Original Waveform Research Articles

Automated microseismic event location by amplitude stacking and semblance

Without the need to pick the arrival times of P- and S-waves, migration-based location methods, such as semblance-based and amplitude-stacking-based location methods, are best applied to microseismic events. By comparing and analyzing the advantages and disadvantages of these two methods, we have developed a new location method using amplitude information and semblance. First, we use the two-point ray-tracing method to calculate the traveltime of body waves from the trial point to each receiver, which determines the time-window positions of the P- and S-waves on all traces. Then, we calculate the semblance of the waveforms and the amplitude stacking of the ratio between the short-time average and the long-time average is computed upon the original waveform over the windows. Finally, the semblance weighted by amplitude stacking is used to image the spatial location of the microseismic events. Using experimental and synthetic data considering different factors that may affect the location result (e.g., the signal-to-noise ratio of the waveforms, the scale of the observation array, and the horizontal and vertical distances from the source to fracture zones), we perform microseismic event location with all three methods. According to the source imaging results from experimental and synthetic tests, the semblance method has great location uncertainty in the radial direction but it has good constraints in the circumferential direction; the amplitude-stacking method exhibits the opposite result; and the weighted-semblance method has good constraints in the circumferential and radial directions because it inherits the advantages of semblance-based and amplitude-stacking-based methods. Therefore, compared with existing migration-based location methods, our weighted-semblance method indicates stronger stability and lower location uncertainty, even when downhole monitoring is conducted with a limited aperture of the receiver array.

Patient asynchrony modelling during controlled mechanical ventilation therapy

Background and objectiveMechanical ventilation therapy of respiratory failure patients can be guided by monitoring patient-specific respiratory mechanics. However, the patient's spontaneous breathing effort during controlled ventilation changes airway pressure waveform and thus affects the model-based identification of patient-specific respiratory mechanics parameters. This study develops a model to estimate respiratory mechanics in the presence of patient effort. MethodsGaussian effort model (GEM) is a derivative of the single-compartment model with basis function. GEM model uses a linear combination of basis functions to model the nonlinear pressure waveform of spontaneous breathing patients. The GEM model estimates respiratory mechanics such as Elastance and Resistance along with the magnitudes of basis functions, which accounts for patient inspiratory effort. Results and discussionThe GEM model was tested using both simulated data and a retrospective observational clinical trial patient data. GEM model fitting to the original airway pressure waveform is better than any existing models when reverse triggering asynchrony is present. The fitting error of GEM model was less than 10% for both simulated data and clinical trial patient data. ConclusionGEM can capture the respiratory mechanics in the presence of patient effect in volume control ventilation mode and also can be used to assess patient-ventilator interaction. This model determines basis functions magnitudes, which can be used to simulate any waveform of patient effort pressure for future studies. The estimation of parameter identification GEM model can further be improved by constraining the parameters within a physiologically plausible range during least-square nonlinear regression.

Switch On-off Automatic Detection Technology Based on Basic Model of Waveform Singularity Detection

During the test of low-voltage switches, numerous waveforms are required to be collected and electrical parameters calculated. Traditional detection methods are inefficient and may result in operational errors. Since the sequencing of the test waveform array is converted to corresponding time points by the sampling frequency conversion, to realize automatic positioning of the cursor in waveform and automatic measurement of the electrical parameters, the key lies in the accurate detection of the singular position in the waveform array. This paper builds the basic model of waveform singularity detection based on the singular value decomposition (SVD) algorithm, integrates automatic detection into the electrical parameter test module, realizing the automatic detection of waveforms and parameters during the low-voltage appliance test. Meanwhile, starting from the automatic detection of the conduction time of the same test current when the alternating current contactor is switched on and off, the anti-interference ability of the model under different types of noise and different noise intensities are tested and analyzed. For all types of noise, reliable, uncertain, and unreliable detection zones exist in the automated detection results. When k (effective value of the noise / effective value of original waveform) is less than 10%, all types of noise can be detected in a reliable manner, which effectively J. van der Geer, J.A.J. Hanraads, R.A. Lupton, The art of writing a scientific article, J. Sci. Commun. 163 (2000) 51-59.

A novel feature representation method based on original waveforms for acoustic emission signals

Abstract One of the most important issues arising in the use of acoustic emission (AE) for nondestructive process monitoring is the accurate identification of potential process malfunctions to avoid premature failure. In some cases, the AE signals from malfunction sources are relatively weak, with high levels of background noises. Thus, these signals could easily be submerged, making it rather difficult to separate them. Therefore, it is of critically importance to find a solution to the problem of weak emission source identification and obtain a correct representation of original waveforms. The present work proposes a new feature representation method based on similarity of probability distributions from raw AE waveforms. The Bhattacharyya coefficient is used for this purpose. A standard procedure for calculating similarity is formulated. Both an instantaneous similarity and a relative similarity are defined. The influences of the choice of some key parameters are discussed in detail. Tests on filament breakage detection in an additive manufacturing process reveal the feasibility and effectiveness of the proposed method. This method is believed to be appropriate when the target malfunction emission signal amplitude is less than the environmental emission signals generated by other stationary sources, and threshold methods fail to perform properly. It could also be used as an alternative feature representation method for AE signals in other fields.

Estimation of the Orientations of the S‐net Cabled Ocean‐Bottom Sensors

ABSTRACT The Seafloor Observation Network for Earthquakes and Tsunamis along the Japan Trench (S‐net) is a novel cabled ocean‐bottom station network covering a broad offshore region east of northeastern Japan. To best use the S‐net data, we estimated sensor orientations of all 150 S‐net stations, because without this information the orientations of measurements in geodetical coordinates cannot be specified. We determined three parameters of the sensor orientation at each station: the tilt angle of the long axis of the cable, the rotation angle around the long axis, and the azimuth of the long axis. We estimated the tilt and rotation angles by using the direct current components of accelerometers recording the gravitational acceleration. The tilt and rotation angles slightly varied within the range of 0.001°–0.1° for most stations during the period from 2016 to 2018 except for coseismic steps of rotation angles greater than 1° because of the 20 August 2016 Mw 6.0 off Sanriku and 20 November 2016 Mw 6.9 off Fukushima earthquakes. The long‐axis azimuths were estimated by the particle motions of long‐period Rayleigh waves. We used the accelerometer records in 0.01–0.03 Hz of 7–14 teleseismic earthquakes with Mw 7.0–8.2. The azimuths were constrained with 95% confidence intervals of ±3°–12°. After correcting original waveforms based on the estimated sensor orientation, we confirmed coherent waveforms within the whole S‐net stations and separation of Rayleigh and Love waves in radial and transverse components. The waveforms were also coherent with those of on‐land broadband stations. We provide the estimated sensor orientations and rotation matrix for conversion from the XYZ to east, north, and up components. The estimated orientation can be a fundamental resource for further seismic and geodetic explorations based on S‐net data.

Frequency Invariant Transformation of Periodic Signals (FIT-PS) for Classification in NILM

This paper presents a new signal representation called frequency invariant transformation of periodic signals (FIT-PSs) in the context of non-intrusive load monitoring (NILM). Compared to former approaches where a conglomeration of different signal forms are used, the presented approach is based on a single signal form containing all information. The core idea of this paper is to use the original current waveform relative to the reference voltage as a signature for NILM. In general, the relation of sampling and grid frequency is subject to continuous fluctuations. Therefore, FIT-PS converts uncorrelated sample data to a fixed multiple of the grid frequency. The advantages are that the information of the current signal, as well as the phase shift between voltage and current signal, are completely contained in the FIT-PS signal representation. For classification, a neural net was applied to the home equipment laboratory dataset 1. Features created by FIT-PS are superior to the standard features. With 18 different devices, a detection rate of up to 90% is achieved. In particular, when several consumers are active at the same time, the new signal representation is much more robust and leads to a better detection rate. However, a long short-term memory net with FIT-PS signal representation provides the best results.

Data augmentation using generative adversarial networks for robust speech recognition

For noise robust speech recognition, data mismatch between training and testing is a significant challenge. Data augmentation is an effective way to enlarge the size and diversity of training data and solve this problem. Different from the traditional approaches by directly adding noise to the original waveform, in this work we utilize generative adversarial networks (GAN) for data generation to improve speech recognition under noise conditions. In this paper we investigate different configurations of GANs. Firstly the basic GAN is applied: the generated speech samples are based on spectrum feature level and produced frame by frame without dependence among them, and there is no true labels. Thus, an unsupervised learning framework is proposed to utilize these untranscribed data for acoustic modeling. Then, in order to better guide the data generation, condition information is introduced into GAN structures, and the conditional GAN is utilized: two different conditions are explored, including the acoustic state of each speech frame and the original paired clean speech of each speech frame. With the incorporation of specific condition information into data generation, these conditional GANs can provide true labels directly, which can be used for later acoustic modeling. During the acoustic model training, these true labels are combined with the soft labels which make the model better. The proposed GAN-based data augmentation approaches are evaluated on two different noisy tasks: Aurora4 (simulated data with additive noise and channel distortion) and the AMI meeting transcription task (real data with significant reverberation). The experiments show that the new data augmentation approaches can obtain the performance improvement under all noisy conditions, which including additive noise, channel distortion and reverberation. With these augmented data by basic GAN / conditional GAN, a relative 6% to 14% WER reduction can be obtained upon an advanced acoustic model.

Principal Component Analysis Reveals the Proximal to Distal Pattern in Vertical Jumping Is Governed by Two Functional Degrees of Freedom.

The successful completion of motor tasks requires effective control of multiple degrees of freedom (DOF), with adaptations occurring as a function of varying performance constraints. In this study we sought to compare the emergent coordination strategies employed in vertical jumping under different task constraints [countermovement jump (CMJ) with arm swing-CMJas and no arm swing-CMJnas]. In order to achieve this, principal component analysis (PCA) was conducted on joint moment waveform data from the hip, knee and ankle. This statistical approach has the advantage of analyzing the whole movement within a time series and reduces multidimensional datasets to lower dimensions for analysis. Both individual and group analyses were conducted. For individual analysis, PCA was conducted on combined hip, knee, and ankle joint moment data for each individual across both CMJnas (thirty-eight participants), and CMJas (twenty-two participants) conditions. PCA was also performed comparing all data from each individual across CMJnas and CMJas conditions. The results revealed a maximum of three principal components (PC) explained over 90% of the variance in the data sets for both conditions and within individual and group analyses. For individual analysis, no more than 2PCs were required for both conditions. For group analysis, CMJas required 3PCs to explain over 90% of the variance within the dataset and CMJnas only required 2PCs. Reconstruction of the original NJM waveforms from the PCA output demonstrates a greater loading of hip and knee joint moments to PC1, with PC2 showing a greater loading to ankle joint moment. The reduction in dimensions of the original data shows the proximal to distal extension pattern in the sagittal plane, typical of vertical jumping tasks, is governed by only 2 functional DOF, at both a group, and individual level, rather than the typically reported 3 mechanical DOF in some forms of jumping.

Determining velocity of detonation using high-resolution time domain reflectometry

Two methods are presented to make measurements of the length of a coaxial cable with submillimeter and submicrosecond accuracy. If the cable is destroyed by the passage of a strong shock, for instance, from the detonation of an explosive, a measure of the phase velocity of the detonation (VOD) can be obtained with high accuracy (≈0.25%). The first method introduces a series of short Gaussian electrical pulses (<300 ps) into the cable that reflect off the cable end and are compared to the time of the original outgoing pulse by using a fast digitizer (100 gigasamples per second). By curve fitting a Gaussian form to the outgoing and reflected pulses, a measurement of the cable length can be made with subsample accuracy. So, as long as the pulses do not overlap, a large number may be present in the cable at any time allowing pulse repetition rates of 40-135 MHz depending on the cable length. Averaging allows accurate VOD measurements to be made that are comparable with the best traditional techniques. The second method replaces the pulser with a wide-spectrum (0.1-10 GHz) amplified random noise source. A similar setup and digitizer are used before autocorrelating short (80-160 ns) subsamples of the total waveform to measure the reflected delay time. That is, a smaller amplitude version of the original waveform appears time shifted by twice the transit time in the cable and this time is estimated via autocorrelation. The advantage of this method is that a new subsample needs only to be selected from a time increment in which new spectral information is present for a new measure of the length to be made. In this manner, many more length estimates can be averaged to establish the VOD than with the Gaussian approach. A detailed description of the implementation of both methods is presented together with an example of the type of data collected and limitations found.

Lower limb sagittal gait kinematics can be predicted based on walking speed, gender, age and BMI

Clinical gait analysis attempts to provide, in a pathological context, an objective record that quantifies the magnitude of deviations from normal gait. However, the identification of deviations is highly dependent with the characteristics of the normative database used. In particular, a mismatch between patient characteristics and an asymptomatic population database in terms of walking speed, demographic and anthropometric parameters may lead to misinterpretation during the clinical process. Rather than developing a new normative data repository that may require considerable of resources and time, this study aims to assess a method for predicting lower limb sagittal kinematics using multiple regression models based on walking speed, gender, age and BMI as predictors. With this approach, we were able to predict kinematics with an error within 1 standard deviation of the mean of the original waveforms recorded on fifty-four participants. Furthermore, the proposed approach allowed us to estimate the relative contribution to angular variations of each predictor, independently from the others. It appeared that a mismatch in walking speed, but also age, sex and BMI may lead to errors higher than 5° on lower limb sagittal kinematics and should thus be taken into account before any clinical interpretation.

An enhanced Lamb wave virtual time reversal technique for damage detection with transducer transfer function compensation

The Lamb wave time reversal method has widely been investigated as a baseline-free damage detection technique for structural health monitoring. Due to the mode tuning effects from the transducer-wave interactions, even for a pristine wave path, the reconstructed signal waveform may differ much from the original excitation waveform. Consequently, it becomes difficult to distinguish the differences between undamaged and damaged wave paths. This article presents an enhanced Lamb wave virtual time reversal (VTR) algorithm with transducer transfer function compensation to eliminate the transducer influence for dispersive, multimodal Lamb waves. This VTR procedure builds upon a complete 2D analytical model for Lamb wave generation, propagation, and reception. The analytical solution shows that, with the transducer transfer function compensation, a perfect reconstruction of the original excitation waveform can be achieved for both symmetric and antisymmetric Lamb modes. In addition, finite element modeling and experimental validations are further performed to verify the enhanced time reversal procedure. Finally, a time reversal tomography experiment is conducted with a piezoelectric transducer array for structural damage imaging. The enhanced VTR method can achieve more accurate and robust damage imaging results. The paper finishes with discussion, concluding remarks, and suggestions for future work.

Fast Detection and Compensation of Current Transformer Saturation Using Extended Kalman Filter

In this paper, an efficient method based on the Kalman filter (KF) theory is proposed for fast and accurate detection and compensation of current transformer (CT) saturation. During the unsaturated regions, the algorithm uses the current sample points together with extended Kalman filter (EKF) for obtaining a model, which suitably represents the current signal behavior. The acquired model is used to reconstruct the original current waveform during the saturated regions. The onset and end points of the CT saturation intervals are detected by defining a criterion based on the estimation error of the EKF algorithm. The proposed method is very profitable as it requires only one sample point to detect the CT saturation and it also copes with the problems of measurement noises and harmonics. These advantages are confirmed by a large number of test cases obtained using PSCAD/EMTDC software as well as by real-time hardware implementation.

Spike Sorting Based on Window-Gradient Feature

Spike sorting is difficult when there is high waveforms similarity between different spike or when there is a large number of superimposed spikes in the sample. A new sample optimization method is proposed in the paper, called window-gradient feature. Every spike waveform is segmented into successive fragments in terms of the width σ, and each segment is called a window. Then calculate the gradient change of each window and use the ratio as the new alternative feature of the window. Finally, all window-gradient features of the spike waveforms are used to replace the original waveform features for spike sporting. The method is verified on the simulation data at different SNR. The experimental results show that when using SVM for spike sorting, the optimization effect of the proposed method is better than PCA, especially for data sets data sets with high noise or large sampling waveform similarity.

Triton identification in the 6Li(n, t)4He reaction measurement with the grid ionization chamber at CSNS Back-n white neutron source

The measurement of 6Li(n, t)4He differential cross section is the first experiment conducted with the Light-charged Particle Detector Array (LPDA) at the Back-n white neutron source of the China Spallation Neutron Source. The LPDA system consists of three kinds of detectors. One of them is the grid ionization chamber (GIC), which has the capability of particle identification. The particle identification performance of the GIC is presented based on the data collected in this experiment. Signals from the anode and the cathode of the GIC are digitized through the readout electronics and DAQ system. The offline data analysis is performed in two major steps. The first step is the processing of the original signal waveform digitized by the readout electronics. A set of digital filter algorithms including RC-CR filter, band-pass filter and digital CFD are developed and implemented in the original signal processing in order to obtain a high signal-noise ratio and a precise timing resolution. The second step focuses on the identification of triton events. The triton identification efficiency is measured to be 98% using a multivariate analysis algorithm implemented in the TMVA package in ROOT to discriminate the triton events and background events. In general, the performance of the GIC has met the baseline requirement of the experiment measurement.

A Seismic Signal Denoising Method Based on Wavelet Comprehensive Threshold

The wavelet comprehensive threshold is used to inherit and develop the advantages of hard threshold and soft threshold denoised method. Based on the small wavelet coefficients of the noise signal and the large wavelet coefficients of the seismic signal, the simulation experiment of the Ricker signal and the noise reduction experiment is carried out. The simulation results show that the MSE of the wavelet comprehensive threshold is the closest to the original signal waveform after noise reduction, and the energy of the high frequency part of the waveform is reduced and the low frequency part is suppressed. Finally, the actual seismic waveform, for example, the noise reduction of the actual waveform can be the first time of the waveform, and the waveform of the effective signal energy, noise signal energy is suppressed.

Model-Based Pileup Events Correction via Kalman-Filter Tunnels

A Kalman-filter-based method is proposed to address challenges in nuclear signal processing, including pileup events and noise cancelation. The state-transition equation is established on the basis of suitable pulse shape models. By regarding the output signal of particle detectors as a measurable variable, the measurement equation is also introduced to correct dynamic processes of the state transition. A linear Kalman filter with $n$ tunnels is constructed to separate pileup events and maintain the original waveforms of pulses. In addition, two terms are defined and addressed for separation operations, namely, Availability , which is used to prepare for upcoming events, and Clearance Mechanism , which is given to remove thumbnail information from the previous events and make occupied tunnels available to subsequent nuclear events. Simulation and field experiments demonstrate that the proposed method can handle critical situations of pulse pileup and minimize count loss and dead time in the X-ray energy spectrum while maintaining the original pulse shapes.

Association Between Magnetoencephalographic Interictal Epileptiform Discharge and Cognitive Function in Young Children With Typical Development and With Autism Spectrum Disorders.

Electroencephalograms of individuals with autism spectrum disorders (ASD) show higher rates of interictal epileptiform discharges (IEDs), which are known to have an inverse association with cognitive function in typically developed (TD) children. Nevertheless, that phenomenon has not been investigated adequately in children with ASD. From university and affiliated hospitals, 163 TD children (84 male, 79 female, aged 32–89 months) and 107 children (85 male, 22 female, aged 36–98 months) with ASD without clinical seizure were recruited. We assessed their cognitive function using the Kaufman Assessment Battery for Children (K-ABC) and recorded 10 min of MEG. Original waveforms were visually inspected. Then a linear regression model was applied to evaluate the association between the IED frequency and level of their cognitive function. Significantly higher rates of IEDs were found in the ASD group than in the TD group. In the TD group, we found significant negative correlation between mental processing scale scores (MPS) and the IED frequency. However, for the ASD group, we found significant positive correlation between MPS scores and the IED frequency. In terms of the achievement scale, correlation was not significant in either group. Although we found a correlative rather than a causal effect, typically developed children with higher IED frequency might better be followed up carefully. Furthermore, for children with ASD without clinical seizure, clinicians might consider IEDs as less harmful than those observed in TD children.

Microseismic event location using multi-scale time reversed imaging

Locating passive seismic sources in a reservoir monitoring system has attracted considerable interest recently due to its ability to image the induced fracture geometry that results from fluid injection and fracturing processes. Time reversed imaging (TRI) techniques are well recognized localization tools, which back-propagates the received microseismic recordings to focus at its real source location without picking various arrival phases. However, the time reversed images are often contaminated due to the strong noise, coherent noise and poor spatial constraints especially in surface microseismic measurements, leading to unreliable location estimations. To minimize these interferences, we propose a multi-scale TRI technique using the shift-invariant dual-tree complex wavelet transform (DTCWT) to decompose the original waveforms into multiple time-frequency domains (different levels). The Birge-Massart threshold is applied to the wavelet coefficients at each level to further attenuate the noise amplitude. The TRI are then applied to the reconstructed waveform component at each level. Both the synthetic and field studies show that the multi-scale cross-correlation of TR images at effective levels substantially improved the quality of the final image of subsurface microseismic events with a much sharper and brighter focus. On the other hand, the images at noise levels may reveal the velocity structure, which is helpful in event location. In addition, the noise components recognized by the multi-scale TRI can be applied to estimate the background noise level, which can be further used in microseismic sensitivity and location uncertain analysis.

Full-intensity waveform inversion

Many full-waveform inversion schemes are based on the iterative perturbation theory to fit the observed waveforms. When the observed waveforms lack low frequencies, those schemes may encounter convergence problems due to cycle skipping when the initial velocity model is far from the true model. To mitigate this difficulty, we have developed a new objective function that fits the seismic-waveform intensity, so the dependence of the starting model can be reduced. The waveform intensity is proportional to the square of its amplitude. Forming the intensity using the waveform is a nonlinear operation, which separates the original waveform spectrum into an ultra-low-frequency part and a higher frequency part, even for data that originally do not have low-frequency contents. Therefore, conducting multiscale inversions starting from ultra-low-frequency intensity data can largely avoid the cycle-skipping problem. We formulate the intensity objective function, the minimization process, and the gradient. Using numerical examples, we determine that the proposed method was very promising and could invert for the model using data lacking low-frequency information.

Spread-Spectrum Method Using Multiple Sequences for Underwater Acoustic Communications

Direct-sequence spread-spectrum-based approaches have been adopted in underwater acoustic communications. In this paper, multiple sequences utilizing the periodic correlation characteristics of an original spreading waveform are superimposed, modulating multiple distinct symbols simultaneously within each block at the transmitter. Among those superimposed symbols, one is used as the pilot for channel estimation on chip level and others carry data. Similar to existing superposition schemes, the proposed approach also has low receiver complexity, only requiring matched filtering operations. However, the innovation of the proposed approach is perfect interference suppression, cyclic prefix (CP) is inserted to eliminate the interblock interference totally. Combined CP operation leading to an equivalent cyclic channel matrix with circularly shifted versions of spreading sequences, interferences can be solved considerably. The proposed technique is evaluated with both simulation and experimental data analysis. Data processing gives a brief discussion about low probability of detection and low probability of interception.