Rupture Failure Analysis of HP Ethylene Cracking Furnace Tube Elbows
Literature Overview and Failure Context
This 1990 study from Dalian University of Technology remains a reference point for anyone who has to explain why a radiant coil elbow let go in service. The authors took a failed HP alloy elbow from an ethylene cracking furnace and ran a full metallurgical autopsy: optical microscopy, scanning electron microscopy, energy-dispersive spectroscopy, and mechanical testing. The conclusion is blunt and worth remembering: the elbow did not fail because the alloy was wrong. It failed because the wall thinned, and the thinning was accelerated by a cocktail of high-temperature oxidation, carburization, thermal stress from temperature gradients, and thermal fatigue. In my thirty years of running plants and shipping pipe, I have seen this exact sequence repeat itself more times than I care to count.
Metallurgical Findings and Failure Sequence
The investigators identified a clear progression of damage that I have since used as a checklist when I walk a furnace during a turnaround. The elbow wall had thinned preferentially on the extrados, and the thinning was not uniform. The damage mechanism can be summarized in the following table.
| Damage Mode | Location on Elbow | Accelerating Factor | Typical Evidence |
|---|---|---|---|
| High-temperature oxidation | Extrados, outer surface | Oxygen-rich flue gas | Oxide scale, metal loss |
| Carburization | Intrados, inner surface | Carbon-rich process gas | Carbide precipitation, embrittlement |
| Thermal fatigue | Both, cyclic zones | Start-stop cycles, temperature gradients | Surface cracking, crack networks |
| Wall thinning | Extrados | Combined oxidation plus erosion | Reduced wall thickness, local thinning |
The key metallurgical observation was that the carburized layer had formed on the inner surface, and the oxide scale on the outer surface. Between them, the remaining wall section carried the full pressure load. As the wall thinned, the local stress rose. The thermal gradient across the thinned section increased, which drove thermal fatigue cracking. Once a crack initiated, it propagated rapidly through the weakened section and the elbow ruptured.
Practical Implications for Furnace Tube Design and Operation
This failure teaches several lessons that I have carried into every furnace tube specification I have signed off on since. First, wall thickness is not just a pressure design issue. In radiant coil service, the wall must be thick enough to accommodate the expected metal loss from oxidation and carburization over the design life. If you specify a minimum wall based only on ASME B31.3 or API 530 pressure calculations, you are designing a failure. Second, the elbow geometry matters more than most engineers think. The extrados of a 90-degree elbow is a high-stress zone because the bending moment concentrates there. Combine that with the thinnest wall from the forming process, and you have a recipe for trouble. Third, the carburization resistance of HP alloy is not uniform. The cast structure of a static-cast elbow has different carburization behavior than a centrifugally cast straight pipe. I have seen elbows fail while the adjacent straight sections were still sound, purely because of this difference.
Engineering Practice Case and Reflection
I once worked on a furnace where the same pattern emerged. The client had specified HP alloy for the radiant coil, but the elbows were static cast while the straight sections were centrifugal cast. After three years of service, the straight sections showed acceptable carburization depth, but the elbows had carburized through nearly half the wall. We replaced all the elbows with centrifugally cast ones and increased the nominal wall thickness by 2 mm. The furnace ran another eight years without a single elbow failure. The lesson is simple: when you specify furnace tube components, you are not just specifying a material grade. You are specifying a manufacturing route, a wall thickness, and a service life expectation. If you do not think about how the elbow will actually degrade in service, you will pay for that oversight in unplanned downtime and a very expensive repair.
Conclusion and Study Insight
This paper is a classic because it demonstrates that failure analysis is not just about identifying the cause. It is about understanding the interaction between material, geometry, and service environment. The HP alloy did not fail because it was inferior. It failed because the elbow geometry created a preferential degradation zone, and the operating conditions accelerated the damage. For any engineer involved in furnace design, procurement, or operation, this study is a reminder to look beyond the material certificate and think about how the component will behave after years of high-temperature exposure. The wall thinning mechanism described here is still the dominant failure mode in ethylene cracking furnaces today, and the countermeasures remain the same: adequate wall thickness, proper manufacturing route selection, and rigorous inspection during turnarounds.
Zhuojin Pipe Fitting Co., Ltd