2D PROFILE SEISMIC WAVES FIRST BREAKS AUTOMATIC PICKING USING MACHINE LEARNING AND FULLY CONNECTED NEURAL NET
DOI:
https://doi.org/10.25635/2313-1586.2025.02.096Keywords:
first breaks, automatic picking, microseism, seismic noise, neural net, machine learning, 2D seismic dataAbstract
The article discusses a new method of automatic detection of the first breaks of seismic waves. The method used is: 2D seismic data as its input with applying the specially trained neural net mode. Existing modern approaches to the first breaks detection are described in short. Some of these approaches use neural nets and some do not. The characteristics of relevance of conducted research is provided, the process of initial field seismic data preparation for the training the model is explained. The researched algorithm of automatic first breaks picking, that uses machine learning and special trained neural net, is given in details. The calculation of accumulated seismic traces energy lies in the basis of the proposed first breaks detection method. In the research it is presumed that the time moment of seismic waves first breaks separates seismic signal into two pieces: the first one, which is noise microseismic part, and the second one that is seismic waves information part.
The accuracy of received during the investigation results is estimated in the paper. The reference first breaks time moments were taken from manual picks. The seismic field data from three of Western Siberia oilfields were used as initial input data to train and test the neural net. Seismic input data were not preprocessed to reduce noise that exists in raw data. This noise was not removed from the input in order to help the neural net adapt its parameters to confidently pick first breaks of seismic waves from raw field data. The research that is described in this text is dedicated only to 2D seismic data, 3D field seismic cubes were not considered.
References
Souza W.E., Cerqueira A.G., Porsani M.J., 2024. First-break prediction in 3-D land seismic data using the dynamic time warping algorithm. Geophysical Journal International, 237: pp. 402–418. DOI: 10.1093/gji/ggae048.
Yin Y., Han L., Zhang P., 2023. First-Break Picking of Large-Offset Seismic Data Based on CNNs with Weighted Data. Remote Sensing, 15(2), 356. DOI: 10.3390/rs15020356.
Trnkoczy A., 2012. Understanding and parameter setting of STA/LTA trigger algorithm. Potsdam, Germany: New Manual of Seismological Observatory Practice 2 (NMSOP-2). DOI: doi.org/0.2312/GFZ.NMSOP-2_IS_8.1.
Sharma B.K., Amod Kumar, Murthy V.M., 2010. Evaluation of Seismic Events Detection Algorithms. Geological Society of India, Vol.75, pp. 533-538. DOI: 10.1007/s12594-010-0042-8.
Vassallo M., Satriano C., Lomax F., 2012. Automatic Picker Developments and Optimization: A Strategy for Improving the Performances of Automatic Phase Pickers. Seismological Research Letters. DOI: 10.1785/gssrl.83.3.541.
Küperkoch L., Meier T., Diehl T., 2011. Automated Event and Phase Identification. Potsdam, Germany: New Manual of Seismological Observatory Practice 2 (NMSOP-2). DOI: 10.2312/GFZ.NMSOP-2_ch16.
Li X., Shang X., Wang Z., Dong L., Weng L., 2016. Identifying P-phase arrivals with noise: An improved kurtosis method based on DWT and STA/LTA. Journal of Applied Geophysics. DOI: 10.1016/j.jappgeo.2016.07.022.
Shen T., Tuo X., Li H., Liu Y., Rong W., 2018. A first arrival picking method of microseismic data based on single time window with window length independent. Springer Nature. DOI: 10.1007/s10950-018-9789-y.
Stampa J., Eckel F., Keers H., Lebedev S., Meier T., AlpArray and SWATH-D Working Groups, 2024. Automated measurement of teleseismic P -, SH - and SV-wave arri-val times using autoregressive prediction and the instantaneous phase of multicomponent waveforms. Geophysical Journal International. DOI: 10.1093/gji/ggae307.
Feng J., Lu Sh., 2019. Performance Analysis of Various Activation Functions in Artificial Neural Networks. Journal of Physics: Conf. Series 1237 (2019) 022030. DOI: 10.1088/1742-6596/1237/2/022030.
Приезжев И.И, Иванов П.Д., Гаврилов С.С., Мамаев Д.А., Калинин А.Ю., Стенина Ю.В., 2022. Автоматическая пикировка первых вступлений с использованием машинного обучения. Геофизика, № 1, С. 90-96. DOI: 10.34926/geo.2022.65.65.001.
Shen Yu-Ju, Wang Ming-Shi, 2005. Apply neural schemes to deformation objects. Taiwan: ICGST-GVIP Journal, Vol. 5, Issue 4.
Приезжев И.И., Мамаев Д.А., Стенина Ю.В., 2021. Использование элементов машинного обучения для автоматической пикировки первых вступлений. Геомодель 2021, 23-я конференция по вопросам геологоразведки и разработки месторождений нефти и газа. Москва: Издательство ООО «ЕАГЕ Геомодель», С.114.
Münchmeyer J., Woollam J., Rietbrock A., Tilmann F., Lange D., Bornstein T., Diehl T., Giunchi C., Haslinger F., Jozinovi´c D., Michelini A., Saul J., Soto H., 2021. Which picker fits my data? A quantitative evaluation of deep learning based seismic pickers. arXiv:2110.13671v1 [physics.geo-ph]. DOI: 10.48550/arXiv.2110.13671.
Johnson Sean W., Chambers Derrick J. A., Boltz Michael S., Koper Keith D., 2021. Application of a convolutional neural network for seismic phase picking of mining-induced seismicity. Geophysical Journal International, Volume 224, Issue 1. DOI: 10.1093/gji/ggaa449
St-Charles Pierre-Luc, Rousseau B., Ghosn J., Nantel Jean-Philippe, Bellefleur G., Schetselaar E., 2021. A Multi-Survey Dataset and Benchmark for First Break Picking in Hard Rock Seismic Exploration. Fourth Workshop on Machine Learning and the Physical Sciences (NeurIPS 2021).
Krakovská H., Kuehn C., Longo I.P., 2024. Resilience of Dynamical Systems. European Journal of Applied Mathematics, 35: 155–200. DOI: 10.1017/S0956792523000141.
Terven J., Cordova-Esparza Diana M., Ramirez-Pedraza A., Chavez-Urbiola Ed-gar A., Romero-Gonzalez Julio A., 2023. Loss Functions and Metrics in Deep Learning. arXiv:2307.02694. DOI: 10.48550/arXiv.2307.02694.
