TT-SAM Realtime System

A real-time seismic intensity prediction system that utilizes deep learning to process seismic waveforms and predict ground motion intensities across Taiwan.

Features

• Real-time seismic waveform processing
• Deep learning-based ground motion prediction
• Integration with Earthworm seismic processing system
• Web-based visualization interface
• MQTT support for real-time notifications
• Multi-station processing capability

Requirements

• Earthworm
• MQTT broker
• Docker

Installation

1. Clone this repository

git clone https://github.com/SeisBlue/TTSAM_Realtime.git

2. Pull the Docker image:

docker pull seisblue/ttsam-realtime

3. Prepare the required data files in the data directory:

   • site_info.txt: Station information

     Station Channel Location Latitude Longitude Elevation Depth Azimuth Start_time End_time Constant 
     ALS HLE 10 23.508380 120.813410 2417.00 0.00 90.0 2018-08-08 2599-12-31 3.27E-6 
     ALS HLN 10 23.508380 120.813410 2417.00 0.00 0.0 2018-08-08 2599-12-31 3.24E-6 
     ALS HLZ 10 23.508380 120.813410 2417.00 0.00 0.0 2018-08-08 2599-12-31 3.25E-6 
     

   • eew_target.csv: Target stations for prediction

     network,station,station_zh,longitude,latitude,elevation
     CWB_SMT,TAP,臺北地震站,121.514,25.038,16
     TSMIP,A024,板橋地震站,121.475,25.019,14
     CWASN,NTS,淡水地震站,121.449,25.164,15
     

   • Vs30ofTaiwan.csv: VS30 data for Taiwan

     x,y,Vs30,x_97,y_97,lon,lat
     287760,2802000,534.43737793,288590.5292444,2801796.6794629595,121.3833232712489,25.324688719187737
     287840,2802000,534.24029541,288670.5304836,2801796.67998464,121.38411791685077,25.324686654878782
     287920,2802000,534.02142334,288750.5317228,2801796.6805063197,121.38491256236529,25.324684586285322
     

4. Place trained model in the model directory:

   • ttsam_trained_model_11.pt (TT-SAM)

5. MQTT configuration file:

   • ttsam_config.json

     "mqtt": {
       "username": "ttsam",
       "password": "ttsam",
       "host": "0.0.0.0",
       "port": 1883,
       "topic": "ttsam"
     }

Usage

Run the system with:

docker run \
-v $(pwd):/workspace \
-v /opt/Earthworm/run/params:/opt/Earthworm/run/params:ro \
--rm \
--ipc host \
--net host \
--name ttsam-cpu \
seisblue/ttsam-realtime \
/opt/conda/bin/python3 /workspace/ttsam_realtime.py [options]

Options:

• --mqtt: Connect to MQTT broker, default: False
• --config: MQTT configuration file, default: ttsam_config.json
• --web: Run the web server, default: False
• --host: Web server IP, default: 0.0.0.0
• --port: Web server port, default: 5000

Update

Pull the latest code:

git pull

Pull the latest Docker image:

docker pull seisblue/ttsam-realtime

System Components

• Wave Listener: Processes incoming seismic waveforms
• Pick Listener: Handles phase picks and triggering
• Model Inference: Runs deep learning prediction
• Web Server: Provides visualization interface
• MQTT Client: Broadcasts predictions

Model Architecture

TT-SAM

The system uses a deep learning model combining:

• CNN for waveform processing
• Transformer for station data integration
• MDN (Mixture Density Network) for uncertainty estimation

References

Münchmeyer, J., Bindi, D., Leser, U., & Tilmann, F. (2021). The transformer earthquake alerting model: A new versatile approach to earthquake early warning. Geophysical Journal International, 225(1), 646-656. (https://academic.oup.com/gji/article/225/1/646/6047414)

Liu, Kun-Sung, Tzay-Chyn Shin, and Yi-Ben Tsai. (1999). A free-field strong motion network in Taiwan: TSMIP. Terrestrial, Atmospheric and Oceanic Sciences, 10(2), 377-396. (http://tao.cgu.org.tw/index.php/articles/archive/geophysics/item/308)

Akazawa, T. (2004, August). A technique for automatic detection of onset time of P-and Sphases in strong motion records. In Proc. of the 13th world conf. on earthquake engineering (Vol. 786, p. 786). Vancouver, Canada. (https://www.iitk.ac.in/nicee/wcee/article/13_786.pdf)

Kuo, C. H., Wen, K. L., Hsieh, H. H., Lin, C. M., Chang, T. M., & Kuo, K. W. (2012). Site classification and Vs30 estimation of free-field TSMIP stations using the logging data of EGDT. Engineering Geology, 129, 68-75. (https://www.sciencedirect.com/science/article/pii/S0013795212000397)

Lee, C. T., & Tsai, B. R. (2008). Mapping Vs30 in Taiwan. TAO: Terrestrial, Atmospheric and Oceanic Sciences, 19(6), 6. (https://www.researchgate.net/profile/Chyi-Tyi-Lee-2/publication/250211755_Mapping_Vs30_in_Taiwan/links/557fa82608aeb61eae262086/Mapping-Vs30-in-Taiwan.pdf)

Huang, H. H., Wu, Y. M., Song, X., Chang, C. H., Lee, S. J., Chang, T. M., & Hsieh, H. H. (2014). Joint Vp and Vs tomography of Taiwan: Implications for subduction-collision orogeny. Earth and Planetary Science Letters, 392, 177-191. (https://www.sciencedirect.com/science/article/pii/S0012821X14000995)