Haijiang ZhangDepartment of Geology and GeophysicsUniversity of Wisconsin-Madison-ECE539 Project Report (Professor Yu Hen Hu)-Application of Multilayer Perceptron (MLP) NeuralNetwork in Identification and Picking P-wave arrivalHaijiang ZhangDepartment of Geology and GeophysicsUniversity of Wisconsin-Madison0AbstractQuickly detecting and accurately picking the first-arrival of a P wave is of greatimportance in locating earthquakes and characterizing velocity structure, especially in theera of large volumes of digital and real-time seismic data. The detector should be capableof finding the onset of the P-wave arrival against the background of microseismic andcultural noise. Normally, P-wave onset is characterized by a rapid change in theamplitude and/or the arrival of high-frequency energy. The Akaike information criteria (AIC) picker has been used to detect and pick the P-wave arrival (Maeta 1986; Maeta 1989). But AIC picker requires an appropriate timewindow, or it will detect the wrong P-wave arrival. The Multilayer Perceptron (MLP)neural network is used to detect the P-wave arrival, from which a time window can bechosen for the AIC picker. This method has been applied to our PASO array data set.About 90% of P first-arrivals are detected correctly. Compared with manual picks, thispicker provides onset times and uncertainties with high confidence. 91% of autopicks arewithin 0.15 seconds of analyst picks for this data set.11. IntroductionQuickly detecting and picking the arrival times for P and S waves from the recordingsof earthquake events are of great importance in event location, event identification,source mechanism analysis, and spectral analysis. Traditionally, this work is did by ananalyst who checking the seismograms and picking out P and S arrivals based on hisindividual experience. This task is time consuming and subjective, especially in the era oflarge volumes of digital and real-time seismic data. There is a need to provide a morereliable and robust alternative, which is less time consuming and perhaps more objective.There have been some techniques in the literature to detect and pick the seismic wavesarrivals. The traditional approach to automatic phase detection has been to apply a seriesof narrow bandpass frequency filters and then use the absolute value as the characteristicfunction (CF). When the ratio between the short term average (STA) and the long-termaverage (LTA) of the CF exceeds a predefined threshold, a detection is declared. Absolutevalues and the envelope function of the seismogram are usually used as CF (Allen, 1982).Artificial neural networks have also been used to construct the characteristic functionto detect and pick the seismic phases (Dai et al., 1995, 1997; Zhao et al., 1999; Wang etal., 1997). It is claimed that ANN method is very successful and promising in detectingand picking seismic phases. There are two different types of input vector fed to the neuralnetwork, which are the associated values of the seismograms such as mean amplitude,spectral properties, planarity, etc., and the absolute values of the seismograms,respectively. Comparatively, the former method may lose information and involve toomuch computing time. Using the full waveforms as the network input might be a betterchoice. ANN is very successful in detecting the seismic phases. However, it is difficult topick the seismic arrival time from the characteristic function. It is not easy to determinewhich point should be chosen as the arrival time because there is a region of thecharacteristic function exceeding the predefined threshold. Multi-term method is tried toshrink this region, but it still requires an empirical value to determine the phase arrival(Zhao et al., 1999). Different from the previous methods, the Akaike InformationCriterion (AIC) picker is used to pick the P-wave arrival in this report. When the time2window is chosen properly, AIC picker can choose the phase arrival very accurately. TheMLP neural network will choose a time window for the AIC picker. This report will review the AIC picker and the Multilayer Perceptron (MLP) neuralnetwork first. Then I will discuss the problem of constructing the MLP neural network todetect the P-wave arrival and how the AIC picker is used to pick the P-wave arrival.Finally the application of this method in the PASO array data is given. 2. AIC PickerSuppose that the seismogram can be divided into locally stationary segments eachmodeled as an Autoregressive (AR) process and the intervals before and after the onsettime are two different stationary processes (Sleeman et al, 1999). The order and the valueof the AR coefficients change when the characteristic of the current segment ofseismogram is different from before. For example, the typical seismic noise is wellrepresented by a relatively low order AR process, whereas seismic signals usually requirehigher order AR process (Leonard, et al., 1999). Akaike Information Criterion (AIC) isalways used to determine the order of the AR process when fitting a time series with ARprocess, which indicates the badness of the model fit as well as the unreliability (Akaike,1974). This method has been used in onset estimation by analyzing the variation in ARcoefficients representing both multi-component and single-component traces ofbroadband and short period seismogram (Leonard et al., 1999). When the order of the ARprocess is fixed, AIC function is a measure for the model fit, and the point where AIC isminimized determines the optimal separation of the two stationary time series in the leastsquares sense, and thus is interpreted as the phase onset (Sleeman et al, 1999). Thispicker is known as AR-AIC picker (Leonard, 2000). Different from AR-AIC picker, Maeta calculates AIC function directly from theseismogram, without using the AR coefficients (Maeta, 1985 and Maeta, 1986). Theonset is the point where the AIC has a minimum value. For the seismogram x, the AICvalue is defined as AIC(k)=k*log(variance(x[1,k]))+(n-k-1)*log(variance(x[k+1,n]))where k goes through all the seismogram.Noted that AIC picker finds the onset point as the global minimum. For this reason, itis necessary to choose a time window that includes only the segment of seismogram of3interest. If the time
View Full Document
Unlocking...