Reduction of Train Noise from Telluric Current Data by Neural Networks Kazuki Joe (System Designer)...

20
Reduction of Train Noise from Telluric Current Data by Neural Networks Kazuki Joe (System Designer) Toshiyasu Nagao (VAN Method Advisor) Mika Koganeyama (Neural Network Implementor) Moyo Sugita (Visualization Implementor) Nara Women’s University Tokai University
  • date post

    20-Dec-2015
  • Category

    Documents

  • view

    215
  • download

    0

Transcript of Reduction of Train Noise from Telluric Current Data by Neural Networks Kazuki Joe (System Designer)...

Reduction of Train Noise from Telluric Current

Data by Neural Networks

Kazuki Joe (System Designer)

Toshiyasu Nagao (VAN Method Advisor)

Mika Koganeyama (Neural Network Implementor)

Moyo Sugita (Visualization Implementor)

Nara Women’s UniversityTokai University

Earthquakes in Japan ( 1990-1994 )

What is short-term earthquake prediction?

Time

Earthquake

Electro-Magnetic Phenomena

Aftershock Aftershock Aftershock

Definition of Earthquake

EarthquakeDec. 1994Feb. 19953 months

90 days

VAN method

Designed by Greek physicists enable to observe SESs

50m 1km

N

Short dipoles ( 30 ~ 200m )Long dipoles ( several km )

electrode

Telluric Current Data(TCD) Feeble current that flows in the earth surface

– potential difference between 2 points by burying electrodes in the earth

observation points– 42 points (Tokai and Hokuriku area)

– 8 channels or 16 channels for each observation point

– observe every 10 seconds (8640 data on one day) seismic electric signals(SESs)

Seismic electric signals(SESs)

Current changes before earthquake– earthquake is a kind of events of

destroying rocks

– current flows before rocks are destroyed

20 ~ 30 minutes one-way amplitude

find the signals by specialists 300 frames ( 50 minutes )

about 160 frames ( 27 minutes )

Case Study

Big earthquake in Greece, Pirgos city ( in March, 1993 )– Seismic electric signal was detected before the earthquake.

– By the prediction, some part of resident are evacuated.• half of buildings (about 4000 ridge) were destroyed completely or partially

• no casualties

effectiveness of TCD

Investigate TCD in Japan

Problem of the use of VAN method in Japan

TCD components in Japan

TCD

train noise( about 90% )

other noise

SESSES

Characteristics of Train Noise

TCD

Timetable (Nagano railway Matsushiro station)

6:10 6:46 7:26 8:06

Regularity of the appearance

Similarity of the shapecan be learned & recognized by Neural Networks

Up-train

Down-train

Train noise reduction filter

- Basic Idea -

train noise reduction filter

Train noise + SES

constructed by neural network

SES

Problem of Constructing the Filter by Neural Networks

NNs require training and supervising samples– the TCD with train noise and SESs are very rare (only

several ten cases)

– no TCD with the same SESs without the train noise

Generate training and supervising samples artificially

Artificial Generation of Training & Supervising Samples

Pre-processed TCD (LF components are cut)

300 frames300 frames120 ~ 250frames ( about 20 ~ 40minutes )

Train noise Natural noise

Artificial Generation of Training & Supervising Samples

Training data Supervising data

+ +

Train noise Natural noise SES

Natural noise SES

300 frames 300 frames

Artificial Generation of Training & Supervising

Samples More shift-tolerant neural network to time series data

– train noise and SES are shifted right for several points as shown below

Train noise

Supervising data

Training data

Experiment Result After the learning, only train noise from unknown

TCD data could be removed.– unknown TCD is generated artificially by train noise and an SES

Demonstration

Artificial generation of TCD with train noise and an SES arbitrarily

Train noise reduction of TCD with SESs Train noise reduction of unknown TCD

Conclusion and Future Work

It turned out that just train noise can be removed from TCD by neural networks

Can be a big progress toward automatic short-term earthquake prediction

More learning with other observation points Design and implementation of SES detector Other method...