooting machine was determined to be 50–60 seconds. The PLC program was locked to prevent operators from shortening the curing time for higher productivity. For the water-jacket core, the central air vent channel is formed by a retractable core rod. If the core rod was removed too early, the uncured sand collapsed and blocked the channel. By delaying the core rod removal until after at least 50 seconds of curing, the blockage rate dropped dramatically.
4.3 Coating and Drying
All cores are dipped in water-based coating and then dried in a surface dryer. The coating type was either Foseco or Guiyang water-based coating. Their properties are compared in Table 11.
| Property | Foseco coating | Guiyang coating |
|---|---|---|
| Specific gravity (Baumé) | 1.30–1.35 | 1.28–1.32 |
| Suspension index (%) | 98 | 95 |
| Gas evolution (ml/g) | 14 | 17 |
To reduce the gas contribution from the coating, we extended the drying time of the water-jacket core from 60 minutes to 90 minutes at the same temperature of 140 °C. Table 12 shows the measured gas evolution of the coated core after different drying times.
| Drying time (min) | Oven temperature (°C) | Coating specific gravity (Baumé) | Gas evolution (ml/g) | Core color |
|---|---|---|---|---|
| 60 | 140 | 1.32 | 18 | White |
| 75 | 140 | 1.32 | 14 | Light yellow |
| 90 | 140 | 1.32 | 12 | Yellow |
Additionally, strict rules were implemented: cores must be used within three days after drying; otherwise they are re-dried. The water-jacket core must exhibit a light yellow surface after drying. During dip coating, the core print openings are covered by hand to prevent coating from entering and blocking the vent channels. These measures effectively reduced the gas evolution of cores and ensured smooth venting, thereby decreasing invasive gas pore formation.
5. Optimization of Melting and Pouring Process
Melting and pouring conditions strongly influence the fluidity of the molten metal and the static pressure during filling, both of which affect the formation of invasive blowholes. The melting facility consisted of an externally water-cooled acid hot-blast cupola duplexed with a channel induction furnace. The base iron composition was adjusted in the induction furnace, and the melt was inoculated during tapping.
5.1 Chemical Composition and Carbon Equivalent
For high-strength gray iron castings, the carbon equivalent (\(CE\)) is an important measure of fluidity. The composition of the cylinder block iron is controlled as follows:
- C: 3.15–3.35%
- Si: 1.75–2.10%
- Mn: 0.7–0.9%
- P: ≤0.12%
- S: ≤0.12%
- Cu: 0.15–0.30%
The carbon equivalent is calculated using:
$$CE = w(\mathrm{C}) + \frac{w(\mathrm{Si}) + w(\mathrm{P})}{3}$$
The target \(CE\) for the cylinder block lies in the range of 3.9 to 4.0. This range provides good fluidity while maintaining the required tensile strength of 250 MPa and Brinell hardness of 190–230.
5.2 Adjusting Pouring Temperature
Lower pouring temperature increases the viscosity of the iron and reduces the ability of gas bubbles to float out. We conducted a series of trials at different pouring temperatures, keeping all other conditions unchanged. The results are presented in Table 13.
| Pouring temperature (°C) | Blocks poured | Blowhole rejections | Blowhole rejection rate (%) | Severe sand burn-on blocks | Severe burn-on rate (%) |
|---|---|---|---|---|---|
| 1340–1360 | 50 | 5 | 10.0 | 0 | 0 |
| 1370–1390 | 50 | 3 | 6.0 | 1 | 2 |
| 1400–1420 | 50 | 1 | 2.0 | 3 | 6 |
| 1430–1450 | 50 | 1 | 2.0 | 6 | 12 |
Considering the tradeoff between blowholes and sand burn-on, we selected a pouring temperature of 1400–1420 °C. To avoid the increased sand burn-on, the upper mold (cope) was sprayed with an additional alcohol-based coating at the critical position.
5.3 Improving the Gating System and Pouring Practice
The original gating system was open bottom-gated with a sprue diameter of 24 mm. The total cross-sectional areas of sprue, runner, and ingates were in the ratio:
$$\Sigma A_{\text{sprue}} : \Sigma A_{\text{runner}} : \Sigma A_{\text{ingates}} = 1 : 2.1 : 2.6$$
This caused a relatively long pouring time of about 28 seconds. To increase the filling rate and thus the static pressure head, the sprue was enlarged from 24 mm to 28 mm. The new ratio became:
$$\Sigma A_{\text{sprue}} : \Sigma A_{\text{runner}} : \Sigma A_{\text{ingates}} = 1.36 : 2.1 : 2.6$$
This semi-closed gating system reduced the pouring time to 20–23 seconds, giving an average reduction of about 5 seconds. In addition, the pouring practice was standardized: the ladle nozzle is centered over the pouring cup, the initial stream is as ample as possible without overflowing, and the pouring cup remains filled throughout the pour. This ensures a higher effective metal static pressure and minimizes gas aspiration.
5.4 Control of Sulfur Content
Sulfur is known to reduce the fluidity of gray iron because it forms manganese sulfide particles. The sulfur comes from the charge materials and from the coke in the cupola. To lower the sulfur content, we adjusted the charge composition, reduced the coke ratio to about 1:10 (coke to metal), and controlled the bed height of the coke at 1.5–1.7 m. The final sulfur content of the iron was held between 0.06% and 0.10%.
After the melting and pouring improvements, the molten iron had higher fluidity and the gas bubbles could escape more easily, resulting in a measurable decrease in blowhole defects.
6. Conclusions
This thesis systematically analyzed the blowhole defects of a diesel engine cylinder block produced on a static-pressure molding line. The following conclusions can be drawn:
- More than 90% of the blowhole defects were located on the water-jacket face. They were classified into two types according to their visual characteristics: one with smooth pear-shaped pores and the other with deep irregular pores. Both types were confirmed to be invasive gas pores, formed by gas intrusion from the mold or core into the molten metal before complete solidification.
- The key molding-sand factors affecting invasive blowholes were the gas evolution and the actual permeability of the sand. By stabilizing the moisture content of reclaimed sand, optimizing the sand mixing process from dry to wet mixing, using higher-quality bentonite, and replacing traditional coal powder with a substitute containing starch and pitch, we successfully reduced the gas evolution and increased the permeability. Additionally, rearranging vent pins, adding blind vent holes, and improving core print sealing further enhanced venting and reduced blowholes.
- For the cores, the resin content and the curing process were the main parameters influencing gas evolution. Reducing resin content from 1.2%–1.4% to 1.0%–1.1%, locking the curing time at 50–60 seconds, and optimizing the coating drying time to 90 minutes effectively reduced the gas generation and maintained reliable core venting.
- In the melting and pouring area, increasing the pouring temperature to 1400–1420 °C, enlarging the sprue to create a semi-closed gating system, ensuring a completely filled pouring cup, and controlling the sulfur content together improved the fluidity and static pressure of the iron, allowing gas bubbles to float out before the metal solidified.
- After implementing these measures over two years, the blowhole rejection rate of the cylinder block dropped from about 2.3% to less than 1.2%, and the welding repair rate also decreased substantially. The annual economic benefit was significant.
The study provides a comprehensive methodology for solving sand foundry defects of the blowhole type in complex thin-wall castings. It demonstrates that a combination of careful theoretical analysis, experimental validation, and strict process control is essential for minimizing casting defects and improving casting quality.