In our foundry, we produce a critical high-pressure hydraulic cylinder casting. The cylinder, with a raw casting weight of 1,350 kg and made from ZG35Cr steel, must operate at 41 MPa, withstand a load of 600 tons, and be completely leak-proof. Despite these stringent requirements, our initial production methods, based on manual molding with sodium silicate-bonded sand, resulted in an unacceptable scrap rate where approximately 10% of the castings exhibited leakage during pressure testing. This persistent issue with casting defects drove us to undertake a thorough investigation, leveraging solidification simulation to diagnose the root causes and develop a robust, optimized process. The new, simulation-validated process has since produced cylinders of consistent, high quality with zero incidents of leakage.
1. Component Specifications and the Challenge
The cylinder is a thick-walled, hollow component. The primary functional requirement is structural integrity under extreme pressure, which demands a sound, dense microstructure throughout the casting wall, particularly in geometrically complex regions. Any internal shrinkage porosity or micro-shrinkage can become a path for fluid under pressure, leading to the leakage casting defects we observed. Our goal was to eliminate these defects by ensuring directional solidification toward the risers, keeping feeding paths open until the entire casting section solidified.
2. Analysis of Initial Process Attempts and Their Defects
We experimented with two distinct casting layouts before resorting to simulation.
2.1 Process Scheme One: Open End Down with Internal Chills
The first process oriented the cylinder with its open bore facing downward for ease of molding and core placement. This created a significant thermal challenge: the two thick side walls at the closed bottom end became isolated thermal centers, or hot spots, too far from the top risers to be effectively fed. To address this, we implemented an array of welded threaded steel bars (Ø10 mm) as internal chills in these regions. The risering was designed using modulus calculations.
The modulus (Mc) of the critical section was calculated as:
$$M_c = \frac{V_c}{A_c} = \frac{33.8 \times 14}{2(14 + 33.8) – 8.4} = \frac{473.2}{87.2} = 5.4 \, \text{cm}$$
Where $V_c$ is volume and $A_c$ is the cooling surface area of the section. Following the principle $M_r = 1.2 M_c$, the riser modulus $M_r$ was determined to be 1.02 cm. Two top open risers (300 mm × 200 mm × 320 mm), with a total mass $G_r$ of 250 kg, were used.
We verified the feeding capacity using the “required vs. available” method, incorporating the casting shape quotient $Q$:
$$Q = \frac{V_c}{M_c^3} = \frac{171975}{5.4^3} \approx 1092$$
For this $Q$ value, a feeding efficiency ($\eta_r$) of 35% was deemed appropriate for a conventional riser (a safety margin below the theoretical maximum for insulated risers). The available feed metal is $\eta_r \times G_r$. The total contraction volume to be compensated is the solidification shrinkage ($\varepsilon$, ~5%) multiplied by the total mass (casting $G_c$ + riser $G_r$).
$$
\begin{aligned}
\text{Avaliable Feed Metal} &= \eta_r \times G_r = 0.35 \times 250 = 87.5 \, \text{kg} \\
\text{Required Feed Metal} &= \varepsilon \times (G_c + G_r) = 0.05 \times (1350 + 250) = 80.0 \, \text{kg}
\end{aligned}
$$
Since Available > Required, the riser theoretically had sufficient capacity. A bottom-gating system via ceramic tubes was used, switching to riser filling at the end. Despite this seemingly sound design, production was unstable, with intermittent leakage casting defects.
2.2 Process Scheme Two: Open End Up with External Chills
Suspecting issues with the internal chills or feeding distance, we inverted the casting. The open end was placed upward, requiring a core hang-up. The internal chills were removed, replaced by external chills on the outer bottom surface to promote faster cooling there. The riser size remained the same but was made “hot” by gating directly into it (top-pouring). This scheme aimed to establish better temperature gradients from the chilled bottom toward the top riser. However, results were still inconsistent, with leakage defects persisting in some castings. It was clear our empirical modifications were not reliably solving the problem. The presence of these casting defects indicated hidden flaws in the solidification sequence not captured by traditional modulus calculations.
3. Solidification Simulation: Diagnosing the Root Causes
To move beyond trial-and-error, we performed computer solidification simulations on both previous schemes. The simulations provided a clear, visual understanding of how and where the casting defects formed.
- Process One Simulation: It revealed that while the internal chills accelerated cooling in the side walls, the feeding path from the main riser to these deep, hot sections closed off prematurely due to the solidification of the thinner connecting walls. This left isolated liquid pools in the thick sections that subsequently formed shrinkage porosity—the source of leakage.
- Process Two Simulation: The simulation showed that inverting the casting created a new problem. The internal fillet radius at the closed bottom (now at the top) became the last point to solidify. Although closer to the riser, the geometry created a hot spot that could not be effectively fed once the surrounding walls froze, leading to micro-shrinkage in this critical area.
The simulations conclusively identified the specific locations and mechanisms of the casting defects: interrupted feeding channels in Scheme One and an unfed isolated hot spot in Scheme Two.
4. The Optimized Process Scheme Three
Armed with insights from the simulation, we designed a third process. The goal was to enforce a clear, uninterrupted directional solidification sequence from the farthest points of the casting toward the risers. We retained the open-end-down orientation of Scheme One for its practical benefits but made several critical modifications informed by the simulation.
| Feature | Process One | Process Two | Process Three (Optimized) |
|---|---|---|---|
| Orientation | Open end down | Open end up | Open end down |
| Chill Type/Location | Internal chills in side walls | External chills on outer bottom | External chills on side & bottom; Core chill inserts |
| Core Design | Straight | Straight | 10° Taper (feeding aid) on upper section |
| Gating System | Bottom gating | Top gating (into riser) | Step gating (middle height) |
| Core Sand | Standard silica | Standard silica | Chromite sand for side cores |
| Key Mechanism | Chill hot spots | Inverse gradient + hot riser | Forced directional solidification + open feeding path |
| Simulation Result | Shrinkage in side walls | Shrinkage in top fillet | Sound casting |
| Leakage Defect Rate | ~10% | ~10% | 0% |
4.1 Key Design Modifications
1. Core Taper (Feeding Aid): The upper section of the main bore core was made with a 10° taper. This effectively creates an artificial thickness gradient in the casting wall, ensuring the metal near the risers remains liquid longest, keeping the feeding channel open.
2. Strategic Use of Chills:
- External Chills: We removed the unreliable internal chills. Instead, external chills were placed on the outer side walls and the entire bottom face of the closed end. This rapidly extracts heat from the farthest points, initiating solidification there first.
- Core Chill Inserts: Cylindrical chills were embedded within the sand cores forming the two side holes. This aggressively cools these protruding thick sections from the inside, aligning their solidification timing with the chilled outer walls.
3. Gating for Thermal Control: The gating was changed to a step-gating system with ingates at approximately the mid-height of the casting. This deliberate choice creates a favorable thermal gradient: the metal entering first (at the bottom) begins to cool first, while the metal near the ingates and risers remains hotter longer. This “washing effect” helps maintain an open feeding path from the bottom hot spots toward the riser throughout the solidification process.
4. Advanced Core Materials: The side cores containing the chill inserts were made with chromite sand. Chromite sand has significantly higher thermal conductivity and capacity than silica sand, further enhancing the chilling effect and promoting faster solidification in these problematic regions.
4.2 Refined Pouring Practice
The pouring procedure was strictly controlled to support the designed thermal gradients:
- Ladle Holding: After tapping, the steel was held in the ladle for 5-8 minutes to allow temperature reduction and slag separation, ensuring a “quieter,” lower-temperature pour.
- Sequential Pouring: The mold was filled via the step gates until the metal level reached the base of the risers. At this point, pouring from the main ladle stopped.
- Hot Riser Topping: A smaller, pre-heated ladle was used to pour hotter metal directly and slowly into the risers until full.
- Insulation & Feeding: Riser tops were immediately covered with exothermic insulating compound. After approximately 5 minutes, a final hot metal feed was added to the risers to compensate for pipe formation.
4.3 Simulation Validation and Production Outcome
The solidification simulation for Process Three showed a perfect solidification sequence. Isothermal lines progressed uniformly from the heavily chilled bottom and side sections toward the risers. No isolated liquid pools were formed. The final shrinkage cavity was fully contained within the riser body. The predicted sound casting was confirmed in production. All subsequent cylinders passed rigorous hydrostatic testing and have performed flawlessly in service without any leakage casting defects. The process yield was calculated as:
$$
\text{Process Yield} = \frac{G_c}{G_c + G_r + G_g} = \frac{1350}{1350 + 250 + 15} \approx 83.6\%
$$
This represents a highly efficient outcome for a heavy steel casting of this complexity.
5. Conclusion and Learnings on Defect Prevention
The journey to solve the leakage problem in this cylinder underscores several critical principles in modern foundry engineering, particularly for preventing shrinkage-related casting defects.
1. The Diagnostic Power of Simulation: Computer solidification simulation transformed our approach from empirical guesswork to scientific diagnosis. It provided an unambiguous visualization of defect formation mechanisms—interrupted feeding and isolated hot spots—that were not apparent from traditional calculation methods. This allows for targeted, effective corrections at the design stage, saving enormous cost and time associated with physical trials and scrap.
2. The System Approach to Directional Solidification: Achieving a sound casting is not about a single “silver bullet” but the synergistic integration of multiple tools:
- Chills are used not just to cool a spot, but to strategically initiate solidification at the correct locations.
- Feeding Aids (Tapers) are designed to artificially create favorable thermal gradients.
- Riser Design & Insulation ensure sufficient feed metal is available and liquid until the end.
- Gating Design is leveraged for thermal management, not just mold filling.
The relationship can be summarized as a systematic enforcement of the Chvorinov’s rule gradient. The goal is to ensure that the solidification time $t_f$ for any section $x$ is less than that of the section closer to the riser:
$$t_{f}(x) < t_{f}(x+\Delta x)$$
where $t_f \propto \left(\frac{V}{A}\right)^2 = M^2$. Our modifications (chills, taper) all worked to manipulate the local modulus $M$ to create this sequence.
3. Process Control is Integral to Process Design: The most perfectly designed mold cannot produce a sound casting if the pouring practice is poor. The controlled low-temperature pour, the sequential filling, and the disciplined riser feeding and topping were essential to realizing the thermal gradients planned in the design. This holistic view—integrating design, simulation, and controlled practice—is key to robust, zero-defect production.
In conclusion, by combining fundamental casting principles with advanced simulation technology and rigorous process control, we successfully identified and eliminated the root causes of a persistent and costly leakage defect. This case demonstrates that even for complex, heavily loaded components, a systematic engineering approach can reliably prevent casting defects and achieve high quality with excellent production efficiency.