Inquiry on Acquiring Risk Values Corresponding to IM from OQ Group

[bahar r]

Hello OQ Group,

I am currently conducting a scenario risk assessment and have calculated the values for 1000 GMF for each point. The results I obtained from the OQ include 1000 IM for each point as well as the mean of my risk values (structural, etc.). However, I noticed that the OQ does not provide the risk value corresponding to each individual IM.

Could you please advise on how I can acquire the risk values for all 1000 IM? 

Thank you for your assistance.

Best regards,

Bahar

[Anirudh Rao]

You'll need a short script like the one described in a previous thread to extract this information from the hdf5 file for your calculation: https://groups.google.com/g/openquake-users/c/wlieWppKKps/m/1aB3uut7AgAJ . This is not provided as a default output because for realistic calculations this output file would be massive, and most users do not need this.

[bahar r]

Hello  Anirudh,

Thank you for the response and the guidance you provided.

In my case, I want to calculate the scenario risk. I added aggregate_by = id to the job initial file. Upon checking the output HDF5 file, I observed that agg_keys has been added to the table groups, which now contains a single column of exposures.

Additionally, I examined the data group of risk_by_event, which includes the following dataframes: agg_id, event_id, loss, loss_id, and variance.

My requirement is to calculate the risk corresponding to their IM for each point (site). For the gmf_data, it includes the sid, which is the site id, and I need something similar for the risk data. Moreover, I calculated 1000 gmfs for each point. 

I would be grateful if you could guide me further on how to achieve this.

Thank you in advance for your assistance.

Best regards,

Bahar

[bahar r]

Hello,

Thank you for your previous guidance. I was successful in obtaining results from various simulations using aggregate_by = id in the job.ini file. However, I used the attached code to save the results from different simulations and the HDF5 file, and upon comparing the average loss value from my results (results from the attached codes) with the Average Asset Losses Statistics provided by OQ outputs, I noticed considerable discrepancies.

The ground motion model (GSC 2020) I used includes 16 branches, each with a weight dependent on Vs30 and IM period. As you may know, OQ simulates the number_of_ground_motion_fields for each ground motion branch. For instance, if the number_of_ground_motion_fields is 100, OQ simulates 16 * 100 scenarios.

One reason for the discrepancy might be because I simply calculated the mean of these 1600 scenarios. I was wondering if OQ calculates the mean of the loss across these 1600 events or considers the weight of ground motion models? Previously, I inquired about this, and it seemed that OQ considers the mean of scenarios (https://groups.google.com/g/openquake-users/c/Ijy5t4dfeas/m/6CbrlIezAwAJ).

Do you have any idea why there is a discrepancy between the mean of the 1600 loss scenarios I calculated and the result of the Average Asset Losses Statistics provided by OQ?

Thank you for your assistance.

Best regards,

Bahar

[Peter Pažák]

Hi,

last time you were asking about the average ground motion values (which seem to be geometric mean),

now you have average asset losses, so it is a bit different thing.

Peter

Dátum: sobota 15. júna 2024, čas: 20:45:37 UTC+2, odosielateľ: bahar...@gmail.com