Multi%2DComponent+Consequence+Modelling+in+Safeti+and+Phast

Current work being performed by the risk modelling team includes improved thermodynamic modelling of mixtures for some of the key consequence models.

Print this page Save as PDF

Introduction
A future release of Safeti and Phast will include an add-on that makes significant improvements in modelling releases of mixtures. This ‘multi-component’ (MC) modelling is founded on the more rigorous calculation of mixture properties and phase equilibria as shown in Figure 1. Along with most other consequence and risk analysis packages, Safeti and Phast currently use a ‘pseudo-component’ (PC) approach for mixtures, where the thermodynamic behaviour and properties of a mixture resemble those of a pure component, whose properties are averaged from the constituent components.
Risk and consequence analyses are frequently used as the basis for decisions resulting in significant investment, and it is important that users have confidence that their results are as accurate as current technology allows. In this respect MC modelling is widely recognised as a key enhancement, and Safeti and Phast users will be among the first to be given this capability.
This is the first stage in a planned implementation over several releases, demonstrating DNV Software’s long-term commitment to provide leading-edge modelling technology for its suite of risk products. Whilst the initial range of scenarios covered will be limited, this first stage will provide improved source term estimation and constitute an essential platform for future consideration of downwind effects.

MC and PC Approaches to Modelling Mixtures
For many mixtures, especially those with wide volatility ranges, PC predictions do not match actual behaviour in key respects. For example, as illustrated in Figure 1, at a given pressure a mixture might be 2-phase across a range of temperatures, but using PC logic the bubble point and dew point temperatures for a mixture are identical. Thus, a 2-phase or 100% liquid mixture may, according to PC logic, appear to be a vapour. The use of MC modelling will mean that an analyst no longer has to adjust the starting conditions to ensure the correct phase split (for example by reducing storage temperature) with the inevitable loss of accuracy.


Figure 1: Phase behaviour as predicted by PC and MC approaches

Moreover, the liquid phase of the 2-phase mixture will preferentially contain the heavier components, but under PC logic the composition of both phases is identical. As an example, consider the liquid leak of a hydrocarbon mixture from a 2-phase vessel. The compositions in mole percent at various stages of the release (as predicted by the MC Phast discharge models) are as shown in Table 1.

COMPONENT

VESSEL
(BULK)

RELEASE
LIQUID

POST-FLASH
LIQUD

POST-FLASH
VAPOUR

Ethane

20

1.5

0.1

5.4

Propane

20

4.9

1.3

15.6

N-Butane

20

13.9

8.6

29.2

H-Pentane

20

30.1

30.0

30.8

H-Hexane

20

49.5

60.0

19.3

Tabel 1: Mole percent compositions of a hydrocarbon mixture using MC discharge modelling

The released liquid (column 3) contains less of the lighter components and more of the heavier. In addition the released liquid flashes on exit from the vessel and the subsequent phase split into liquid (column 4) and vapour (column 5) results in heavier components becoming even more concentrated in the liquid. The droplets eventually contain three times more of the heaviest component than the bulk mixture.

As well as the impact on discharge and dispersion results of differences in thermodynamic properties, compositional changes will also have an effect through changes in flammable and toxic properties: for example, jet fires and distance to LFL. Extensions to the MC modelling in future releases of Safeti and Phast will be able to model downwind effects such as the rainout and re-evaporation of heavier components in droplets, whilst lighter components remain as vapour as seen in Figure 2. All these effects will have an impact on the risk associated with a scenario.


Figure 2: Compositional effects in modelling discharge and dispersion of mixtures

Implementation in future version of Phast
The enhanced models will be available as an add-on to a future version of the standard Safeti and Phast releases. Not every user will wish to use the new features, but more advanced users will recognise the benefits that multi-components can bring to their analyses. However, all users will benefit from a complete overhaul of the physical property system and discharge models.

The most noticeable changes for users of the multi-component add-on will be in the materials tab, as seen in Figure 3, and in a new ‘smart’ vessel and pipe source model for setting up and running MC cases.


Figure 3: Materials tab and input for future version of Phast

It is planned for the MC calculation enhancements are added in phases to Safeti and Phast. The first release of MC capability will contain only the first stage in the full implementation of MC modelling. The most widely used continuous and instantaneous discharge scenarios will be included: leak, catastrophic rupture, line rupture, relief valve, disk rupture and 10 minute release. For dispersion, the first stage includes only equilibrium cases with no rainout. With MC thermodynamics in place in the first implementation, the priority for future releases will be to extend MC to include rainout and non-equilibrium modelling.

Contact us

Phone+47 67 57 76 50

E-maildnv.software@dnv.com

Related links