In the development of a fuel control system for a turbofan engine, a critical component—a high-pressure fuel cavity housing—was manufactured as a shell castings from ZL105A aluminum alloy. During ground-based engine integration testing, a persistent oil leakage was observed from this specific casting after approximately 20 hours of operation under a continuous fuel pressure of 2.2 MPa. The leaking section was a machined wall with a nominal thickness of 6 mm. This failure presented a significant reliability concern, prompting a comprehensive investigation into its root cause and the development of a robust corrective action for the shell castings design and processing.
The initial step involved constructing a logical failure tree to systematically identify all potential contributors to the leakage event. The analysis branched out from the top event (“Fuel Cavity Leakage”) to intermediate events like “Excessive Operational Load,” “Design Deficiency,” “Machining Defect,” “Inspection Lapse,” and ultimately to basic physical and procedural causes. A summary of the fault tree analysis for the shell castings is presented below:
| Top / Intermediate Event | Basic Event / Finding | Conclusion |
|---|---|---|
| Fuel Cavity Leakage | ||
| Excessive Operational Load | X1: System Pressure Exceeds Design Value | Ruled Out (Pressure was at specified 2.2 MPa) |
| Design Deficiency | X3: Incorrect Material Selection X5: Inadequate Wall Strength |
Ruled Out (ZL105A was suitable; theoretical strength was sufficient) |
| Machining Defect | X6: Dimensional Non-conformance | Ruled Out (Dimensions met drawing specifications) |
| Inspection Lapse | X7: Lack of/Inadequate NDT Specification | CONFIRMED (No post-machining NDT was required) |
| Casting Defect | X14: Presence of Penetrating Micro-porosity | CONFIRMED (Root Cause) |
The analysis pinpointed two linked primary causes: the physical existence of penetrating micro-porosity (X14) and the procedural lack of a non-destructive testing (NDT) requirement to detect it after machining (X7). To confirm this, the failed housing was sectioned. The cross-section revealed interconnected clusters of shrinkage porosity that spanned nearly the entire wall thickness. Crucially, both the inner and outer surfaces of this section were machined faces.
The leakage mechanism in these shell castings is a progressive failure under sustained pressure. In the as-cast state, the outer and inner surfaces have a densified “skin.” The machining process removed this skin on both sides, exposing the subsurface pore networks. Initially, these pores were separated by thin walls of metal matrix. Under the constant service pressure of 2.2 MPa, stress concentrations at the tips of these pore networks induce micro-plastic deformation. The brittle silicon phases within the ZL105A alloy, having low ductility, are prone to forming micro-cracks under this strain. Over time, under the cyclic and sustained pressure, these micro-cracks propagate and link adjacent pores. This process can be described by a simplified model for crack extension from a pore tip, related to the stress intensity factor:
$$
K_I = Y \sigma \sqrt{\pi a}
$$
Where \(K_I\) is the mode I stress intensity factor, \(\sigma\) is the applied remote stress (from fuel pressure), \(a\) is the characteristic size of the pore or flaw, and \(Y\) is a geometry factor. When \(K_I\) exceeds the material’s fracture toughness over time (sub-critical crack growth), the flaw extends. Repeated cycling leads to the eventual interconnection of pores, forming a continuous leakage path through the wall. The pressure-driven flow through such a connected porous network can be analogized to flow through a capillary or a porous medium, where the leakage rate \(Q\) is governed by the pressure differential \(\Delta P\), the fluid viscosity \(\mu\), and the complex permeability \(k\) of the defect path:
$$
Q \propto \frac{k A \Delta P}{\mu L}
$$
Here, \(A\) and \(L\) represent the effective cross-sectional area and length of the leakage path, respectively. The “permeability” \(k\) increases dramatically as isolated pores connect.
The investigation concluded that the fundamental design flaw was specifying machined surfaces on both sides of a thin-walled, pressure-containing section of the shell castings. This practice destroyed the inherent dense casting skin, the primary barrier against leakage from internal micro-porosity. Furthermore, the absence of post-machining X-ray and fluorescent penetrant inspection meant defective castings could not be screened out.
Based on this failure analysis, a two-pronged corrective action was implemented to enhance the integrity of the shell castings:
- Structural Design Change: The critical fuel cavity wall surfaces were redesigned to remain in the as-cast condition. One side was designated a non-machined, as-cast surface, preserving its densified layer.
- Enhanced Non-Destructive Testing (NDT) Protocol: A mandatory 100% NDT regimen was instituted for the shell castings:
- X-ray Radiography: To detect and reject castings with significant internal shrinkage porosity in critical areas.
- Fluorescent Penetrant Inspection (FPI): Applied to all machined and critical as-cast surfaces to reveal surface-breaking defects like interconnected porosity.
The table below contrasts the initial and improved design and processing parameters for the shell castings:
| Parameter | Initial Design/Process | Improved Design/Process |
|---|---|---|
| Critical Wall Surfaces | Both sides machined | One side as-cast (non-machined) |
| Post-Machining NDT | Not required | Mandatory X-ray & FPI |
| Primary Sealing Layer | Removed by machining | Preserved (as-cast skin) |
| Defect Screening | None for internal porosity | Active screening for pores/defects |
The effectiveness of these modifications was rigorously validated through a series of tests on newly manufactured shell castings. The test sequence and results are summarized as follows:
| Test Sequence | Test Condition / Standard | Acceptance Criterion | Result (Improved Casting) |
|---|---|---|---|
| 1. NDT Screening | Fluorescent Penetrant Inspection | No surface defects in critical areas | PASS (No indications) |
| 2. Static Pressure Test | 7.5 MPa Hydraulic Pressure, 3 min hold | Zero visible leakage or pressure drop | PASS (No leakage) |
| 3. Pressure Cycle Endurance | 2.2 MPa to 0 MPa, 10,000 cycles | Zero leakage throughout test | PASS (No leakage) |
| 4. High-Pressure Leakage Test | 6.5 MPa Hydraulic Pressure, sustained | Measured leakage rate = 0 | PASS (Zero leakage) |
The successful passage of all validation tests, especially the sustained high-pressure and pressure cycling tests which simulated and exceeded the original failure conditions, conclusively proved the correctness of the structural and procedural improvements. The mathematical rationale for the improvement lies in significantly reducing the probability of a continuous defect path. By keeping one surface as-cast, the length \(L\) of any potential leakage path originating from internal porosity is effectively doubled, as the path must now traverse through the entire thickness from an internal pore to the opposite machined surface, rather than connecting two closely spaced machined surfaces. According to the flow equation, for the same \(\Delta P\) and fluid, increasing \(L\) drastically reduces \(Q\). Furthermore, the preserved dense skin acts as a high-strength barrier with a much higher fracture toughness, increasing the critical flaw size \(a_c\) required for failure, as derived from the fracture toughness \(K_{IC}\):
$$
a_c = \frac{1}{\pi} \left( \frac{K_{IC}}{Y \sigma} \right)^2
$$
This means the component can tolerate larger inherent defects without failure.
In conclusion, the investigation into the leaking fuel cavity shell castings underscores several critical principles for the design and manufacture of high-integrity pressure-containing aluminum castings. First, the surfaces of shell castings that contain fluid pressure should, whenever possible, be designed as non-machined, as-cast surfaces to preserve the defect-blocking densified surface layer. Second, when machining of pressure walls is unavoidable, a comprehensive NDT strategy including volumetric (X-ray) and surface (FPI) methods must be mandated to screen out components with harmful defect networks. The synergistic combination of a prudent design that works with the casting process’s characteristics and a stringent quality assurance protocol is essential for ensuring the reliability and leak-tightness of critical shell castings in demanding aerospace hydraulic and fuel systems. This case study provides a validated framework for analyzing and mitigating leakage failures in thin-walled shell castings.