In modern engine manufacturing, the crankshaft stands as a pivotal component, enduring significant bending and torsional stresses during operation. Its primary failure modes are fatigue fractures stemming from these loads. Consequently, crankshafts must exhibit high rigidity, superior fatigue strength, and excellent wear resistance. The selection of material and casting process is therefore critical. The iron mould sand coating (IMSC) process has emerged as a highly suitable method for producing crankshafts from nodular cast iron, also widely known as ductile iron. This process is characterized by rapid cooling rates and high mould rigidity. For nodular cast iron, these conditions promote the formation of pearlite, refine the microstructure, and leverage graphite expansion to achieve self-feeding, ultimately yielding castings with dense microstructures, high strength, and good toughness. However, despite these advantages, the production of high-grade nodular cast iron components like crankshafts is susceptible to certain casting defects, with slag inclusion being particularly prevalent and detrimental.
Slag inclusions are a common but serious defect in nodular cast iron castings. When present in critical stress-bearing areas such as the crankpin (connecting rod journal) or main journal of a crankshaft, these inclusions act as potential crack initiation sites. They disrupt the continuity of the metallic matrix, significantly reducing the component’s plasticity, toughness, and, most importantly, its fatigue strength. The risk is exacerbated in high-strength grades of nodular cast iron; the higher the required strength, the more pronounced the damaging effect of slag inclusions on the matrix continuity, thereby drastically increasing the risk of fatigue failure in service.
In our production facility, we employ the IMSC process to manufacture a specific model (A-type) of four-cylinder engine crankshaft from a high-grade nodular cast iron (QT850-5). During machining at our customer’s end, slag inclusion defects were discovered on the outer cylindrical surfaces of both the crankpins and main journals. The defects manifested as isolated or clustered fine points or irregular linear marks. Due to their sub-surface nature and specific locations, these defects are undetectable prior to machining and are often only revealed during post-machining magnetic particle inspection or under intense light during visual inspection. This late-stage discovery prevents timely intervention and leads to significant scrap losses and economic damage.
The defect occurrence rate was historically unstable, fluctuating around 1.2% to 2.0%. However, in April 2019, a specific production batch experienced a sudden spike in scrap rate. Statistical analysis revealed that out of 11,970 machined A-type crankshafts, 584 were scrapped due to slag inclusions, corresponding to a scrap rate of 4.9%. This incident caused substantial financial loss and severely disrupted the customer’s production line, elevating quality risks. In response, we initiated a thorough investigation. Defective samples were subjected to detailed metallographic analysis and energy-dispersive X-ray spectroscopy (EDS) at a partner institution. This report synthesizes our findings, analyzing the root causes from the perspectives of slag formation mechanisms, technical design, raw material quality, process control, and inspection protocols. We detail the systematic countermeasures implemented and their subsequent validation.
The foundry process for nodular cast iron crankshafts via IMSC can be divided into several key stages: melting and treatment, moulding, pouring, and cleaning. The stages most critically influencing slag formation are melting, nodularizing/innoculation, and pouring.
Foundry Production Process Flow
The entire manufacturing chain must be meticulously controlled to ensure the integrity of the final nodular cast iron component.
1. Melting and Treatment
This stage encompasses iron melting followed by nodularizing and inoculation treatment.
a) Iron Melting: We utilize a 1-ton medium-frequency induction furnace. The charge materials consist of pig iron specifically designed for nodular cast iron, steel scrap (low-to-medium carbon steel), returns (gates and risers), and alloying elements like copper and manganese. These are charged sequentially per the established工艺. The accuracy and purity of these charge materials directly impact the final chemical composition and melt cleanliness. Impurities inherent in the charge become direct sources of slag. Therefore, strict control is mandated over harmful elements (S, P, Ti), sand adherence on returns, foreign materials in scrap, and factors like rust and moisture on all charge materials.
b) Nodularizing and Inoculation Treatment: The treatment is performed using a covered ladle (teapot ladle) method. The reaction chamber is loaded with a compacted mixture of nodularizer (typically a Fe-Si-Mg alloy), ferrosilicon inoculant, and iron chips. The chips act as a buffer, preventing direct contact between the hot metal and the nodularizer to reduce Mg loss. Once the melt reaches the target temperature and chemistry, it is tapped into the ladle to initiate the treatment. After the reaction subsides, thorough slag skimming is performed to obtain a clean metal ready for pouring.
2. Moulding
The IMSC process involves using compressed air in a shooting machine to blow resin-coated sand (phenolic resin sand) into the cavity between a pre-heated metal pattern and the iron mould. The sand cures upon contact with the heated surfaces, forming a thin (5-8 mm) sand layer on the mould wall. After curing, the pattern is removed, leaving a precise mould cavity. The process steps include mould cleaning, sand shooting, curing, pattern stripping, mould repair, mould closing, and placement of the pouring cup.
3. Pouring
This stage involves transferring the treated nodular cast iron melt into the mould cavity.
a) Pouring: Key controlled parameters are pouring temperature and pouring speed.
b) Secondary (In-stream) Inoculation: Due to the rapid cooling inherent in the IMSC process, there is a tendency for cementite formation during solidification. A secondary inoculation, performed as a fine stream of inoculant added to the metal stream during pouring, is crucial. It increases graphite nodule count, refines graphite structure, improves nodularity, and suppresses chill. The key parameters are inoculant particle size and addition rate.
c) In-mould Cooling: The solidification rate, which governs the final microstructure, is controlled primarily by the time before shakeout.
4. Cleaning
This final step removes fins, flash, and any residual sand adhering to the casting surface.
Deep Dive into Slag Inclusion Defects
Our investigation focused on two fundamental aspects: the sources of slag and the pathways through which slag enters and remains in the casting.
2.1 Mechanisms of Slag Formation
Slag in nodular cast iron originates from multiple chemical and physical reactions. The primary chemical reactions leading to slag formation involve magnesium, a key nodularizing element:
Reaction with Sulfur: $$ \text{Mg} + \text{S} \rightarrow \text{MgS} $$ This magnesium sulfide has a low density and can form solid inclusions.
Reaction with Oxygen: $$ 2\text{Mg} + \text{O}_2 \rightarrow 2\text{MgO} $$ Magnesium oxide is also a common inclusion.
Furthermore, silicon from the inoculant can oxidize: $$ \text{Si} + \text{O}_2 \rightarrow \text{SiO}_2 $$
The total potential slag volume can be conceptually related to the excess reactive elements. A simplified mass balance for slag-forming oxides and sulfides can be represented as:
$$ m_{\text{slag}} \propto [\text{Mg}]_{\text{excess}} + [\text{Si}]_{\text{oxidized}} + [\text{Impurities}]_{\text{charge}} $$
where $[\text{Mg}]_{\text{excess}}$ is the magnesium content beyond what is needed for nodularization and sulfur neutralization.
The main identified slag sources are:
- Charge Materials: Rust (Fe2O3·nH2O) on pig iron/scrap, sand on returns, and other impurities introduce oxides and silicates into the melt.
- Oxidation during Melting: Exposure to air leads to oxidation of iron and alloying elements.
- Excess Magnesium: Over-addition of nodularizer leads to high residual Mg, which readily reacts with S and O to form MgS and MgO slag particles.
- Inoculant-related Slag: Unsuitable inoculant particle size or excessive addition rate can lead to un-dissolved or agglomerated inoculant particles, which act as exogenous inclusions.
The pathways for slag entry into the casting are:
- Ineffective Gating System: If the gating system design does not promote effective slag trapping, inclusions are carried into the mould cavity.
- In-stream Inoculant Bypass: If the secondary inoculant particle size is smaller than the mesh size of any in-gate filters, they can enter the cavity and potentially form clusters.
2.2 Root Cause Analysis of the Problem Batch
A systematic audit of our process during the problematic period revealed several non-conformances:
1. Charge Material Management: Inspection showed that the pig iron and steel scrap used had noticeable surface rust. The steel scrap contained miscellaneous impurities. Furthermore, the returns (gates and risers) from the A-type crankshaft production were being recycled without prior cleaning to remove the adherent sand coating. This directly increased the slag load in the furnace.
2. Nodularizer Addition Rate: To ensure a high nodularity grade, the practice had shifted towards an excessive addition of nodularizer. The excess residual magnesium in the melt subsequently formed sulfide and oxide inclusions during solidification. EDS analysis of the defect sites provided conclusive evidence. The spectra showed pronounced peaks for magnesium, with a calculated local concentration as high as 0.67 wt% in the inclusion area, significantly higher than the desired bulk residual level (typically 0.03-0.05%). This confirmed slag formation from excess Mg.
3. Secondary Inoculant Specification: The FeSi75 inoculant specified for in-stream addition had a nominal size range of 0.2 mm – 0.85 mm. However, sieve analysis revealed an inconsistent distribution, with about 70% of particles in the finer 0.2 mm – 0.4 mm range. These extremely fine particles are prone to oxidation in the metal stream before complete dissolution. Their low density causes them to float and accumulate at the metal-mould interface or within the solidifying metal, forming micro-inclusions.
4. Gating System Design Deficiencies: The original gating system was analyzed for its slag-trapping efficiency. Two critical design flaws were identified:
- Runner Cross-section Ratio: The total cross-sectional area of the ingates was larger than that of the sprue. In such a configuration, the flow in the runner is not pressurized, reducing its ability to force lighter slag particles to the top of the runner where they can be trapped. Instead, slag can be drawn into the ingate stream.
- Ingate Design Causing Turbulence: Conversely, if the total ingate area is too small relative to the sprue, metal velocity through the ingates becomes very high, causing jetting and splashing inside the mould cavity. Each droplet oxidizes, creating copious amounts of dross (re-oxidation products).
The original design suffered from the first issue, leading to poor slag separation in the runner.
The interaction of these factors—high slag load from charge materials, excessive Mg, fine inoculant, and an inefficient gating system—created a perfect storm resulting in the 4.9% scrap rate.
Implemented Corrective and Preventive Actions
Our countermeasures targeted both slag source reduction and pathway interruption.
3.1 Enhanced Raw and Auxiliary Material Management
We instituted stricter controls and specifications:
| Material | Previous State | New Control Measure | Target |
|---|---|---|---|
| Pig Iron & Scrap | Used with rust and impurities | Visual and periodic chemical inspection; rejection of rusty/contaminated batches. | Minimize introduction of Fe2O3 and impurities. |
| Returns (Gates/Risers) | Recycled with adherent sand | Mandatory shot blasting before charging into furnace. | Eliminate SiO2 from sand entering melt. |
| Buffer Iron Chips | Potential contamination from tool bits | Magnetic separation and screening prior to use. | Remove tungsten carbide, HSS fragments. |
| Nodularizer | Fixed over-addition based on historical practice | Addition rate dynamically adjusted based on daily residual Mg analysis and microstructural results (nodule count & shape). Target residual Mg: 0.035-0.045%. | Minimize excess Mg available for slag formation: $$ \text{Mg}_{\text{added}} = f(S_{\text{initial}}, \text{Target Mg}_{\text{residual}}) $$ |
| In-stream Inoculant (FeSi75) | Size: 0.2-0.85 mm, poorly distributed | New specification: 0.4-0.85 mm. Mandatory sieve check upon receipt. | Ensure adequate particle size for dissolution, avoid fine oxidized particles. |
3.2 Optimization of the Gating and Feeding System
We redesigned the gating system based on hydraulic principles for slag trapping. The primary goals were to reduce flow velocity, promote laminar flow, and create effective slag traps.
1. Runner Redesign:
- Increased the slope from the lower drag runner to the upper cope runner to facilitate the upward movement of slag.
- Added a flow-dampening chamber (expansion well) near the sprue base in the cope runner.
- Increased the height of the primary slag trap by 10 mm to increase the distance slag must float against the flow, enhancing capture probability.
The efficiency of a slag trap can be approximated by considering the buoyant velocity of a slag particle versus the horizontal flow velocity in the runner. A particle is trapped if it has enough time to float to the top before being carried past the trap. This is governed by the relationship:
$$ \frac{h_{\text{trap}}}{v_{\text{horizontal}}} < \frac{h_{\text{runner}}}{v_{\text{buoyant}}} $$
where $v_{\text{buoyant}} = \frac{2r^2 g (\rho_{\text{iron}} – \rho_{\text{slag}})}{9\eta}$ (Stokes’ Law for a sphere), $r$ is slag particle radius, $g$ is gravity, $\rho$ are densities, and $\eta$ is metal viscosity.
2. Two-Stage Filtration: In addition to the existing ceramic fiber filter placed at the ingate, we introduced a porous foam ceramic filter in the pouring cup. This provides primary filtration of larger slag particles and dampens the metal stream, reducing turbulence. The filtration efficiency for particles larger than the filter pore size $d_p$ is very high.
3. Ingate Geometry Optimization: This required empirical iteration. The original ingate dimensions were 13 mm (W) × 14 mm (L) × 7 mm (H), giving a cross-sectional area $A_g = 13 \times 7 = 91 \text{ mm}^2$ (assuming rectangular flow area).
First Modification: Increased to 16 mm × 17 mm × 8 mm ($A_g = 128 \text{ mm}^2$). The aim was to reduce metal velocity. However, this made the total ingate area larger than the sprue area, creating an unpressurized system that impaired slag trapping in the runner.
Second Modification (Successful): Reduced height to find the optimal balance. Final dimensions: 16.2 mm (W) × 17 mm (L) × 6 mm (H). This gave $A_g \approx 97.2 \text{ mm}^2$. The goal was to achieve a pressurized system (total ingate area < choke area, usually the sprue base) while avoiding excessive velocity. The design principle follows:
$$ \frac{A_{\text{sprue base}}}{ \sum A_{\text{ingates}} } = k $$
where $k$ is typically >1 for pressurized systems (e.g., 1.1 to 1.5) to ensure runner fullness and slag trapping.
| Design Version | Dimensions (W×L×H) mm | Cross-sectional Area (mm²) | System Pressure State | Theoretical Metal Velocity at Ingate* | Primary Slag Trapping Mechanism |
|---|---|---|---|---|---|
| Original | 13×14×7 | 91.0 | Near-balanced / Slightly unpressurized | High | Poor |
| Modification 1 | 16×17×8 | 128.0 | Clearly Unpressurized (k < 1) | Lower | Very Poor (slag entrainment) |
| Modification 2 (Final) | 16.2×17×6 | 97.2 | Pressurized (k ≈ 1.2) | Optimal | Good (full runner, buoyant separation) |
* Velocity is inversely proportional to area for a constant flow rate: $v_g = Q / A_g$.
Validation of Corrective Measures
The effectiveness of each set of actions was monitored by tracking the machining scrap rate due to slag inclusions at the customer’s end.
4.1 Validation of Material Management Improvements
After implementing the enhanced material controls (Table 1), but before major gating changes, we observed an immediate improvement. The scrap rate dropped from the peak of 4.9% to an average of 1.97% over a batch of 19,281 pieces (381 scrapped). This confirmed that source reduction was vital. However, the rate was still above the customer’s target of less than 1.5%, indicating that pathways for remaining slag still existed.
4.2 Validation of Gating System and Filtration Optimization
Implementing the redesigned runner and the two-stage filtration system yielded further improvement, but results were unstable, with scrap rates fluctuating between 0.5% and 1.8%. This variability pointed towards the sensitivity of the ingate design, leading to the focused ingate experiments.
4.3 Validation of Optimal Ingate Geometry
The two ingate modification experiments provided clear insights:
Experiment 1 (Modification 1: Larger Ingate): Produced alongside the original design in the same melt batch to isolate variables. Results:
- Original Ingates (5,520 pcs): 61 scrap pieces → Scrap Rate = 1.10%
- Modification 1 Ingates (4,617 pcs): 67 scrap pieces → Scrap Rate = 1.45%
The increased scrap rate with the larger ingate validated the theory that an unpressurized system ($\sum A_{\text{ingates}} > A_{\text{sprue}}$) compromises slag trapping in the runner. Slag particles within the “draw zone” of the ingates were carried into the mould.
Experiment 2 (Modification 2: Adjusted Height): Using the final dimensions (16.2×17×6 mm) for a batch of 8,256 pieces resulted in only 58 scrap pieces, a rate of 0.70%. This was significantly better than concurrent production using other designs.
Final Combined Solution Performance: Implementing all measures—enhanced material control, optimized runner design with two-stage filtration, and the final ingate geometry—consistently yielded scrap rates below 0.5% in subsequent high-volume production. This represents an order-of-magnitude improvement from the initial crisis level of 4.9%.
| Phase / Action Implemented | Batch Size (approx.) | Slag Inclusion Scrap Count | Scrap Rate (%) | Key Learning / Cause |
|---|---|---|---|---|
| Pre-Crisis Baseline | – | – | 1.2 – 2.0 (fluctuating) | Chronic, uncontrolled sources. |
| Crisis Batch (April 2019) | 11,970 | 584 | 4.9 | Convergence of all root causes. |
| Post Material Controls Only | 19,281 | 381 | 1.97 | Source reduction effective but insufficient. |
| Gating & Filtration + Mod1 Ingate | Varied | Varied | 0.5 – 1.8 (unstable) | Gating helps, but ingate design is critical. |
| Final Solution (All measures + Mod2 Ingate) | Ongoing Production | Consistently Low | < 0.5 (Stable) | Holistic approach addressing both source and pathway is essential for high-quality nodular cast iron. |
Conclusion
The journey to mitigate slag inclusion defects in high-grade nodular cast iron crankshafts produced via the IMSC process underscores several fundamental principles in foundry engineering.
1. Slag Source Control is Paramount: The most direct and effective strategy is to prevent slag formation at its origin. This necessitates rigorous control over all charge materials for chemical composition and physical cleanliness. For nodular cast iron production, special attention must be paid to minimizing elements that lead to stable oxide/sulfide formation and to carefully calculating the nodularizer addition to achieve the target residual magnesium without excess. The relationship $ \text{Mg}_{\text{added}} = f(S_{\text{initial}}, \text{Target Mg}_{\text{residual}})$ must be diligently applied. Once slag particles are dissolved or suspended in the melt, their complete removal before casting is exceedingly difficult.
2. Gating System Design is a Critical Enabler: A well-designed gating system is not merely a channel for metal delivery; it is the primary defense line against slag entry into the casting cavity. The design must be based on hydrodynamic principles to ensure a pressurized flow in the runners, promoting the flotation and trapping of inclusions. The use of filters, especially multi-stage filtration (e.g., foam ceramic in the cup plus filters in the system), significantly enhances melt cleanliness by physically intercepting particles and calming the flow. The empirical optimization of the ingate dimensions proved that small geometric changes can have a profound impact on the system’s slag-trapping efficiency, governed by the fundamental area ratios $A_{\text{sprue}} / \sum A_{\text{ingates}}$.
3. A Holistic, Systems-Based Approach is Essential: No single measure was sufficient to solve the problem. The high scrap rate was the result of multiple weaknesses interacting. Success was achieved only by addressing the entire chain: from raw material yard to poured casting. This includes process discipline (e.g., cleaning returns), precise treatment control (Mg, inoculation), and engineered design (gating, filtration). Continuous monitoring through methods like residual Mg analysis and EDS of defects provides vital feedback for maintaining control.
In summary, the production of premium nodular cast iron components like crankshafts demands an integrated strategy that aggressively minimizes slag sources through material science and process chemistry, while simultaneously employing mechanical and hydraulic principles in gating design to intercept any remaining inclusions. For foundries specializing in nodular cast iron, mastering this dual approach is key to achieving consistent high quality, reliability, and customer satisfaction in demanding applications. The stability of the scrap rate below 0.5% after implementing these comprehensive measures stands as a testament to the effectiveness of this systematic, root-cause-based methodology.