The failed unit was an ANSUL RED LINE hand-portable dry chemical extinguisher, a welded three-piece steel pressure vessel consisting of a cylindrical shell, a top closure, and a bottom closure. According to the manufacturer’s specification, it operates at a nominal pressure of 200 psi (13.7 bar), is hydrostatically tested at 600 psi (41.4 bar), and is specified to withstand no less than six times the operating pressure. During service, the extinguisher was frequently placed directly on floor surfaces rather than wall-mounted, no protective rubber pad was fitted at the bottom, and periodic maintenance was limited to external repainting. Because the instantaneous internal pressure at the moment of failure was not recorded, the analysis did not attempt to reconstruct the exact failure pressure.

Wall-thickness measurements revealed the scale of the damage. Using ultrasonic thickness gauging where surfaces allowed reliable coupling, and image-based cross-sectional measurement or microscopy-based measurement where severe corrosion prevented it, the team mapped thickness around the bottom circumference in four regions with three measurement points each. The new extinguisher of the same model showed a uniform shell thickness of about 2.40 mm and a bottom thickness of about 5.50 mm. The failed cylinder, by contrast, retained shell thicknesses as low as 0.22 to 0.26 mm in the B–C region and 0.30, 0.18, and approximately 0.05 mm at three points in the C–D region. The minimum remaining ligament of roughly 0.05 mm, measured about 10 mm from the rupture edge, corresponds to only about 2.1 percent of the nominal thickness — a local wall loss of roughly 97.9 percent.

The thinning was strikingly non-uniform. In the A–B and D–A regions, minimum thicknesses remained substantially higher, between roughly 1.10 and 2.06 mm, showing that corrosion was strongly localized rather than spread evenly around the circumference. This pronounced gradient meant the vessel could appear externally acceptable while its pressure boundary had been critically undermined in a narrow band. The authors emphasize that localized corrosion can progressively reduce wall thickness at the bottom section without producing obvious external deformation, which is precisely why visual inspection alone may not provide a reliable indication of structural integrity.

Metallographic examination of five sections extracted at the same nominal offset of about 5 mm from the rupture edge told a consistent story. The base material retained its normal ferrite–pearlite microstructure throughout service, and no metallurgical abnormalities or manufacturing-related discontinuities that could independently explain the failure were identified in the examined regions, although the authors note the observations do not exclude undetected defects outside the sampled areas. Degradation extended inward from the external surface, with irregular penetration fronts, corrosion cavities, and remnants of corrosion products — evidence of gradual material removal during service rather than sudden fracture. At the weld-transition region, damage was more extensive than in adjacent base metal, and in several locations the original weld profile became difficult to distinguish because of advanced corrosion-assisted material loss, an interpretation the authors attribute to long-term degradation rather than an original welding defect.

Scanning electron microscopy of the shell-bottom transition revealed a rough, porous, stratified corrosion morphology with layered products extending through a substantial portion of the wall thickness, and near-zero residual thickness in places at the fracture surface. No evidence of significant plastic deformation such as necking or extensive plastic flow was observed, supporting the interpretation that sudden global mechanical overload was not the primary mechanism. X-ray diffraction of corrosion products collected from the severely corroded region identified goethite [FeO(OH)] together with α-Fe (ferrite) from the substrate, using ICDD reference patterns and indexing only experimentally observed peaks. The authors caution that XRD phase identification alone does not establish the initiation site or cause of the corrosion, but the crystalline iron oxyhydroxide is consistent with environmental exposure of carbon steel under humid, marine conditions.

Chemical composition analysis by optical emission spectroscopy showed carbon, manganese, phosphorus, and sulfur contents of 0.09, 0.37, 0.004, and 0.007 weight percent respectively, all within the specified limits for EN 10130 grade DC03. The authors are careful to stress that DC03 was adopted as an engineering-based reference material because its drawing-quality characteristics are relevant to the forming requirements of the shell, not as a confirmed identification of the steel grade. Micro-Vickers hardness testing at five locations across the shell–weld–bottom section, with five indentations per location, produced mean values from 141.2 to 222.4 HV0.2, with the highest values in the weld-affected region consistent with normal welding thermal effects and no hardness anomalies suggesting abnormal hardening contributed to the failure.

The structural assessment used classical thin-walled cylindrical-shell relationships, justified by the cylinder’s diameter-to-thickness ratio of 75, to calculate nominal hoop and longitudinal membrane stresses under the documented operating pressure of 1.38 MPa. At the nominal 2.40 mm thickness, hoop stress was about 51.75 MPa. At the minimum measured 0.05 mm, the calculated hoop stress rose to about 2484 MPa, a value the authors interpret as an indicator of lost elastic load-carrying capacity rather than a physically sustainable local stress. A first-order pressure-capacity screening estimate based on the Barlow relationship, the measured thicknesses, and an assumed reference ultimate tensile strength of 270 MPa — a standard-based input, since the actual tensile strength of the shell was not measured — showed that all measured thicknesses in the severely degraded B–C and C–D regions fell below the operating pressure. At 0.05 mm, the screening estimate dropped to roughly 0.15 MPa, only about 11 percent of the operating pressure, and the screening curve intersected the operating pressure at a model-specific thickness of about 0.46 mm.

The authors repeatedly stress the limits of these calculations. The thin-walled and Barlow relationships assume a relatively uniform cylindrical pressure boundary and do not account for localized perforations, irregular residual ligaments, weld-transition geometry, local bending, stress concentrations, welding residual stresses, material plasticity, or fracture propagation. The calculated values are therefore first-order screening indicators, not predictions of the actual rupture pressure and not necessarily conservative bounds. The principal conclusion that corrosion-induced wall thinning governed the failure rests primarily on the combined physical and metallurgical evidence, with the analytical work providing complementary quantitative context. The authors also acknowledge that because the instantaneous failure pressure was not recorded and original manufacturing records were unavailable, an abnormal pressure event or an undetected manufacturing defect cannot be conclusively excluded.

The proposed failure sequence begins with repeated ground contact mechanically damaging the external coating at the shell-bottom region, allowing moisture and oxygen to penetrate beneath the remaining protective layer and drive localized under-coating corrosion around the transition. Corrosion then propagated circumferentially, producing severe thinning and perforation holes, until the corroded band became highly susceptible to rupture under normal pressurized service, with the fracture most likely propagating along the circumferentially weakened shell–bottom transition and the stored energy releasing in rapid separation. Comparisons with a rejected extinguisher showing similar corrosion morphology but no rupture, and a new unit showing coating damage but no corrosion, support the mechanism, though the authors caution the three units had different service histories and should not be read as a directly observed chronological sequence. The practical lessons are concrete: inspection programs should combine visual examination with periodic ultrasonic thickness measurements around the lower circumferential region, prioritize minimum locally measured thickness over averages, repair damaged coatings promptly, and consider replacing vessels with localized under-coating corrosion near welded transitions even when the remaining shell appears satisfactory.

Subject of Research: Corrosion-induced structural failure of a thin-walled fire extinguisher pressure vessel

Article Title: Structural integrity failure of a thin-walled fire extinguisher induced by localized under-coating corrosion

Article References: Permana, M. S., Widyanto, B., Rabbika, A. I., & Supriyono, T. (2026). Structural integrity failure of a thin-walled fire extinguisher induced by localized under-coating corrosion. Results in Engineering, 32, Article 113754. https://doi.org/10.1016/j.rineng.2026.113754

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113754

Keywords: fire extinguisher, under-coating corrosion, pressure vessel failure, wall thinning, coating degradation, failure analysis, marine environment, metallography, X-ray diffraction, structural integrity, shell-bottom separation, ultrasonic thickness measurement