In July 2024, extraction of a small volume of bitumen from a heated and insulated storage tank at a UK port facility resulted in partial internal vacuum and significant shape distortion of the tank shell when the tank in-breather failed to operate as designed.
Following the quick actions of the local operating team, manual venting of the tank relieved the vacuum conditions and seemingly allowed the tank to return to its original shape. Integrity concerns were raised regarding localized permanent shell distortion, residual plastic deformation, and potential cracking, all requiring extensive insulation removal or complete draining and thorough internal cleaning to investigate.
We proposed an analytical screening approach using finite element analysis (FEA) to evaluate the likelihood of permanent plastic strain under the suspected vacuum conditions as well as identification of potential damage locations for focused inspection.
The absence of visible distortion on the external insulation cladding inhibited visual identification of potential areas of plastic deformation.
This necessitated an extensive insulation removal campaign to achieve high inspection coverage to confirm the absence of any detrimental plastic deformation.
Alternatively, the client could have completely drained and thoroughly cleaned the tank for safe access in support of visual inspection by a human or drone. Both inspection approaches would have required significant down time, incurred substantial preparation and inspection costs, and generated large quantities of waste products with a significant environmental impact.
We proposed the use of a simplified finite element model of the tank to simulate the prevailing conditions during the event, which was then used to determine the likelihood of plastic deformation occurring during the vacuum event.
The associated buckling analysis completed on the model was expected to indicate the most-likely location of plastic deformation to allow for targeted insulation removal and inspection, resulting in minimized preparation, inspection, and asset down-time.
A simplified and representative 3-dimensional finite element model of the tank was created from the as-built drawings, using the specified (and varying) design thicknesses for individual shell strakes.
As there was product in the tank at the time of the event, the floor was considered a rigid surface, while the roof and support structure were modelled in detail to ensure the stiffness they provided to the tank shell was represented accurately. Submerged internal accessories in the bottom of the tank, such as heating coils, were not expected to affect the buckling behavior of the tank and so were omitted, while appropriate additional factors for external walkways and other assets were added to the model.
The team performed buckling analysis and linear-elastic stress analysis on the tank model for various loading scenarios, following the general guidance of ASME VIII Div 2 Part 5, but without the application of design load factors, to simulate actual in-service conditions.
For each assessed bitumen fill height, the internal vapor space pressure was reduced until the load multiplier in the buckling assessment reduced to a value of one, indicating the onset of buckling instability. Subsequently, engineers used linear-elastic analysis at the onset of buckling point to check for maximum stress levels exceeding the material yield strength, if any, indicating the onset of permanent plastic deformation.
Unfortunately, due to the short duration of the vacuum event, key operational data such as the bitumen fill height and vapor space pressure were not accurately recorded. The team was able to perform sensitivity analysis to address uncertainties by varying key operational data to capture all possible scenarios.
Although the buckling analysis indicated the tank had very limited capacity to withstand partial vacuum, the partial vacuum at which bucking instability occurs was shown to be orders of magnitude greater than the original design requirements, confirming adequacy of the original design.
The sensitivity analysis provided a range of maximum allowable vacuum pressure conditions for different tank fill heights, with the highest vacuum pressure associated with the lowest fill height and largest vapor space, as expected.
In contrast, by applying Boyle’s law, it was demonstrated that the risk of initiating larger vacuum conditions was greatest at higher fill heights where smaller changes in bitumen volume would result in greater increases in vacuum conditions.
All assessment scenarios demonstrated that the stress levels in the tank shell at the onset of buckling instability were well below the material yield strength. It was also demonstrated that substantial increases in negative pressure, well beyond the point of buckling instability, would be required to initiate plastic deformation, suggesting catastrophic collapse of the tank would occur well before the onset of significant permanent deformation. Furthermore, it was shown that the first plastic deformation in such an (unlikely) scenario would occur at the tank-to-roof interface, rather than in any of the areas of observed shell distortion.
Our innovative analytical approach provided quantification of the likelihood of permanent damage to the tank from the vacuum event, which enabled the client to proceed with a risk-based decision to return the tank to service following a monitored and controlled initial filling.
The rapid turnaround time of the analytical approach, completed in a matter of days, reduced the tank’s down-time to the absolute minimum. Furthermore, elimination of extensive inspection requirements resulted in a substantial direct cost saving as well as a significant reduction in environmental impact with zero insulation waste, disposed product, or contaminated water.
Contact Us
Our team of experts is ready to help you get started.
Oceaneering offers best-in-class products and solutions, backed by unparalleled quality, reliability, and global support. Fill out the inquiry form to start a conversation today.