As a materials engineer specializing in failure analysis, I was tasked with investigating the fracture of a QT700-2 ductile iron casting crankshaft from an automobile engine during road testing. The crankshaft, a critical component subjected to complex cyclic loads, failed at the fourth connecting rod journal. This analysis aims to identify the root cause and propose mitigation strategies. Ductile iron casting, particularly QT700-2 grade, is widely used for such applications due to its good strength and wear resistance, but its performance hinges on precise manufacturing and processing. In this report, I will detail my first-person examination using metallography, chemical analysis, residual stress measurement, and mechanical testing, emphasizing the role of material integrity in ductile iron casting components.
The failed crankshaft sample exhibited a clean fracture at the fourth connecting rod journal. Upon macroscopic inspection, I observed that the crack initiated at multiple points along the rolled fillet edge, propagating radially. The fracture surface showed signs of wear near the crack origins and the final rupture zone, but no evident impact or abrasive damage was present nearby. This suggested a fatigue-driven failure, common in ductile iron casting parts under torsional stresses. The overall geometry appeared intact, with no visible deformations elsewhere.
To delve deeper, I performed scanning electron microscopy (SEM) on the fracture surface. The crack initiation sites revealed quasi-cleavage features, indicative of brittle fracture modes under cyclic loading. No casting defects like inclusions, porosity, or pre-existing cracks were found, which often plague ductile iron casting processes. The propagation zone displayed classic fatigue striations and beach marks, confirming fatigue as the failure mechanism. Interestingly, the rolled fillet near the fracture appeared smooth at the base but retained circumferential machining marks at the edges, a potential stress concentrator. For comparison, I examined the first connecting rod journal’s fillet, which showed similar topography. This implied that the manufacturing process, including rolling and machining, might have introduced inhomogeneities critical to fatigue life in ductile iron casting.
Metallurgical analysis was crucial. I sectioned samples from both the first and fourth connecting rod journals to assess the quenched layer and microstructure. The quenched layers were uneven in width and depth; for instance, the fourth journal had unquenched zones of 5.5 mm and 4.5 mm, with a quench depth varying from 1.5 mm to 2.0 mm. Similarly, the first journal showed disparities. Such non-uniformity in heat treatment can lead to residual stress gradients and reduced fatigue resistance in ductile iron casting components. The fillet morphology was round and defect-free, but the graphite structure analysis revealed spheroidization grade 1 (≥90%) and graphite size grade 6 (3–6 mm at 100×), consistent with standard QT700-2 ductile iron casting specifications. However, the microstructure comprised ferrite and pearlite, with pearlite content ≥98%, carbide content ≈0.5%, and phosphide eutectic ≈1%. No significant deformation layer from rolling was detected superficially, suggesting the rolling process might not have induced beneficial compressive strains uniformly.
Chemical composition analysis, conducted via spectroscopy, yielded results that raised concerns. While elements like silicon, manganese, sulfur, and phosphorus were within typical ranges for ductile iron casting, the carbon content was measured at 3.24 wt%, below the standard 3.60–3.85 wt% for QT700-2. Although carbon in ductile iron casting largely exists as graphite, which may not fully combust during testing, this deviation could affect the matrix strength and fatigue behavior. Lower carbon might reduce hardness and increase ductility, but in fatigue scenarios, it could compromise load-bearing capacity. I tabulated the findings for clarity:
| Element | Standard Range | Measured Value |
|---|---|---|
| C | 3.60–3.85 | 3.24 |
| Si | 1.90–2.60 | 2.37 |
| Mn | 0.20–0.70 | 0.40 |
| S | ≤0.030 | 0.0071 |
| P | ≤0.05 | 0.026 |
Surface residual stress measurements using X-ray diffraction revealed compressive stresses at the fillets, which are generally beneficial for fatigue resistance. However, the values varied significantly: the first connecting rod journal fillet showed -407.6 ± 22.8 MPa, the fourth journal’s intact side (B) had -333.8 ± 24.7 MPa, and the fractured side (C) exhibited only -63.3 ± 8.5 MPa, likely due to stress relief post-fracture. This gradient indicates that the rolling process did not impart uniform compressive residual stress, a key factor in enhancing the fatigue life of ductile iron casting parts. The stress state can be modeled using the following relationship for residual stress ($\sigma_r$) and fatigue strength:
$$ \sigma_a = \sigma_{fat} + k \cdot \sigma_r $$
where $\sigma_a$ is the allowable stress amplitude, $\sigma_{fat}$ is the fatigue limit of the material, and $k$ is a factor dependent on geometry and loading. In ductile iron casting, achieving high compressive $\sigma_r$ is critical to mitigate crack initiation.
Mechanical properties were assessed through hardness and tensile tests. Hardness values met specifications: core hardness ranged 268–291 HB, and surface hardness was 54.5–56.0 HRC. Tensile tests on two samples gave ultimate strengths of 970 MPa and 999 MPa, yield strengths of 589 MPa and 617 MPa, and elongations of 7.5% and 8.0%, all exceeding the QT700-2 requirements (≥770 MPa, ≥480 MPa, ≥3%). This confirms that the ductile iron casting material itself had adequate baseline properties. I summarized these results:
| Sample ID | Tensile Strength (MPa) | Yield Strength (MPa) | Elongation (%) |
|---|---|---|---|
| 1# | 970 | 589 | 7.5 |
| 2# | 999 | 617 | 8.0 |
| Requirement | ≥770 | ≥480 | ≥3 |
To understand the fatigue failure, I considered the loading conditions. Crankshafts experience complex multiaxial stresses, primarily torsional and bending moments. The stress amplitude ($\sigma_a$) under torsion can be expressed as:
$$ \sigma_a = \frac{16T}{\pi d^3} $$
where $T$ is the torque and $d$ is the journal diameter. For ductile iron casting, fatigue life ($N_f$) often follows the Basquin equation:
$$ \sigma_a = \sigma_f’ (2N_f)^b $$
with $\sigma_f’$ as the fatigue strength coefficient and $b$ as the exponent. Given the multiple crack origins at the fillet edge, I inferred high stress concentrations there. The stress concentration factor ($K_t$) for a fillet in torsion depends on geometry, and if machining marks are present, they can act as micro-notches, elevating local stresses. The effective stress range ($\Delta \sigma_{eff}$) becomes:
$$ \Delta \sigma_{eff} = K_t \cdot \sigma_a $$
This likely exceeded the endurance limit of the ductile iron casting under cyclic loading, initiating cracks. Moreover, the non-uniform quenched layer and residual stress distribution exacerbated the issue, as compressive stresses were insufficient to counteract the tensile peaks during torsion.
Microstructural homogeneity plays a vital role in ductile iron casting performance. The presence of pearlite (>98%) provides strength but can reduce toughness if not balanced with ferrite. Carbides and phosphide eutectics, though within limits, may act as crack initiators under cyclic loads. I calculated the theoretical fatigue limit based on hardness using the empirical relation for ductile iron:
$$ \sigma_{fat} \approx 0.5 \cdot \text{HB (in MPa)} $$
For a hardness of 270 HB, $\sigma_{fat}$ approximates 135 MPa. However, with stress concentrators, the actual endurance drops significantly. The fracture mechanics approach using stress intensity factor range ($\Delta K$) is also relevant:
$$ \Delta K = Y \Delta \sigma \sqrt{\pi a} $$
where $Y$ is a geometry factor and $a$ is crack length. For small cracks at fillet edges, $\Delta K$ might quickly surpass the threshold $\Delta K_{th}$ for ductile iron casting, leading to rapid propagation.
In my analysis, I also evaluated the role of manufacturing processes. The rolling operation intended to induce compressive residual stress but left machining marks at fillet edges, creating stress risers. Additionally, the uneven quench layer suggests inconsistent heat treatment, which can cause microstructural gradients and residual stresses. For ductile iron casting, optimal austempering or quenching processes are essential to achieve a uniform bainitic or martensitic matrix, but here, the matrix remained pearlitic-ferritic, possibly due to suboptimal cooling rates.
To quantify the fatigue life, I estimated the number of cycles to failure using the Coffin-Manson relation for low-cycle fatigue, though high-cycle fatigue is more applicable here. The strain-life equation is:
$$ \frac{\Delta \epsilon}{2} = \frac{\sigma_f’}{E} (2N_f)^b + \epsilon_f’ (2N_f)^c $$
where $\Delta \epsilon$ is the strain range, $E$ is Young’s modulus, and $\epsilon_f’$ and $c$ are ductility parameters. For ductile iron casting, $E \approx 170$ GPa. Given the torsional loading, I converted strains to stresses. The actual road test conditions involved variable amplitudes, but for simplicity, I assumed constant amplitude torsion. The crack initiation life likely dominated, as multiple origins suggest high local stresses.
Based on my findings, I concluded that the failure resulted from multi-origin fatigue cracking under large torsional cyclic loads, initiated at the fourth connecting rod journal’s rolled fillet edge. Key contributing factors included: non-uniform heat treatment leading to inconsistent microstructures and residual stresses, machining marks acting as stress concentrators, and possibly suboptimal carbon content affecting the matrix. The ductile iron casting material met basic mechanical specs, but processing flaws undermined its fatigue resistance.
To prevent recurrence, I recommend several improvements for ductile iron casting crankshafts: First, optimize the heat treatment cycle to ensure a uniform quenched layer and microstructure, perhaps using controlled atmosphere furnaces. Second, refine the machining and rolling processes to eliminate edge marks and achieve smoother fillet transitions, reducing stress concentration. Third, implement surface enhancement techniques like shot peening or deep rolling to impose higher compressive residual stresses uniformly. The beneficial stress can be modeled as:
$$ \sigma_{peening} = -A \cdot \frac{P}{d_p} $$
where $A$ is a constant, $P$ is peening pressure, and $d_p$ is shot diameter. Fourth, tighten chemical composition controls, especially for carbon, to maintain optimal matrix properties in ductile iron casting. Regular non-destructive testing of fillet regions during manufacturing could also detect early anomalies.
In summary, this failure analysis highlights the importance of holistic quality control in ductile iron casting production. While QT700-2 offers excellent properties, its performance in dynamic applications like crankshafts depends critically on manufacturing consistency. By addressing the identified issues, manufacturers can enhance fatigue life and reliability, ensuring that ductile iron casting components meet the rigorous demands of automotive engines. Future work could involve finite element analysis to simulate stress distributions and experimental validation of improved processes.