10th May 2017

Abercus presents at the IMechE evening meeting in Aberdeen
An introduction to modelling explosions using CFD, and how to use the predictions to design for blast

News_20170510_Banner

Abercus principal consultant Steve Howell presented at the IMechE (Aberdeen branch) evening meeting in May. His paper focussed on the use of computational fluid dynamics for simulating hydrocarbon explosions, and on how to use the CFD predictions to derive blast loads for structural design.

The first part of this talk introduced the benefits of using CFD to model explosions, such as the ability to capture the effects of congestion and confinement, and the influence of the ignition parameters, which make this a niche CFD application. Whilst modelling individual explosion scenarios to determine the blast loads for that scenario forms an important part of the process, it alone cannot be used to determine blast loading required for structural design, since this deterministic approach generally leads to design blast loads that are excessively conservative which cannot be realistically designed for. Instead a risk based approach is preferred, which requires an understanding of the probability of occurrence of the level of blast loading. This is the basis of the so-called probabilistic explosion method, which is outlined in NORSOK Standard Z-013 and the international standards ISO 19901-3 and ISO 13702. Since the conception of the method in the late 1990s, the industry has steadily moved towards a probabilistic approach for explosion risk.

The second part of the presentation focussed on the probabilistic approach and the generation of exceedence curves, from which the design explosion loads for the permissible frequency of occurrence are derived. The probabilistic approach involves the simulation of a large dataset of individual scenarios covering each stage in the sequence of events leading up to an explosion. It is often a substantial undertaking, and for this reason it is often performed by a CFD team that is entirely separate from the structural design team that will eventually use the design explosion loads derived from the modelling. In Abercus’ experience, the interface between the two parties is often primitive, and as a consequence it is usually necessary to reduce the huge amount of predictive data generated as part of the probabilistic assessment into just a few exceedance curves describing the predicted explosion risk at selected targets of interest, so that the information can be easily transferred to the structural team.

With a little thought and imagination, the predictive data accumulated as part of the probabilistic assessment can be used to provide improved insight, which can lead to a better, safer and more cost-effective design. The key to this is simulation data management (SDM), and the concluding part of the presentation considered some novel benefits and approaches that are enabled through the effective use of SDM, specifically:

  • the democratisation of the probabilistic approach
  • 3D risk assessment
  • probabilistic structural response, enabled through one-to-one coupling between the explosion CFD and structural FEA codes
  • consistency across industry.
Download the full presentation here.