Calculate consequences from multiple perils (ground shaking and sep) for each asset

[catarinaquintela]

Context:

In network risk assessments, it is fundamental to estimate the damage state and the non-functionality of nodes and edges not only due to ground shaking but also due to earthquake-induced hazards (liquefaction and landslide). The effects of these three sources of damage need to be considered simultaneously for each asset, as one asset might be operational if only ground shaking is considered, but non-operational due to liquefaction effects.
Currently, OQ supports the estimation of damage from ground shaking, liquefaction and landslide independently, through the definition of fragility functions that use the corresponding IMs (e.g., PGA for ground shaking; LiqOccur or LSE for landslide; Disp or DispProb for landslide). However, for each taxonomy class the user can define only one FF, and therefore the damage state is calculated considering one IM only. This means that it is not possible to estimate the damage state and consequences from ground shaking and liquefaction (and/or landslide) at the same time for one specific asset/taxonomy class.

Implementation:
For each asset, OQ needs to estimate the consequences due to different sources of damage and select the most severe one.
This implementation will require the following updates:

• Allow multiple fragility functions for the same asset, as long as they are based on different IMs (e.g., one FF for PGA, one for LiqOccur and one for DispProb).
• Update the consequence file to include not only the taxonomy class but also the corresponding IM.

Calculation steps:

• Calculate the damage state for all the FFs defined for each asset
• Calculate the corresponding consequence for each asset, for each FF
• Select the most severe consequence as the final output

[micheles]

Thanks for the very clear description, @catarinaquintela, but I also need example calculations that you can send me privately. Best if they are very small, say with a single asset.

[micheles]

There is also the problem of managing multiple loss types independently from consequences, i.e. the
scenario_damage tests case_7 and case_4b.

[micheles]

Closing since we have #10162 subsuming this issue.