Logging Signals Research Articles

Leader Confirmation Replication for Millisecond Consensus in Private Chains

The private chain-based Internet-of-Things (IoT) system ensures the security of cross-organizational data sharing. As a widely used consensus model in private chains, the leader-based state-machine replication (SMR) model meets the performance bottleneck in IoT blockchain applications, where nontransactional sensor data are generated on a scale. We analyzed IoT private chain systems and found that the leader maintains too many connections due to high latency and client request frequency, which results in lower consensus performance and efficiency. To meet this challenge, we propose a novel solution for maintaining low request latency and high transactions per second (TPS): replicate nontransactional data by followers and confirm by the leader to achieve nonconfliction SMR, rather than all by the leader. Our solution, named leader confirmation replication (LCR), uses the newly proposed future log and confirmation signal to achieve nontransactional data replication on the followers, thereby reducing the leader’s network traffic and the request latency of transactional data. In addition, the generation replication strategy is designed to ensure the reliability and consistency of LCR when meeting membership changes. We evaluated LCR with various cluster sizes and network latencies. The experimental results show that in ms-network latency (2–30) environments, the TPS of LCR is 1.4X-1.9X higher than Raft, the transactional data response time is reduced by 40%–60%, and the network traffic is reduced by 20%–30% with acceptable network traffic and CPU cost on the followers. In addition, LCR shows high portability and availability since it does not change the number of leaders or the election process.

Lithology prediction from well log data using machine learning techniques: A case study from Talcher coalfield, Eastern India

Coal exploration in the Indian scenario is challenging due to plenty of carbon contents and dirt bands within a coal seam. Manual interpretation of geophysical logging data in such conditions is time-consuming and tedious due to the non-linear behaviors of well log signals. This paper intends to apply supervised machine learning techniques such as Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Multi-Layer Perceptron (MLP), and Extreme Gradient Boosting technique (XGBoost) for meticulously interpreting such banded coal seams from geophysical logs. To investigate the efficacy of the above-mentioned five ML techniques in the Indian scenario, we have considered Gamma-ray, Density, and Resistivity logs from four boreholes drilled in Talcher coalfield, Eastern India. ML model training results indicate that all the prediction models have scored more than 88% accuracy scores in classifying carbonaceous and non-coal lithofacies. Receiver Operating Characteristics (ROC) curves of different ML models suggest that the obtained area under the curve (AUC) is positive and above the main diagonal for all lithotypes. Finalized ML models with appropriate hyper-parameters are also applied on the nearby drilled wells, and the outputs are validated against geological core data. It is found that all ML models have acquired an overall accuracy score greater than 80% in all three test wells, conveying that ML techniques are a potential solution for dealing with banded coal seam problems in the Indian scenario.

Compression Behavior of EBM Printed Auxetic Chiral Structures.

In this study, the cyclic compression and crush behavior of chiral auxetic lattice structures produced from titanium alloy (Ti6Al4V) metallic powder using electron beam melting (EBM) additive manufacturing technology is investigated numerically and experimentally. For material characterization and understanding the material behavior of EBM printed parts, tensile and three-point flexural tests were conducted. Log signals produced during the EBM process were investigated to confirm the stability of process and the health of the produced parts. Furthermore, a compressive cyclic load profile was applied to the EBM printed chiral units having two different thicknesses to track their Poisson’s ratios and displacement limits under large displacements in the absence of degradation, permanent deformations and failures. Chiral units were also crushed to investigate the effect of failure and deformation mechanisms on the energy absorption characteristics. Moreover, a surface roughness study was conducted due to high surface roughness of EBM printed parts, and an equation is offered to define load-carrying effective areas to prevent misleading cross-section measurements. In compliance with the equation and tensile test results, a constitutive equation was formed and used after a selection and calibration process to verify the numerical model for optimum topology design and mechanical performance forecasting using a non-linear computational model with failure analysis. As a result, the cyclic compression and crush numerical analyses of EBM printed Ti6Al4V chiral cells were validated with the experimental results. It was shown that the constitutive equation of EBM printed as-built parts was extracted accurately considering the build orientation and surface roughness profile. Besides, the cyclic compressive and crush behavior of chiral units were investigated. The regions of the chiral units prone to prematurely fail under crush loads were determined, and deformation modes were investigated to increase the energy absorption abilities.

Empirical mode decomposition-refined composite multiscale dispersion entropy analysis and its application to geophysical well log data

Oil and gas exploration activities often face great challenges due to the nonlinear behavior of the reservoir's physical properties, which is commonly defined as “heterogeneity”. Currently, well log data analysis techniques are a novel approach to unravel such nonlinear behavior, because well log data incorporates considerable geological information that determines reservoir property. However, the current complexity analysis techniques face two challenges: 1) the spatiotemporal multiscale nature of complex geological systems and 2) the superposition of the trends in the geophysical well log data on the analysis results. To fill the research gap, we propose an empirical mode decomposition-refined composite multiscale dispersion entropy analysis (EMD-RCMDEA) to eradicate trends and obtain the complexity results with spatiotemporal characteristics. The proposed method produces more accurate results, and its effectiveness, stability, and efficiency are also verified by the simulation signals and the gamma-ray (GR) well log signals. Compared to previous refined composite multiscale entropy analysis (RCMSEA), the EMD-RCMDEA enhances the stability by 69.3% and efficiency by 53.5%. Additionally, using the GR well log data for reservoirs, this method is also applied to explore the heterogeneity of strata with diverse depositional environments and different composite patterns and acquire the following results. 1) The EMD-RCMDEA values of the GR well log data are positively correlated with the heterogeneity of the strata. 2) The reservoir developed in a delta-front depositional environment has the strongest heterogeneity. 3) The heterogeneity of the composite patterns is much stronger than that of the single heterogeneity patterns. 4) Among the heterogeneities of the composite patterns, the pattern consisting of different facies is stronger than that for single facies. • A novel heterogeneity evaluation method is proposed based on EMD-RCMDEA. • New metrics for evaluating the stability of the multiscale entropy method are proposed. • The EMD-RCMDEA eliminates the drawback of multiscale entropy methods. • Trends superimposed in signals affect complexity analysis results. • The EMD-RCMDEA values of GR well log data positively correlate with reservoir heterogeneity.

Characteristics of the interface acoustoelectric conversion waveform in fluid-filled boreholes

When acoustic waves meet the side wall media in a fluid-filled borehole, the electromagnetic field at the borehole interface between the formation and borehole fluid called the interface acoustoelectric conversion wave signal is converted. The propagation of the refracted P wave, refracted S wave, Stoneley wave, and other borehole-mode waves in the fluid-filled borehole generates accompanying electromagnetic wave signals corresponding to various borehole-mode waves, called accompanying acoustoelectric conversion wave signals. In this study, we examined the interface acoustoelectric conversion wave. First, we numerically simulated the acoustoelectric wave fields in fluid-filled boreholes using a method of real axis integration and analyzed the effects of the borehole radius, porosity, and pore fluid salinity on the interface acoustoelectric conversion waves. The arrival time of the interface acoustoelectric conversion wave increased as the borehole radius increased, and the amplitude of the interface acoustoelectric conversion wave increased as the porosity increased and decreased as the pore fluid salinity increased. Using the self-developed acoustoelectric effect logging tool, we conducted an actual downhole test in a sand shale well in northern China and observed the interface and accompanying acoustoelectric conversion waves. The observations followed the acoustoelectric effect logging theory. Further, to characterize the interface acoustoelectric conversion wave, we calculated its arrival time using a method of long-short time window energy ratio and calculated its energy by taking the summation of squares of the waveform amplitude in a window, respectively. The arrival time approximated to the time in which the acoustic wave propagated from the radiating surface of the acoustic radiator to the side wall interface, and the energy was consistent with the energy ratio of the accompanying conversion Stoneley wave to the acoustic Stoneley wave. The energy curve of the interface acoustoelectric conversion wave and conventional logging curves were highly correlated. The interface acoustoelectric conversion wave in acoustoelectric logging signals can provide a new reference for evaluating the formation parameters associated with open-hole wells.

MODERN ALGORITHMS AND SOFTWARE FOR INTERPRETATION OF RESISTIVITY LOGGING DATA

The electrodynamics of geological media investigates the interrelations of resistivity logging signals and properties of fluid-containing rocks and creates innovative well logging technologies. Its development is inextricably linked with modern techniques for mathematical modeling and quantitative interpretation of high-precision data. In order to increase the information content of galvanic and electromagnetic logging, we have developed algorithms and software for numerical simulation and inversion of field data. In our study of the Cretaceous and Jurassic deposits of West Siberia, a quantitative interpretation of high-frequency electromagnetic and lateral logging signals was carried out. To create geoelectric models, we interpreted the field resistivity logging data by an unconventional quantitative technique based on their joint numerical inversion and estimations of the vertical resistivity of permeable deposits. Another line of our research was aimed at a scientific substantiation of a new technology for mapping and spatial tracking of lateral heterogeneities and oil-promising zones in the Bazhenov Formation. The aim was achieved by using the TEM sounding data on a spatially distributed system of directional and horizontal wells.

Influence of Recorded Pattern on Background Interference impact in Magnetic Recording

This paper discusses the impact of background interference on a recorded pattern for heat-assisted magnetic recording technology (HAMR). Several patterns of the background track were examined, with the log bit error rate and signal to noise ratio measured via a spin-stand tester using HAMR head and media. It was found that the low frequency pattern gave the highest BER and SNR loss due to the strong magnetic field from the adjacent tracks. Similar to its practical use, the PRBS pattern also showed high interference. These observations may be used to support HDD areal density growth.

FEATURES OF JOINT INVERSION OF NON-SIMULTANEOUS SIGNALS OF THE RUSSIAN LATERAL SOUNDING AND ELECTROMAGNETIC LOGGING IN CRETACEOUS RESERVOIRS

The article presents the results of the study of the effect of the non-simultaneous measurement of electrical and electromagnetic logging signals on the determination of the parameters of the geoelectric model by joint numerical inversion. The radial profile of the electrical resistivity is calculated for different times after drilling in the program for modeling the process of mud invasion into a porous permeable formation with parameters characteristic of the BS cretaceous reservoir. For this continuous resistivity profile electric log signals are calculated for the reservoir of unlimited thickness.

Tracking small-scale tropical forest disturbances: Fusing the Landsat and Sentinel-2 data record

Information on forest disturbance is crucial for tropical forest management and global carbon cycle analysis. The long-term collection of data from the Landsat missions provides some of the most valuable information for understanding the processes of global tropical forest disturbance. However, there are substantial uncertainties in the estimation of non-mechanized, small-scale (i.e., small area) clearings in tropical forests with Landsat series images. Because the appearance of small-scale openings in a tropical tree canopy are often ephemeral due to fast-growing vegetation, and because clouds are frequent in tropical regions, it is challenging for Landsat images to capture the logging signal. Moreover, the spatial resolution of Landsat images is typically too coarse to represent spatial details about small-scale clearings. In this paper, by fusing all available Landsat and Sentinel-2 images, we proposed a method to improve the tracking of small-scale tropical forest disturbance history with both fine spatial and temporal resolutions. First, yearly composited Landsat and Sentinel-2 self-referenced normalized burn ratio (rNBR) vegetation index images were calculated from all available Landsat-7/8 and Sentinel-2 scenes during 2016–2019. Second, a deep-learning based downscaling method was used to predict fine resolution (10 m) rNBR images from the annual coarse resolution (30 m) Landsat rNBR images. Third, given the baseline Landsat forest map in 2015, the generated fine-resolution Landsat rNBR images and original Sentinel-2 rNBR images were fused to produce the 10 m forest disturbance map for the period 2016–2019. From data comparison and evaluation, it was demonstrated that the deep-learning based downscaling method can produce fine-resolution Landsat rNBR images and forest disturbance maps that contain substantial spatial detail. In addition, by fusing downscaled fine-resolution Landsat rNBR images and original Sentinel-2 rNBR images, it was possible to produce state-of-the-art forest disturbance maps with OA values more than 87% and 96% for the small and large study areas, and detected 11% to 21% more disturbed areas than either the Sentinel-2 or Landsat-7/8 time-series alone. We found that 1.42% of the disturbed areas indentified during 2016–2019 experienced multiple forest disturbances. The method has great potential to enhance work undertaken in relation to major policies such as the reducing emissions from deforestation and forest degradation (REDD+) programmes.

A Plio-Pleistocene (c. 0–4 Ma) cyclostratigraphy for IODP Site U1478 (Mozambique Channel, SW Indian Ocean): Exploring an offshore record of paleoclimate and ecosystem variability in SE Africa

The paleoclimate and ecosystem variability in Africa during the Plio/Pleistocene has received considerable attention due to its potential links to hominid evolution. However, the reconstruction of this variability hinges critically upon highly temporally resolved proxy data from continuous, well-dated sediment archives. In light of these requirements we use a new XRF core-scanning record from International Ocean Discovery Program (IODP) Site U1478 off the Limpopo River mouth (Mozambique Channel, SW Indian Ocean) spanning the past c. 4 Ma to identify the climate variability in SE Africa. Our results show that the elemental distribution in the Site U1478 cores is mainly controlled by the rate of terrigenous input and – to a lesser extent – by bottom-current transport and post-depositional processes such as propagation of paleoredox boundaries and diagenesis across some intervals. The log(Ti/Ca) ratio, which is used as a tracer of terrigenous sediment input, shows quasi-cyclical variability across the entire record that closely matches the periods of orbital parameters. However, the cyclical behaviour of the log(Ti/Ca) signal varies through time, with the uppermost 106 m of the sequence (0–1.07 Ma) displaying a mix of precession and obliquity signals, the intervals 106–223 m (1.07–2.80 Ma) and 240–257 m (3.68–4.05 Ma) being dominated by precession, and the interval 223–240 m (2.80–3.68 Ma) being controlled by eccentricity. To refine the available chronology for Site U1478, which is based on shipboard biostratigraphic and paleomagnetic data, we have tuned the log(Ti/Ca) record to the LR04 benthic oxygen isotope record, summer insolation at 25° S, and orbital eccentricity depending on the dominant cyclicities in the XRF dataset across individual time intervals. The resulting chronology enables us to evaluate the XRF data as well as the previously available shipboard sedimentological and geochemical datasets within a regional and global climatic context. This allows the connection of a c. 7-m-thick contourite deposit and a prominent paleoredox boundary to hydroclimate and ocean-circulation changes during the early Pleistocene and across the Mid-Pleistocene Transition, respectively. Moreover, a decoupling of the log(Ti/Ca) and the log(Ti/K) records, with the latter indicating the degree of sediment weathering, from 3.2 to 2.8 Ma points to an increased delivery of highly weathered sediments to Site U1478. We attribute this to temporarily wetter and warmer conditions in the catchment of the Limpopo River and/or a change in the sediment source, perhaps associated with the tectonically driven enlargement of the Zambezi River catchment during the late Pliocene/early Pleistocene.

Development and Validation of a Novel Interpretation Algorithm for Enhanced Resolution of Well Logging Signals

This work presents a novel algorithm that achieves enhanced resolution of well logging signals, e.g., from 1 ft of a pulsed neutron mineralogy tool to 0.04 ft of an imaging tool. The algorithm, denoted as “Digital Core,” combines mineralogical and sedimentological information to generate a high‐resolution record of the formation mineralogy which can be consequently applied to thin bedded environments. The keystone to the philosophy of this algorithm is that the spectral information recorded by mineralogy tool is a weighted average of the mineralogy of each lithological component in the analyzed volume. Therefore, by using a high‐resolution image log to determine the proportion of each lithological component, their composition can be determined from the mineralogy log data. A field case from a well located in South Australia is presented in this work, and the results validate the feasibility of an integrated core‐level petrophysical analysis in a cost‐effective and timely manner compared to conventional core measurements.

Improving facies prediction by combining supervised and unsupervised learning methods

Facies classification from well logs is an indispensable part of seismic interpretation and is important in the determination of sequence stratigraphy and ultimately reservoir characterization. Although there have been improvements in the tools used to perform this task, it remains laborious, subjective, and error-prone. Achieving a proper classification is complicated by increasing dataset sizes as well as the need for correlated multidisciplinary models. Recent developments in machine learning provide an opportunity to assist interpreters in accomplishing this task while also improving the accuracy of classification results. Applications of machine learning methods for automating facies classification from well logs have previously been explored, however these have largely focused on evaluations or comparisons of individual algorithms or of ensembles of homogeneous agents. The proposed method combines heterogeneous agents to enhance prediction accuracy. Specifically, supervised learning, which provides a direct mapping between the data domain and the solution domain while introducing bias to generalize the mapping, is combined with unsupervised learning, which does not depend on similar generalization bias or training data but also does not provide a direct mapping between the data and solution domains. The combination is accomplished via the joint probability density function (PDF) of the supervised classification, which is used to guide identification of clusters delineated by unsupervised learning. This multi-agent approach can reduce bias introduced during training and provides a basis for generating a probability distribution for each sample rather than a discrete classification. The distribution, in turn, can be used to more accurately model the continuous nature of well log signals, which reflects continuity in lithological regimes.

Evaluation of the uncertainties in two-dimensional inversion of the Russian electric logging signals using statistical modeling

Based on previously presented Russian electric logging measurements inversion system AlondraWL, we developed software prototype for estimating one- and two-dimensional inversion results uncertainties suitable for galvanic and induction sounding measurements. The algorithm is based on solving a statistically significant number of inverse problems from various initial models in the automatic mode. The software was tested on synthetic and practical high-frequency electromagnetic and galvanic practical logs. The test results show a significant advantage of the developed algorithm over the approach based on the analysis of sensitivity functions.

On the use of CPTs in stratigraphy: recent observations and some illustrative cases

During processing and interpretation of about 200 Cone Penetration Tests (CPT) with electrical cones in a limited area of 60 km² in NE Belgium, a close analogy was observed between the vertical trend and signature of two CPT derived parameters (normalized soil behavior index Ic and hydraulic conductivity kSBTn) and borehole gamma ray (GR) measurements. This close analogy could be repeatably observed on five locations with a CPT executed along a cored and logged borehole. Using this feature proves to be a considerably help for stratigraphic interpretation, especially for sandy units displaying low lithological contrast. Lithostratigraphic units and their boundaries are determined by the specific log response patterns of these CPT derived parameters (kSBTn and/or Ic). These log responses prove to be more consistent and better to correlate than for the classical parameters (cone resistance qc and friction ratio Rf). This paper describes the case of a Neogene setting in the NE of Belgium where this feature was first observed. An informal stratigraphy based on CPT log signal is proposed for the geological units for the research area. Some other cases are briefly described to illustrate the usability of this feature for other geological settings or issues.

Identifying the lithology of volcanic rocks by using the time-frequency features of array acoustic logging data

The traditional acoustic logging signal processing method is computing the slowness of each component wave by time-domain or frequency-domain methods. But both of the two methods are limited. To combine the signals’ times, frequencies, or amplitudes, we have analyzed the array acoustic logging signals by the fractional Fourier transform and the Choi-Williams distribution. First, we apply the fractional Fourier transform on an array acoustic logging waveform with proper [Formula: see text], then the Choi-Williams distribution analysis method is used to process the signal in the fractional Fourier domain, and finally the result will show in the fractional Fourier time-frequency domain. The results show the following. The array acoustic logging signal is received earlier in the mudstone and diabase formation than in the tuff and breccia formations. The basic frequencies of the compressional wave (P-wave) are not very different, but the basic frequency of the shear wave (S-wave) is highest in the tuff formation and is lowest in the diabase formation. The relative energies of each component wave in the diabase, mudstone, tuff, and breccia formation can be summarized as: for the P-wave, diabase > mudstone ≈ tuff ≈ breccia; for the S-wave, diabase ≈ mudstone > breccia > tuff; and for the Stoneley wave, diabase > mudstone > tuff > breccia. The signal processing method combining the fractional Fourier transform and the Choi-Williams distribution can comprehensively research the time, frequency, and amplitude, thereby improving the segmentation of the time and frequency domains and providing a new method for interpretation of array acoustic logging.

Comparison and research of multidimensional analysis for array acoustic logging in fractured formations

Currently, array acoustic logging data processing is mainly achieved by calculating only the velocity variation of each component wave. This provides a little formation information; thus, it is difficult to identify all fractures. We have analyzed the array acoustic logging signals by combining their times, frequencies, or amplitudes. First, we decomposed the array acoustic logging signals into main and secondary signals by improving the singular value decomposition. We then analyzed the characteristics of the array acoustic logging signals in fractured formations and compared them using three multidimensional analysis methods: the 2D Fourier transform, the 2D wavelet transform, and the Choi-Williams distribution. The results showed that the frequency characteristics of the compressional waves (P-waves), shear waves (S-waves), and Stoneley waves in the [Formula: see text]-[Formula: see text] domain of the fracture formations can be obtained by the 2D Fourier transform method. The multiresolution time characteristics of these waves can be realized by the 2D wavelet transform method, and their time-frequency characteristics can be observed via the Choi-Williams distribution. The times, frequencies, and amplitudes of the P-waves, S-waves, and Stoneley waves change from the tight formation to the fractured formation. The three methods introduced here can comprehensively use eight arrays, time, frequency, or amplitude to improve the information usage rate of array acoustic logging and provide new ways to interpret it.

Estimating reservoir permeability with borehole radar

In oil drilling, mud filtrate penetrates into porous formations and alters the compositions and properties of the pore fluids. This disturbs the logging signals and brings errors to reservoir evaluation. Drilling and logging engineers therefore deem mud invasion as undesired and attempt to eliminate its adverse effects. However, the mud-contaminated formation carries valuable information, notably with regard to its hydraulic properties. Typically, the invasion depth critically depends on the formation porosity and permeability. Therefore, if adequately characterized, mud invasion effects could be used for reservoir evaluation. To pursue this objective, we have applied borehole radar to measure mud invasion depth considering its high radial spatial resolution compared with conventional logging tools, which then allows us to estimate the reservoir permeability based on the acquired invasion depth. We investigate the feasibility of this strategy numerically through coupled electromagnetic and fluid modeling in an oil-bearing layer drilled using freshwater-based mud. Time-lapse logging is simulated to extract the signals reflected from the invasion front, and a dual-offset downhole antenna mode enables time-to-depth conversion to determine the invasion depth. Based on drilling, coring, and logging data, a quantitative interpretation chart is established, mapping the porosity, permeability, and initial water saturation into the invasion depth. The estimated permeability is in a good agreement with the actual formation permeability. Our results therefore suggest that borehole radar has significant potential to estimate permeability through mud invasion effects.

Research on Analog Signal Source of Gamma Ray Spectrum Logging Technology

Gamma ray spectrum logging tool acquires reservoir information by detecting gamma ray from formation. It has strict requirements for preservation and use, which is not convenient for college teaching. This paper introduces the design of an analog signal source for gamma ray spectrum logging tool, which generates random number based acceptance-rejection method. These random numbers are converted into electronic pulse signal by data acquisition system. It can simulate the electric pulse signal obtained from gamma ray through stratum. The statistical law of energy spectrum from simulation signal source accords with the probability distribution law of pulse signal generated gamma ray from formation. The analog signal source of gamma ray spectrum logging tool can generate simulation gamma ray logging signal without formation model and radiation source, which is of great value for efficient teaching.

A Comparative Study of Five Networks for Reservoir Classification Based on Geophysical Logging Signals

Unconventional reservoir classification suffers low accuracy because of the complex geophysical properties. With good performance and moderate cost, geophysical logging is considered to be of great potential as the compromise solution between seismic and core. Since the “sweet-spot” depending on the sands' properties are implied not in the logging “point” but “segment”, 1D-CNN is supposed to be a better fit for reservoir classification: It doesn't pay too much attention to the logging signal sequence itself like LSTM, but focuses on feature extraction from the whole signals combination. Moreover, samples would have various sizes because of different layer thickness. While 1D-CNN requires all samples must be converted to a uniform size before input. 1D fully convolutional network (1D-FCN) can receive any size input of layer thickness without manually aligning it to the same size. For the first time in this paper, structures of the five networks including two fully connected networks (global mapping artificial neural network, GM-ANN and point-to-point mapping artificial neural network, PPM-ANN), two fully convolutional networks (1D fully convolutional network with decision-level fusion, 1D-FCN-DEF and 1D fully convolutional network with data-level fusion, 1D-FCN-DAF) and the common 1D convolutional neural network (1D-CNN) are compared and evaluated the suitability for processing logging data in detail. Results on tight gas in Ordos basin of China illustrated by receiver operating characteristics (ROC) curves show that 1D-FCN-DEF and 1D-FCN-DAF have achieved the higher area under the curve (AUC) with the values of 0.8889 and 0.9107 respectively in average comparing to 0.7515, 0.8006 and 0.8364 of GM-ANN, PPM-ANN and 1D-CNN. Case study proves that 1D-FCNs are more accurate than other networks. This paper provides a suitable new idea for logging interpretation and expands the application scope of DL in geophysical logging signals processing.

Classification of Well Log Data Using Vanishing Component Analysis

This study reports the application of the novel supervised learning approach called vanishing component analysis (VCA) for the classification of lithologies from well log signal data. Geophysical well log data is always non-linear due to anisotropy and heterogeneity of the earth. The main purpose of this study is to test the applicability of the VCA algorithm on non-linear geophysical data of Siraj South-01, Middle Indus Basin, Pakistan for classification of lithologies/facies. We demonstrate the performance and stability of the novel approach on a case study before applying it on well log data. Our analysis demonstrates that VCA algorithm is able to linearly separate such a complex non-linear well log data and clearly distinguish between different classes of well log data coming from different rock units. Furthermore, we show that the average accuracies of the classification methods of linear support vector machines, eXtreme gradient boosting, random forest, neural network and linear discriminant analysis on the VCA feature space are much better than the average accuracy obtained by the same methods on the original data.