SoniceResearches - OpenQuake

My experience with OpenQuake - SMTK Module

October 10 - 2021

OpenQuake Engine is an open-source application dedicated to compute seismic hazard and risk analysis. This application is based on python language and compiles numerous tools, methods, and input-output techniques. Users can use its stand-alone web version or cmd-version to use provided tools. With some expertise with Python, user can develop their tools for OpenQuake. Numerous modules help to select the correct inputs (such as GMPEs, catalogs, source model, etc.) for OpenQuake analysis. Some of them are listed below:

Hazard Modeler ToolKit (HMTK) - For Seismic Hazard Input Preparation
Risk Modeler ToolKit (RMTK) - Fragility curves, vulnerability analysis
Strong Motion ToolKit (SMTK) - For ground motion data analysis
Catalogue ToolKit - For catalog compilation, b-value calculation, and many more.

The graph of acceleration vs time during earthquakes is called strong motions. Such ground motions and their analysis are very important in Earthquake engineering. SMTK Module of OpenQuake compiles numerous methods to compute intensity measures, such as spectral acceleration, peak ground acceleration, peak ground velocities, to name a few. These intensity measures are the backbone of earthquake engineering in the sense that these measures combine with location data, building data, and many more for hazard and risk maps. But unlike OpenQuake, for using this module, one must develop python skills. Hence, the first step in my project was learning basic python programming from YouTube and online courses, and then keep on improving them. We had 36 records from Gurkha Earthquakes, which we converted into various intensity measures using this module.

Figure. Spectral Acceleration (top) and Spectral Velocity (Bottom) of ground motion

There are 400+ ground motion prediction equations installed in OpenQuake, but finding the best combination of equations is always the big challenge. With a very small number of records available, the challenge has only increased. We created a database from a flat file containing all details of 36 events, which includes location parameters (epicenter distance, hypocenter distance), source parameters (Magnitude), and site parameters (Vs30, depth of bedrock). SMTK module provides tools to compare the fitting of records and equations using advanced techniques such as Log-Likelihood (LLH), Euclidean Distance R.. (EDR). We created some python files, which calculates LLH-value, EDR-value, residues, and predicted values to find the best fitting equation for Nepal based on these 36 records. The python files will be available in the files (sharing and learning) section shortly.

However, our experience was not as easy as described above. Being a novice in the earthquake engineering field, we had to learn every term, tool, and technique. We learned a lot from the provided user manual, but we mostly learned from understanding the source codes of the SMTK module. While writing the codes in Python for using these tools, we read with the help of some examples, and a simple technique - when in doubt, run the code.

Some suggestions while learning OpenQuake and its modules are:

Join OpenQuake Forum. This group is very active and helpful.
Read the user-manuals and other relevant papers. Understand the method before applying.
My best way of learning any techniques is by replicating the codes in Excel, or MATLAB. In the case of complex methods, I observe variables using the print function to understand the flow of the codes.
While writing your own codes, use this simple technique - When in doubts, run the codes.

Figure. Learning by Coding

Page updated

Google Sites

Report abuse