Geophysical Logs Research Articles

Machine learning algorithms for real-time coal recognition using monitor-while-drilling data

ABSTRACT Accurate coal seam identification is crucial in coal mining to prevent resource wastage and potential damage to coal seams from misplaced explosives. The current industry standard involves drilling past the seam and refilling the hole, a resource-intensive process. Manual seam detection is error-prone, and geophysical logging, performed for only a subset of drill holes, is costly and time-consuming. Monitor-While-Drilling (MWD) data captures drill response metrics like rotary speed and torque, influenced by local geology. These MWD measurements offer insights into geology, including hardness and rock type; They can be used for real-time rock recognition using advanced artificial intelligence techniques. This study focuses on developing tools for precise coal recognition and identification of the top of coal seams using MWD data. Several Machine Learning classifiers are employed, each providing unique data interpretations, and their results are integrated into a more reliable prediction. An artificial neural network is used for rock density regression, which is then used to correct depth offset between geophysical loggings and drill MWD data. The research demonstrates that MWD data can enable real-time coal seam identification, reducing the reliance on time-consuming and expensive geophysical logging. The integrated model accurately identifies the top of coal seams within a ± 20 cm margin.

Stress state estimation – new data and variability assessment of model results

Abstract. The upper crust of the Earth is used more and more to transport and extract raw materials and energy. It is also used for the final disposal of radioactive waste in deep geological repositories with a particularly crucial requirement for long-term safety over 1×106 years. An accurate knowledge of the hosting rock and surrounding formation coupled with comprehensive modelling is fundamental to demonstrate the geological site has the required properties for safe and long-term underground storage. Besides other criteria, geomechanics play an important role. Even if the determination of rock properties is already a challenge, the estimation of the contemporary stress state in the upper crust is even more challenging. To predict the stress state 3-D geomechanical–numerical models are used, but high-quality underground structure, rock properties and in situ stress data for model calibration are needed. In particular, data of maximum and minimum horizontal stress magnitudes (SHmax and Shmin) are required to find suitable lateral displacement boundary conditions to produce a best-fit model. During the recent exploration phase for a deep geological repository for radioactive waste in Switzerland, a unique dataset of stress magnitude data has been acquired from eight cored boreholes. Rock mechanical properties were constrained from geophysical logging and laboratory testing. The empirically correlated rock properties were not simply averaged, but a probability distribution was provided. The stress field was explored by conducting more than 120 micro-hydraulic fracturing (MHF) and dry sleeve re-opening (SR) tests in different stratigraphic units, to estimate the magnitudes of Shmin and SHmax. We present the results of a 3-D geomechanical–numerical model that shows the best fit with respect to the measured stress magnitudes. Considering the uncertainties of the MHF/SR tests and the ones resulting from rock property variability, the model can reproduce most of the measurements. However, we do show not only the best-fit result but also a bandwidth of individual stress components within a P05–P95 probability range. We also show that there are still outliers for this bandwidth, but we will show that these can be explained by scaling effects.

A 3D thermal conductivity prediction method for deep strata based on temperature field estimation

Accurate predictions of earth’s interior thermal conductivity are essential for interpreting the mechanisms involved in geothermal systems and delineating geothermal resource targets. Existing thermal conductivity prediction methods, which typically involve measurements of rock samples and predictions based on geophysical well logging, cannot accurately predict regional thermal conductivity distributions. We develop a 3D thermal conductivity prediction based on temperature field estimation. First, the finite-element method is used for temperature forward modeling. Subsequently, the inexact Gauss-Newton method is implemented in the prediction algorithm. Synthetic studies indicate that the method can recover the true model well. The subsurface temperature distribution of the Xiong’an New Area is predicted using the coefficient correction method of the optimal temperature. Next, we obtain the 3D distribution of thermal conductivity of deep formation in the Xiong’an New Area with the predicted temperature field. The distribution of terrestrial heat flow also is estimated. In addition, the law of thermal accumulation of strata at different depths is analyzed, and the internal relation among resistivity, temperature, and thermal conductivity is studied. Thus, the 3D thermal conductivity prediction method provides the theoretical basis for interpreting mechanisms for the origin of geothermal systems and heat accumulation patterns. It could have significant utility for the analysis of geothermal fields and the sustainable development and use of geothermal resources.

Secondary porosity and pore aspect ratios integrated with permeability and FMI logs to characterize flow zones at a Port Mayaca aquifer in South Florida

The complex pore structure of carbonate aquifers presents a challenge to interpreters analyzing geophysical logs and geologic data. A significant task is to develop physical rock models to determine the microstructure that provides information about flow-zone paths within the aquifer. In an attempt to achieve this task, an algorithm is devised to predict the secondary porosity formed by stiff macropores, compliant micropores, touching vugs, and pore aspect ratios from sonic logs. The pore aspect ratios are classified in intervals that delineate permeable and low permeability zones of aquifers. This provides information about the presence of isolated vugs, which do not contribute to the flow and connect or touch vugs that are part of the flow zones. The inversion results determined that permeable channels have pore aspect ratios 0.01–0.2. Alternatively, vugs with aspect ratios 0.5–1 are not forming permeable paths because they are isolated and are not contributing to flow zones. The inversion of P- and S-wave velocity logs using density and total porosity logs obtains the secondary porosity and pore aspect ratios. It found optimum correlation coefficients of 0.9975 for S-wave and 0.9405 for P-wave velocities by constraining the solution with the natural relation of the total porosity versus primary plus secondary porosities. The Port Mayaca aquifer includes Stoneley-wave permeability, formation microimager logs, and microresistivity logs, which together with pore aspect ratio and secondary porosity logs delineate and characterize the flow zones. In addition, data integration demonstrates that the vug porosity log detects fractures in resistive, permeable, and dense cemented zones. This new finding creates vug signatures, with their maximums resembling the shapes of the corresponding resistivity heights in the microresistivity log. In conclusion, it is shown that the anomaly zones correspond to water production regions, and their presence is confirmed with flow meter logs.

Hydraulic Tomography Estimates Improved by Zonal Information From the Clustering of Geophysical Survey Data

AbstractHydraulic tomography (HT) has been demonstrated as a robust approach to characterize subsurface heterogeneity through the inverse modeling of multiple pumping data. However, smooth or even erroneous tomograms can result when insufficient observations are involved in the inversion. In this study, the feasibility of integrating geophysical survey data into HT analysis is investigated. First, k‐means clustering is utilized to extract zonal information from borehole geophysical logs, and a new type of spatial constraints containing geological knowledge is proposed to obtain improved hydrostratigraphic boundaries along boreholes. Next, zonation models are constructed by applying clustering‐based zone geometry and populating zonal estimates of hydraulic conductivity (K) from analyzing pumping data. Afterwards, zonation models are treated as the initial guess of spatial variability in the geostatistical inversion of HT analysis. Additionally, local K measurements can be utilized to further improve HT estimates. Comparative cases of HT analyses are designed for a numerical sandbox experiment to highlight the HT performance integrated with geophysical surveys, in which the geostatistical inversion is initialized with: (a) a homogeneous K field; (b) zonation models built by the clustering of disparate geophysical surveys with/without spatial constraints; and (c) zonation improved by incorporating local K measurements. Based on ln K field comparisons and validation through predictions of drawdowns and tracer plume migration from independent tests not used in the calibration effort, we find that integration of geophysical surveys into HT analysis by clustering with spatial constraints is demonstrated as an effective approach, and local K measurements can further improve HT estimates.

New Insights into Fracture Porosity Estimations Using Machine Learning and Advanced Logging Tools

Fracture porosity is crucial for storage and production efficiency in fractured tight reservoirs. Geophysical image logs using resistivity measurements have traditionally been used for fracture characterization. This study aims to develop a novel, hybrid machine-learning method to predict fracture porosity using conventional well logs in the Ahnet field, Algeria. Initially, we explored an Artificial Neural Network (ANN) model for regression analysis. To overcome the limitations of ANN, we proposed a hybrid model combining Support Vector Machine (SVM) classification and ANN regression, resulting in improved fracture porosity predictions. The models were tested against logging data by combining the Machine Learning approach with advanced logging tools recorded in two wells. In this context, we used electrical image logs and the dipole acoustic tool, which allowed us to identify 404 open fractures and 231 closed fractures and, consequently, to assess the fracture porosity. The results were then fed into two machine-learning algorithms. Pure Artificial Neural Networks and hybrid models were used to obtain comprehensive results, which were subsequently tested to check the accuracy of the models. The outputs obtained from the two methods demonstrate that the hybridized model has a lower Root Mean Square Error (RMSE) than pure ANN. The results of our approach strongly suggest that incorporating hybridized machine learning algorithms into fracture porosity estimations can contribute to the development of more trustworthy static reservoir models in simulation programs. Finally, the combination of Machine Learning (ML) and well log analysis made it possible to reliably estimate fracture porosity in the Ahnet field in Algeria, where, in many places, advanced logging data are absent or expensive.

Effect of spatial variability of downhole geophysical logs on machine learning exercises

Machine Learning (ML) based estimation of petrophysical properties from downhole geophysical logs has attracted a significant attention from across the geoscience community in recent years. The main focus of the studies in the area is to identify the most accurate ML algorithm for petrophysical property prediction using logging data often from a very limited number of boreholes (e.g. only one or two). Despite having vertical variability in one borehole, the horizontal variability cannot be captured using limited number of boreholes thus the effect of spatial variability on ML based models' performance needs a specific attention. We have therefore developed an innovative workflow that assesses the impact of spatial variation in training geophysical logs data on the accuracy of ML models in predicting formation properties e.g. compressional wave velocity. To build this systematic workflow, a large dataset containing geophysical logs from 50 boreholes in a coal mine in Bowen Basin, QLD, Australia was selected for the analysis. Boreholes mainly intersected geological formations from Rangal coal measure rocks. The data included density, compressional wave velocity (VP), gamma ray, resistivity and depth. The workflow consists of i) a clustering analysis using the K-mean algorithm to determine if the data could be partitioned into smaller, more homogeneous clusters based on density, gamma ray and measured depth, ii) putting aside five boreholes for validation (these boreholes are not seen by the models, named test boreholes) and iii) selectin training datasets randomly from the remaining boreholes grouping them into 1, 5, 10, 20, 30 and 40 boreholes. The data selection from each group was repeated randomly five times to capture the potential randomness; and iv) the Least Squares Support Vector Regression model of each group was trained with 10, 20, 30, …, 90 and 100% of the data of each group forming 300 predictive models. The accuracy of developed models to predict VP from other logs was evaluated by calculating R2 value comparing the estimated and the actual VP values for five test boreholes. The results from the proposed workflow interestingly suggest that the model should prioritize incorporating data from various boreholes rather than collecting more data points from a limited number of boreholes. This observation is mainly linked to high lateral variability in the study area. Consequently, utilizing a smaller portion of data from a larger number of boreholes yields a simpler yet more accurate predictive model, which also reduces the computational cost.

Application of audio-frequency magnetotelluric data to cover characterisation – validation against borehole petrophysics in the East Tennant region, Northern Australia

The characterisation of the thickness and geology of cover sequences significantly improves targeting for mineral exploration in buried terrains. Audio-frequency Magnetotelluric (AMT) data is applicable to characterise cover sequences, where their conductivity (inverse resistivity) can be differentiated. We present a regional study from the under-cover East Tennant region in the Northern Territory (Australia) where we have applied deterministic and probabilistic inversion methods to derive 2D and 1D resistivity models. We integrated these models with information of co-located basement penetrating boreholes (lithological and geophysical logs) to ground-truth and validate the models and to improve geophysical interpretations. In the East Tennant region, borehole lithology and wireline logging demonstrate that the modelled AMT response is largely controlled by the mineralogy of the cover and basement rocks. The bulk conductivity is due primarily to bulk mineralogy and the success of using the AMT models to predict cover thickness is shown to be dependent on whether the bulk mineralogy of cover and basement rocks are sufficiently different to provide a detectable conductivity contrast. Our investigation of a range of geological scenarios that differ in thickness, complexity and geology of the cover and basement rocks suggests that in areas where there is sufficient difference in bulk mineralogy and where the stratigraphy is relatively simple, AMT models predict the cover thickness with high certainty. In more complex scenarios interpretation of AMT models may be more ambiguous and requires integration with other data (e.g. drilling, wireline logging, potential field modelling). Overall, we conclude that the application of the method has been validated and the results compare favourably with borehole stratigraphy logs once geological (i.e. bulk mineralogical) complexity is understood. This demonstrates that the method is capable of identifying major litho-stratigraphic units with resistivity contrasts. Our results have assisted with the planning of regional drilling programs and have helped to reduce the uncertainty and risk associated with intersecting targeted stratigraphic units in covered terrains.

Geophysical characteristics of a fault system in the northern Central Range of Taiwan and its applications for geothermal energy exploration

The northern Central Range of Taiwan is a high-potential geothermal region. Since the formations are mainly tight metasandstone and slate, permeable structures associated with faults are commonly considered as conduits of geothermal fluids. This study determines the characteristics and orientations of the permeable fault zones by analyzing the geophysical logs and microresistivity formation image log (FMI) of the JT-4 well in Jentse, an important geothermal area in the northern Central Range. Between 720 and 1480 m measured depth (MD), the effective porosity of the intact host rock is mostly below 3% calculated by the geophysical log. Zones with porosity greater than 5% are only clustered within a few thin intervals. The FMI interpretations show these porous zones are in the interior of the fractured and faulted intervals. These porous fault zones comprise fault damage zones with a high density of open fracture planes and fault cores with porous fault breccias. There is a highly brecciated fault core in 1334–1339 m MD, which would be the most permeable interval of the well. Additionally, some healed fault zones with sealed fractures are observed. The picked drilling-induced tensile fractures signify that the direction of the present-day maximum horizontal principal stress is N40–50°E, and most of the open fractures also strike parallel to the NE–SW direction. The study results show that the open fractures are concentrated in the four fault zones belonging to one major normal fault system. After integrating the orientations and locations of the fault zones, we propose that the permeable normal fault system is about 200 m wide, trends N50–70°E, and dips 70–80° to the NW. The development of the open fractures and the permeable fault system in the northern Central Range may be controlled by the current rifting of the Okinawa Trough offshore northeastern Taiwan. The study exhibits the characteristics of fractured fluid conduits of the regional geothermal system, which will benefit future geothermal exploration in northeastern Taiwan.

Fault Zone Spatial Stress Variations in a Granitic Rock Mass: Revealed by Breakouts Within an Array of Boreholes

AbstractThe in situ stress state within fault zones is technically challenging to characterize. At the Bedretto Underground Laboratory in the Swiss Alps, the breakouts observed in an array of eight inclined boreholes penetrating a fault zone offer a unique opportunity to characterize the fault‐associated spatial stress variations. Synthesizing multiple geophysical logs, natural geologic structures intersecting these boreholes are identified, revealing a hierarchy of a major fault zone along with secondary structures. Within the boreholes, breakout rotations occur over multiple scales, spanning individual fractures and the entire major fault zone. We first estimate and rule out the effect of the fracture‐induced anisotropy on the breakout rotations, which are attributed mainly to the stress variations. Based on the stress field around a circular borehole and Mohr‐Coulomb failure criterion, the observed breakout azimuths are used to invert the stress information. Results show that the stress field outside the fault zone features a stress ratio (quantifying the relative stress magnitude) of about 0.9, an inclined overburden stress (inclination: 12°∼18°), and a maximum horizontal principal stress (SHmax) oriented N100∼120°E. Within the fault zone, a substantial reduction of the stress ratio and complicated stress rotations are constrained, likely induced by the stress drop on local fractures. As a result, less critical stress state inside the major fault zone is expected. Our work provides a semi‐quantitative estimation of the in‐situ stress variations around fault zones in the absence of direct stress measurements, which is beneficial to a number of scientific and engineering applications.

Groundwater Potential Mapping at Highly Populated Sub-districts of Thimphu District Using Fuzzy Analytic Hierarchy Process

Groundwater is important when there is an insufficient spring water supply for daily needs. The population growth and expansion of the cities demand more water. Thimphu is the capital city of Bhutan and facing water shortage in recent times due to urban expansion and population growth. This study aims to find the potential area for groundwater using the Fuzzy Analytic Hierarchy Process. The factors used for this study are elevation, slope, curvature, topographic wetness index, drainage density, normalized difference vegetation index, distance from fault density, rainfall distribution map, soil texture map, and drainage density. The rating for the factors and classes of the factors were carried out with the help of field experts. As per the fuzzy weight, the highest contributing factor for the groundwater is rainfall with a fuzzy weight of 0.215. Only 1.19% of the study area falls under very high potential zone for groundwater and the area under the curve of the groundwater potential map is 70.37%. The groundwater potential map will help concerned officials to explore the subsurface groundwater using geophysical investigation using test drilling and borehole geophysical logging techniques.

Geological structure and occurrence conditions of the Bazhenov formation within THE South Tambey gas condensate field

Background. In the 21st century, exploration works in the previously discovered Jurassic deposits has been launched in the north of the Yamal Peninsula. The South Tambey field was no exception, where 3D seismic surveys and exploration drilling have been carried out over the past 10 years to study of Jurassic sediments and rocks of the Bazhenov formation. Due to the lack of a data sufficient sample in the northern regions of the Western Siberia, formation was poorly studied. The main ideas were formed on wells scattered at great distances from each other without the possibility of a clear correlation. After receiving new drilling materials, it became possible to examine the deposits of the Bazhenov formation in more detail.Aim. Determination of the structural features and occurrence conditions of the Bazhenov formation.Materials and methods. The research was based on the logging data of exploration and production drilling. A detailed correlation of well sections was performed, which was compared and analyzed with the core description and interpretation of seismic survey materials.Results. The lithological characteristics of rocks obtained by core description are presented, the atypical neighborhood of stromatolites and radiolarians in the section is noted. Based on a detailed correlation, the Bazhenov formation was divided into three intervals with characteristic geophysical logging values and distribution features over the area. The clinoform occurrence with development in the north-west direction is determined. Based on the areal variability, it is proposed to distinguish two types of the Bazhenov formation, and approaches to their tracing are considered.Conclusion. The study of the Bazhenov formation revealed a co-directionality of the spatial orientation of the faults, the progradation of the Achimov clinoforms, and changes in the thickness of the Bazhenov formation. A hypothesis has been put forward about a single tectonic regime that prevailed in the Upper Jurassic and Lower Cretaceous. It is proposed to consider flower (keyboard) tectonics as the main factor that determined the structure and occurrence conditions of the Bazhenov formation.

Biostratigraphy v. geophysics; correlation of Middle Turonian chalks in the Anglo-Paris Basin

High-resolution correlations between boreholes are commonly based on a combination of geophysical well-log data and biostratigraphy, and when these conflict resolution can be contentious and difficult. Such a situation exists in the Middle Turonian chalks of the Anglo-Paris Basin, where electrical resistivity logs, registering thin but seemingly laterally continuous beds of clay-rich chalk (marls) in mostly uncored boreholes, provide apparently straightforward correlations across broad regions. In contrast, detailed biostratigraphical analysis using microcrinoids provides very different results, showing the presence of a major hiatus associated with a hardground and differing significantly from the geophysical correlations. Detailed re-examination of this contentious problem, utilizing exposures adjacent to boreholes, demonstrates the supremacy of detailed biostratigraphical analysis, independently supported by regional thickness patterns, cyclostratigraphy and geochemistry. Although closely spaced geophysical well logs can afford convincing correlations, their use in correlation over broader geographical regions leads to erroneous and misleading conclusions.

Seismic monitoring of strength in stabilized foundations by P-wave reflection and downhole geophysical logging for drill borehole core

Seismic monitoring of strength in stabilized foundations by P-wave reflection and downhole geophysical logging for drill borehole core

Seismic monitoring of strength in stabilized foundations by P-wave reflection and downhole geophysical logging for drill borehole core

Abstract Evaluating the subground properties during the initial stage of a construction of building is important in order to estimate the suitability of soil quality to the technical requirements of bearing capacity, resistance to stress, and strength. This study presented the evaluation of the geotechnical properties of soil intended for the construction of Max IV facility of Lund University, performed in fieldwork and laboratory. The in situ methods included drilling boreholes, core sampling and assessment, crosshole measurements, and borehole logging. The laboratory-based measurements were performed at Swedish Geotechnical Institute and combined seismic measurements of drill cores, determination of the uniaxial compressive strength (UCS), and examination of material property: sieve analysis and natural moisture content. UCS was evaluated with regard to velocities of elastic P-waves. The synchronous light test by X-ray diffraction was performed for qualitative analysis of mineral composition of samples. The study applied integrated approach of the diverse geophysical methods to solve practical tasks on the evaluation of foundation strength and geotechnical parameters. This study demonstrated the benefits of integrated seismic and geophysical methods applied to soil exploration in civil engineering for testing quality of foundation materials.

Innovative hydrogeophysical approach for detecting the spatial distribution of hydraulic conductivity in Bahri city, Sudan: A comparative study of Csókás and Heigold methods

Hydraulic conductivity is one of the crucial parameters used to identify the potentiality and productivity of groundwater aquifers. This research employs an integrated approach using geophysical well logging, exploratory factor analysis and surface electrical resistivity methods to detect the vertical and horizontal variation of hydraulic conductivity in Bahri city, Sudan. Based on the geophysical well logs of Spontaneous potential (SP), natural gamma ray (GR), and electrical resistivity (RS), Csókás method is used to determine the continuous variation of hydraulic conductivity along the aquifer. Csókás method is an experimentally modified version of the Kozeny–Carman equation and is based on the formation factor of the groundwater aquifer and the effective grain size. This approach is performed in three groundwater boreholes, and the obtained hydraulic conductivities showed a close agreement with that of the pumping test analysis. Furthermore, the hydraulic conductivity is measured using multivariate statistical factor analysis. This statistical approach relies on the correlation between the extracted factors and petrophysical and hydrogeological parameters. In this research, a strong negative linear correlation between the first factor and hydraulic conductivity is indicated. Consequently, a site-specific equation is suggested for continuous estimation of hydraulic conductivity along the aquifer. In the last stage, the results obtained from the Csókás method are interpolated with vertical electrical sounding (VES) measurements using to detect the horizontal variation of hydraulic conductivity throughout the studied area. This was achieved by combining the hydraulic conductivities of geophysical well logging and vertical electrical soundings to obtain a consistent estimation. As a result, the variation of hydraulic conductivity is obtained, and the average was 1.9 m/day which shows a close agreement with the average of the previous investigations (1.5 m/day). This approach is highly recommended since it can enhance data coverage, cutting down the expense of hydrogeological investigations and lowering the uncertainty of the hydrogeological models.

Petrophysical and hydrogeological characterization of coastal aquifer using geophysical logs in Lekki Peninsula, Lagos, Nigeria

The coastal aquifers of Lekki Peninsula, Lagos are increasingly threatened by saline water invasion. In this study, the characterization of the petrophysical properties of the coastal aquifers in Lekki Peninsular through analysis of geophysical logs was carried out. The methodology involved editing, interpolation, depth shifting and normalization of gamma ray and resistivity logs from twelve water wells. Log analysis was employed to map aquiferous units and determine their petrophysical properties which include clay volume, porosity, permeability, water resistivity, total dissolved solid and water quality. These properties were further subjected to cross plot analysis. The results identified three aquifers. Aquifer 1 is mostly unconfined shallow phreatic aquifer characterized by 11–35% porosity. The TDS ranges from of 4.63 mgL−1 – 50.33 mgL−1. The intermediate Aquifer 2 is characterized by significant thickness (25–75 m), clay volume (0.22–0.54) excellent porosity (16–27%), permeability (99–110 mDarcy), water resistivity (25–175 Ωm) and total dissolved solids of <30.00 mgL−1. The deep Aquifer 3 lies at 153–255 m depth and has total dissolved solids value of 10.63 mgL−1 and 26% porosity. Cluster analysis shows 67% of wells that penetrated Aquifer 2 encountered saline to brackish water while 40% found aquitard with poor petrophysical properties. Water wells in the study area should target aquifer 3 as Aquifer 2 is prone to saltwater intrusion due to over-pumping. This study furnishes information that enables sustainable future exploration, exploitation, monitoring and management of groundwater reserve in Lekki Peninsula.

A borehole porosity prediction method with focusing on local shape

Porosity is a valuable parameter reflecting petroleum storage performance and plays an important role in reservoir exploration. Various studies have confirmed the feasibility of establishing a mapping between geophysical logging and porosity by using machine learning methods. The existing works are generally in the form of depth point-to-point. However, due to the temporal and spatial characteristics of the stratum deposition, the correlation under continuous depths of logging curves should be considered in addition to the absolute values. Attempts have been made to achieve the memory for low-frequency information with continuous depths by introducing the concept of time in recurrent neural networks. This logging porosity prediction strategy has disadvantages due to two reasons: (i) periodicity is present in the logging data, but is weak, and (ii) small-scale information of the logging curves have a guiding meaning on the detail of porosity. In this paper, we present the first attempt to describe the high-frequency variation in the local shape of the logging curves as spatial small-scale features and propose a ConvLSTM-based framework in the porosity prediction task. Specifically, the local shape features of the curves are further extracted by convolutional operations as the contextual information is memorized. We form 1-dimensional (1D) images in a segment-to-point manner (fixing segment to 16 depth points) to make the logging data suitable for convolution operations while increasing the volume of data. Further, segment-to-point samples are automatically intercepted and feed into model with segment-to-segment prediction patterns according to the view window size. Various experiments are carried out on a real dataset from the Shengli Oilfield, Jiyang Depression, and the results validate the effectiveness of ConvLSTM for local shape information extraction and the superiority in porosity prediction.

Pore structure characterization and deliverability prediction of fractured tight glutenite reservoir based on geophysical well logging

Pore structure characterization and deliverability prediction of fractured tight glutenite reservoir based on geophysical well logging

Machine learning - a novel approach to predict the porosity curve using geophysical logs data: An example from the Lower Goru sand reservoir in the Southern Indus Basin, Pakistan

Porosity estimation is one of the essential issues in oil and natural gas industries to evaluate the reservoir characteristics properly. Therefore, it is imperative to predict porosity with the optimum way to reduce logging tests. In this paper, fuzzy logic (FL) and neural networks (NNs) are considered effective approaches to predict the Lower Goru sand reservoir's porosity curve. The input dataset for the study contained four known logs, gamma ray (GR), neutron porosity (NPHI), density (RHOB), and sonic (DT) of five wells drilled. For the fuzzy logic model, ten bins were used. The closeness of fit (Cfit) curves were computed using the most likely and second most likely curves. The weighted average final probability Pi, or the most likely solution, was also calculated. The curve histogram distribution and set of curve bin distribution cross plots were built using a fuzzy model. In the fuzzy logic model, the Gaussian membership function provided the optimum match for the examined geophysical log data. Fuzzy logic models indicate Cfit values ranging from 94 to 100% for Sawan-01, Sawan-02, Sawan-03, Sawan-07, and Sawan-08, with standard deviations of 1.248, 1.241, 1.254, 1.336, and 1.374, respectively. The neural networks model was trained using the backpropagation (BP) algorithm. The neural networks model has a Cfit_nn of 88%–100% across five wells with standard deviations of 0.016, 0.014, 0.015, 0.017, and 0.018. The results show that the predicted modeling evaluations using fuzzy logic and neural networks techniques fit the geophysical log data quite well. The multiple linear regression (MLR) assessments were conducted using the same geophysical log datasets of five studied boreholes for comparison. The coefficients of determination (R2) for the fuzzy logic (PHIT_ml) and neural networks (PHIT_nn) models were 0.960127, and 0.973039, respectively, whereas the values of the PHIT curve for multiple linear regression (PHIT_mlr) 0.926329. The high R2 values show that fuzzy logic and neural networks are more effective methods for PHIT curve prediction than the multiple linear regression approach. The relevant correlation was derived by comparing synthetic log values to actual log values. The evaluations between recorded and predicted values applying the two distinct approaches fuzzy logic and neural networks revealed that both are effective at synthesizing PHIT logs. The confirmation of this efficiency was further verified by the low values obtained in the root mean square error (RMSE) analysis. The study conducted on the Sawan Gas Field wells revealed that both fuzzy logic and neural networks are reliable approaches for predicting the PHIT curve. By using a composite of GR, RHOB, NPHI, and DT logs, these techniques can provide a realistic fit for both actual and synthesized PHIT curves. The findings of this study suggest that the implementation of these methods can contribute to the improvement of hydrocarbon exploration and production in the region by reducing uncertainty in predicting the PHIT curve. Moreover, the methods used in this study have the potential for wider application beyond the Sawan Gas Field. These methods can be applied globally to predict the PHIT curve and evaluate the reservoir prospects. The successful application of fuzzy logic and neural networks in this paper provides a solid foundation for future research on using machine learning techniques in reservoir characterization and modeling.