Conventional Log Data Research Articles

Discussion on the Method of Evaluating Oil and Water Layers by Using Ratio of Deep Resistivity and Flush Zone Resistivity

The depth-shallow resistivity ratio method for the same measuring point is an important method to identify oil and water reservoirs, But its application effect in conventional logging data is not good. By studying the invasion characteristics of resistivity curves of reservoirs, the ratio of deep resistivity and flush zone resistivity at the same measuring point is used to evaluate oil and water layers with a good effect. Further research shows that the ratio of deep resistivity to flushing zone resistivity can reflect the oil-bearing property of reservoir. The fact that the ratio of deep resistivity and flush zone resistivity of water layers in the same well is basically a constant supports this view, which can be deduced from Archie's formula. That is, the ratio of deep resistivity to flushing zone resistivity is Rw/Rmf, which is the ratio of formation water resistivity to mud filtrate resistivity. On this basis, a method of evaluating oil and water reservoir by using the ratio of deep resistivity and flush zone resistivity is proposed. One is logging crossplot evaluation method, the other is apparent water saturation calculation and evaluation method. These two methods are not affected by the changes of lithology and resistivity of the reservoir, and do not need data such as porosity, rock-electro parameters and formation water resistivity, which can quickly identify oil and water layers. As the ratio between the deep resistivity and the resistivity of the flush zone is not related to the resistivity of the reservoir, it can be used to evaluate the low resistivity oil reservoir. The results show that it is effective to identify low resistivity reservoirs and it is a reliable method to evaluate low resistivity reservoirs.

Study on recognition of reservoir fluid based on apparent formation water resistivity spectrum

According to characteristics of carbonate fracture-cave reservoir and applying imaging logging data with high resolution and abundant information, combining calculated cementation exponent of conventional logging data, the thesis calculates qualitati

Pore structure classification and logging evaluation method for carbonate reservoirs: A case study from an oilfield in the Middle East

ABSTRACTBased on the data of 121 thin sections, 24 mercury injection and physical properties of the carbonate reservoir in A oilfield of the Middle East, the reservoir in the study area is divided into four pore-throat systems by analyzing the pore-throat volume verses permeability-contribution curve of core mercury injection and corresponding depth NMR logging data. Taking into account the contribution of each pore-throat system to the rock, a new pore structure parameter P based on NMR logging data is proposed. On this basis, the P and flow porosity calculated from NMR logging data are used as variables, and the pore structure of carbonate reservoirs is divided into four types by using the K-means clustering method in combination with the characteristics of capillary pressure curves and thin sections. With the input of NMR logging data and conventional logging data, the classification model of pore structure is established by Rotation Forest algorithm. The accuracy of the classification model based on NMR logging is 98.56%, and the accuracy of the classification model based on conventional logging is 89.9%. Compared with the Random Forest algorithm and the Fisher discriminant method, the Rotation Forest algorithm has high prediction accuracy and strong stability. The research shows that the pore structure classification method proposed in this paper is in good agreement with the interpretation results, which can provide some reference value for finding effective reservoirs in the future.

India National Gas Hydrate Program Expedition-02: Operational and technical summary

Abstract The India National Gas Hydrate Program is being steered by the government of India's Ministry of Petroleum and Natural Gas (MoPNG) with participation of Directorate General of Hydrocarbons (DGH), Oil and Natural Gas Corporation Limited (ONGC), and the National Oil Companies and Research Institutes of India. The India National Gas Hydrate Program Expedition 01 (NGHP-01) established the presence of gas hydrate in the Krishna Godavari (KG) and Mahanadi Basins and in the offshore area of the Andaman Sea Basin. However, the gas hydrates discovered during NGHP-01 were mainly distributed as fracture-filling material in fine-grained clay-rich sediments. The India National Gas Hydrate Program Expedition 02 (NGHP-02) was carried out with an objective to discover gas hydrate in sand-rich sediment along the eastern offshore margin of India. ONGC planned and executed NGHP-02 on the behalf of the MoPNG. NGHP-02 started on March 3, 2015 and was completed on July 28, 2015 (total 147 days) using the Japanese scientific Drilling Vessel Chikyu (D/V Chikyu). During NGHP-02, 42 holes at 25 sites were drilled, cored, and/or surveyed with downhole logging tools. These sites were located in four areas along the eastern margin of India and formally named Area A (Mahanadi Basin, three sites), Area B (northern part of the KG-Basin, twelve sites), Area C (central part of the KG-Basin, six sites), and Area E (southern part to the KG-Basin, four sites). All 25 sites established during NGHP-02 were first drilled and logged with logging-while-drilling (LWD) tools and an additional 17 holes were then drilled and/or cored with conventional coring tools (HPCS/ESCS) or pressure coring tools (PCTB). Wireline logging was conducted in 10 holes and formation tests using a dual packer Modular Formation Dynamics Tester (MDT) tool were carried out in two holes. The onboard science team used the laboratory facilities on the D/V Chikyu to examine and analyse the physical properties, geochemistry, and sedimentology of all the cores collected during the expedition. Core samples were also analysed in additional post-expedition shore-based studies conducted in numerous domestic and international gas hydrate research laboratories. The NGHP-02 sediment cores were archived at the National Gas Hydrate Core Repository in Mumbai (India), which is associated with the ONGC Gas Hydrate Research and Technology Centre (GHRTC). The necessary data for characterizing the occurrence of gas hydrate, such as interstitial water chlorinities, core-derived gas chemistry, physical and sedimentological properties, thermal images of the recovered cores, pressure core and downhole measured logging data (LWD and/or conventional wireline log data), were obtained from most of the drill sites established during NGHP-02. Almost all the drill sites yielded evidence for the occurrence of gas hydrate; however, the inferred in situ concentration of gas hydrate varied substantially from site to site. For the most part, the interpretation of downhole logging data, core thermal images, interstitial water analyses, and pressure core images from the sites established during NGHP-02 indicate that the occurrence of concentrated gas hydrate is mostly associated with coarser grained (sand-rich) sediments. This paper presents the operational and technical summary of NGHP-02.

Exploring the link between the flow zone indicator and key open-hole log measurements: an application of dimensional analysis

Dimensional analysis has been used as a tool for developing a non-linear empirical model giving the flow zone indicator (FZI) as a function of open-hole log measurements. Dimensional analysis confirms that interval transit time (Δ t ), true resistivity ( R t ), bulk density ( ρ b ), the apparent water resistivity ( R wa ) and the photo-electric absorption ( P e ) are essential well-log measurements for estimating the FZI in sandstone formations. A unique power-law relationship has emerged between a dimensionless FZI group ( λ ) and a dimensionless resistivity group ( Ω ) for distinct hydraulic flow units. Petrotyping, using either the discrete rock type (DRT) approach or the global hydraulic elements (GHE) approach, appears to provide a credible framework for comparative hydraulic flow unit description. Dimensionless groups permitted the analysis of the effect of several operational variables. The FZI increases with increasing bulk density, and with increasing interval transit time. On the other hand, the FZI decreases with increasing true resistivity ( R t ). The relationship between the FZI and the photo-electric absorption is more intricate, though, since it depends on the value of the power-law exponent ϖ . Conventional log data from an oil well (well B), penetrating two distinct sandstone oil reservoirs in an onshore sandstone oilfield in the Middle East, were used to validate the dimensionless groups. For the field case presented, the FZI empirical model prediction capability does not appear to be diminished by the existence of a capillary transition zone. The FZI-based model, used for estimating permeability, represents an improvement in the prediction of permeability provided that reliable estimates of the FZI are available. Dimensional analysis proved to be a powerful modelling tool capable of revealing fine relationships that are invisible to exhaustive data mining techniques.

The effect of carbonate reservoir heterogeneity on Archie's exponents (a and m), an example from Kangan and Dalan gas formations in the central Persian Gulf

Calculating hydrocarbon in place is one of the most important aspects to be considered in reservoir evaluation. Carbonate rocks with high heterogeneity, exhibit significant changes in petrophysical and lithological properties. Archie's exponents (m and a) are basic parameters for saturation calculations. Uncertainties in obtaining these exponents lead to considerable errors in the saturation assessment. In this study, various pore and rock typing methods were carried out on a Permian–Triassic carbonate reservoir in the central Persian Gulf, Iran. The used methods include pore typing, pore facies classification, velocity deviation log, reservoir quality index, Winland R35, Pittman R35 and ranges of core permeability. Archie's exponents were obtained for different classes based on extracted data from petrographical studies, conventional core analysis and wire line log data. Results showed that determination of rock types based on pore typing has the greatest effect on the precise determination of Archie's exponents. On the other hand, the most unreliable results were related to the velocity deviation log. Pore types directly control the connectivity of the fluid pathways and so have the most important effect. Combining pore types in various groups such as velocity deviation or pore facies classification, reduce the measurement accuracy of Archie's exponents and resulted water saturation. Finally, it can be concluded that, in addition to the high impact of pore type on the accuracy of the obtained exponents, the permeability and pore throat radius are less effective in determining the Archie's exponents.

Modeling and inversions of acoustic reflection logging imaging using the combined monopole–dipole measurement mode

In this paper, we theoretically and numerically study a combined monopole–dipole measurement mode to show its capability to overcome the issues encountered in conventional single-well imaging, i.e., the low signal-to-noise ratio of the reflections and azimuth ambiguity. First, the azimuth ambiguity, which exists extensively in conventional single-well imaging, is solved with an improved imaging procedure using combined monopole–dipole logging data in addition to conventional logging data. Furthermore, we demonstrate that the direct waves propagating along the boreholes with strong energy, can be effectively eliminated with the proposed combined monopole–dipole measurement mode. The reflections are therefore predominant in the combined monopole–dipole data even before the signals are filtered; thus, the reflections’ arrival times in each receiver are identified, which may help minimize the difficulties in filtering conventional logging data. The optimized processing flow of the combined measurement mode logging image is given in this paper. The proposed combined monopole–dipole measurement mode may improve the accuracy of single-well imaging.

A modified Kozeny-Carman model for estimating anisotropic permeability of soft mudrocks

Abstract This paper quantifies the effects of clay fraction and fabric on the permeability of soft mudrocks with emphasis on micro-level mechanisms. Soft mudrocks are treated as a mixture composing the non-clay minerals and the clay-water composite as two constituents. A simplified approach to estimate the fabric orientation distribution of soft mudrocks based on readily measured parameters such as clay fraction and porosity is presented. A modified Kozeny-Carman model is proposed to estimate the anisotropic permeability of soft mudrocks. An effective porosity is used in the proposed model to consider the effect of clay-bound water on the flow behavior. The anisotropy in permeability is studied by using an anisotropic tortuosity which is quantitatively linked to the data of fabric orientation distribution and porosity. In the proposed theoretical model, the clay fraction data is included in parameters of effective porosity, anisotropic tortuosity, and external specific surface area. An increase in clay fraction results in a significantly decrease in vertical permeability which is due to the increases of external specific surface area and tortuosity along the vertical direction (perpendicular to the bedding plane), and the decrease of effective porosity. The proposed permeability model has been validated using published permeability data on soft mudrock samples with different geological origins. The model can be used to predict anisotropic permeability in soft mudrocks or hard clay soils from conventional geophysical logging data with minimal core analysis.

Estimation of total porosity in shale formations from element capture logging and conventional logging data

The Herron model does not consider the effect of organic matter and gas on total porosity. In this paper, the shale rock is equivalent to three parts: skeleton, organic matter, and pore. According to the volume model, a new total porosity model is established. The new model is applied to estimate the total porosity of a shale gas well in South Sichuan area from element capture logging and conventional logging data. The results show that in the formations with high gas content and organic matter, the total porosity calculated by the new model is smaller than that calculated by the Herron model. To verify the accuracy of the new model in evaluating the total porosity of shale, the advantages and disadvantages of Gas Research Institute (GRI) measurement method and helium injection measurement method for determining porosity are discussed in terms of theoretical analysis and experimental measurement. It is considered that the GRI porosity is closer to the total porosity of shale and it can be used to test the total porosity of logging evaluation. The total porosity of the ten shale samples calculated by the new model and the Herron model is compared with the GRI porosity. Results show that the average relative errors calculated by the Herron model and the new model are 61 and 10%, respectively. It indicates that the accuracy of the total porosity calculated by the new model is higher.

Generating porosity spectrum of carbonate reservoirs using ultrasonic imaging log

Imaging logging tools can provide us the borehole wall image. The micro-resistivity imaging logging has been used to obtain borehole porosity spectrum. However, the resistivity imaging logging cannot cover the whole borehole wall. In this paper, we propose a method to calculate the porosity spectrum using ultrasonic imaging logging data. Based on the amplitude attenuation equation, we analyze the factors affecting the propagation of wave in drilling fluid and formation and based on the bulk-volume rock model, Wyllie equation and Raymer equation, we establish various conversion models between the reflection coefficient β and porosity ϕ. Then we use the ultrasonic imaging logging and conventional wireline logging data to calculate the near-borehole formation porosity distribution spectrum. The porosity spectrum result obtained from ultrasonic imaging data is compared with the one from the micro-resistivity imaging data, and they turn out to be similar, but with discrepancy, which is caused by the borehole coverage and data input difference. We separate the porosity types by performing threshold value segmentation and generate porosity–depth distribution curves by counting with equal depth spacing on the porosity image. The practice result is good and reveals the efficiency of our method.

Application of Fracture Distribution Prediction Model in Xihu Depression of East China Sea

There are different responses on each of logging data with the changes of formation characteristics and outliers caused by the existence of fractures. For this reason, the development of fractures in formation can be characterized by the fine analysis of logging curves. The well logs such as resistivity, sonic transit time, density, neutron porosity and gamma ray, which are classified as conventional well logs, are more sensitive to formation fractures. In view of traditional fracture prediction model, using the simple weighted average of different logging data to calculate the comprehensive fracture index, are more susceptible to subjective factors and exist a large deviation, a statistical method is introduced accordingly. Combining with responses of conventional logging data on the development of formation fracture, a prediction model based on membership function is established, and its essence is to analyse logging data with fuzzy mathematics theory. The fracture prediction results in a well formation in NX block of Xihu depression through two models are compared with that of imaging logging, which shows that the accuracy of fracture prediction model based on membership function is better than that of traditional model. Furthermore, the prediction results are highly consistent with imaging logs and can reflect the development of cracks much better. It can provide a reference for engineering practice.

Application of PSO-ELM Algorithm in Porosity Prediction of Tuffaceous Sandstone Reservoir

Aiming at the problem that the conventional logging data processing method has low accuracy and large error for complex reservoirs, this paper presents a PSO-ELM algorithm based on particle swarm optimization for reservoir porosity prediction. The prediction model is established by the limit learning machine (ELM). The output weight of ELM is optimized by particle swarm optimization algorithm, and the upper and lower limits of the optimal prediction interval are obtained, and the advantages of ELM learning speed and generalization ability are fully utilized to forecast the reservoir porosity. The PSO-ELM algorithm is used to test the tuff sandstone reservoirs in a certain area, and the results are in good agreement with the core data.

A new method for lithology identification of fine grained deposits and reservoir sweet spot analysis: A case study of Kong 2 Member in Cangdong sag, Bohai Bay Basin, China

Abstract Based on systematic coring of 500 m of Kong 2 Member of the Paleogene Kongdian Formation in Cangdong sag of Bohai Bay Basin, identification and XRD (X-Ray Diffraction) analysis of over 1000 thin sections, a simplified method to quantitatively calculate contents of fine grained minerals with conventional logging data such as acoustic travel time (AC) and density log (DEN) has been proposed, and a quick lithologic identification “green mode” has been worked out in this study. By fitting the relationship between normalization of logging curves and mineral content measured by XRD, the mineral contents of sections or wells not cored can be calculated to identify lithology. With this method, several dolomite sweet spot intervals and one sandstone sweet spot interval have been found in the Kong 2 Member of Cangdong sag, where high production oil and gas flows have been tapped from drilled wells. The study shows that the dolomite is in band distribution and enriched in local parts of the study area. This method is applicable to lithologic identification of fine grained deposits in front delta-lake basin center, especially lithologic identification of mud and dolomite dominated fine grained deposits with low sand content of semi-deep, deep lake facies.

Productivity Prediction of Tight Sandstone Reservoir Based on BP Neural Network

To survey He-8 member tight sand reservoir with low porosity and permeability in Mizhi gas field in Ordos basin, using the conventional well log data, this paper proposes the tight sand reservoir productivity prediction model and classification criterion based on BP neural network, getting quick classification of gas well productivity. We can predict sand reserve quantitatively instead qualitatively with the methods．Applications show that the methods of productivity prediction are effective and practical．

Sequence stratigraphy, sedimentary systems and implications for hydrocarbon exploration in the northern Xujiaweizi Fault Depression, Songliao Basin, NE China

Recent drilling data from the northern Songliao Basin reveal considerable hydrocarbon exploration potential in deep-seated small-scale faulted basins. In this paper, the Shahezi Formation of the northern Xujiaweizi Fault Depression within the Songliao Basin was investigated as a case study. Sequence stratigraphy, sedimentary systems and the hydrocarbon exploration potential of the small-scale faulted basin were investigated through observations and descriptions of cores; analyses of conventional log curves, imaging data and logging data; and interpretations of 3-D seismic profiles. Four third-order sequences, mainly controlled by sediment supply and fault activity, were identified based on unconformities at the basin margins and in sedimentary successions. System tracts in the third-order sequences were also recognized through lithological associations and geophysical response characteristics. The distributions of sedimentary systems, including fan deltas, braided deltas, lacustrine deposits, subaqueous fans, volcaniclastic deposits and incised-valley deposits, in the sequence framework are associated with basin evolution. During the early stage of the Shahezi Formation (SQ1 and SQ2), the sedimentary area was mainly covered by fan deltas, and lacustrine deposits were not very developed. Additionally, incised-valley systems formed on basin margins and subaqueous fans formed along basin-controlled faults. During the middle stage (SQ3), fan delta deposits on gentle slopes were gradually replaced by braided delta deposits, and the sizes of lacustrine facies increased with basin expansion. In the late stage (SQ4), the basin and lacustrine deposit areas reached their maximum sizes, and braided deltas became the dominant sedimentary systems on gentle slopes. With continuous sediment supply, fan and braided deltas migrated forward, and the lacustrine system was locally distributed. In this small-scale faulted basin, the lacustrine mudstones and coal seams widely developed in SQ4, especially in SQ4-HST, and they also exhibited good source rock properties for gas generation. Sandbodies of fan deltas and braided deltas in the upper Shahezi Formation, exhibiting good contact relationships with underlying source rocks, can be favourable reservoirs for hydrocarbon accumulation, as confirmed by the newly drilled Well Songs9H (with a daily gas production of 20×104 m3). However, due to low porosity and permeability (averages of 4.7% and 0.28×10−3µm2, respectively), these reservoirs may need fracturing for commercial hydrocarbon production.

Well logging evaluation of water-flooded layers and distribution rule of remaining oil in marine sandstone reservoirs of the M oilfield in the Pearl River Mouth basin

In the marine sandstone reservoirs of the M oilfield the water cut is up to 98%, while the recovery factor is only 35%. Additionally, the distribution of the remaining oil is very scattered. In order to effectively assess the potential of the remaining oil, the logging evaluation of the water-flooded layers and the distribution rule of the remaining oil are studied. Based on the log response characteristics, the water-flooded layers can be qualitatively identified. On the basis of the mercury injection experimental data of the evaluation wells, the calculation model of the initial oil saturation is built. Based on conventional logging data, the evaluation model of oil saturation is established. The difference between the initial oil saturation and the residual oil saturation can be used to quantitatively evaluate the water-flooded layers. The evaluation result of the water-flooded layers is combined with the ratio of the water-flooded wells in the marine sandstone reservoirs. As a result, the degree of water flooding in the marine sandstone reservoirs can be assessed. On the basis of structural characteristics and sedimentary environments, the horizontal and vertical water-flooding rules of the different types of reservoirs are elaborated upon, and the distribution rule of the remaining oil is disclosed. The remaining oil is mainly distributed in the high parts of the structure. The remaining oil exists in the top of the reservoirs with good physical properties while the thickness of the remaining oil ranges from 2–5 m. However, the thickness of the remaining oil of the reservoirs with poor physical properties ranges from 5–8 m. The high production of some of the drilled horizontal wells shows that the above distribution rule of the remaining oil is accurate. In the marine sandstone reservoirs of the M oilfield, the research on the well logging evaluation of the water-flooded layers and the distribution rule of the remaining oil has great practical significance to the prediction of the distribution of the remaining oil and the optimization of well locations.

Prediction of Naturally Fractured Reservoir Performance using Novel Integrated Workflow

Generation of naturally fractured reservoir subsurface fracture maps and prediction its production potential are considered complex process due to insufficient data available such as bore hole images, core data and proper reservoir simulator model. To overcome such shortcomings, the industry has relied on geo-statistical analyses of hard and soft data, which are often referred to as static data. This paper presents an integrated workflow that models and predicting fractured reservoirs performance, through the use of gradient-based inversion techniques and discrete fracture network modelling (DFN), which—through the inversion of well test data (i.e., dynamic data)—aims to optimise fracture properties and then predicting of the reservoir production potential. The first step in the workflow is to identify flow contributing fracture sets by analysing available core descriptions, borehole images, conventional log data and production data. Once the fracture sets are identified, the fracture intensity is statistically populated in the inter-well space. In the second step, 3D block-based permeability tensors are calculated based on flow through discrete fractures, and the fracture intensity is then propagated away from the wellbore, i.e., by relating to permeability tensors with fracture intensity. In the final step (fracture optimisation), the fracture properties are computed by DFN modelling, which includes distribution, orientation and geometry in different realisations. Fluid flow is simulated in these discrete fractures to estimate pressure change and pressure derivatives. The production rate associated with drill stem test that performed within this reservoir area has been successfully simulated using the optimised subsurface fracture map that has been generated from the first step.

A new method for assessing Young's modulus and Poisson's ratio in tight interbedded clastic reservoirs without a shear wave time difference

Tight clastic reservoirs are an important aspect of hydrocarbon exploration and development worldwide. The Xu5 section of the Upper Triassic Xujiahe Formation in the western Sichuan Basin of central China contains tight terrestrial clastic reservoirs. The depth of these reservoirs is greater than 3000 m, while the thickness is generally greater than 500–580 m; the tight sandstone and shale reservoirs are frequently interbedded. Based on the proportion of sandstone and shale, the reservoirs can be divided into four types: sand-rich, interbedded Type I, interbedded Type II, and shale-rich. Horizontal wells and multistage fracturing are used due to the complexity and significant heterogeneity of the tight clastic reservoirs in the study area. The Young's modulus and Poisson's ratio of the reservoir rocks are the two most significant parameters that must be evaluated when implementing these key techniques. Based on mechanical and acoustic tests, a new assessment system for the Young's modulus and Poisson's ratio based on the Hoek-Brown criterion was proposed. This evaluation system does not require the shear wave time difference; thus, the increase in error caused by the unreliability of the shear wave time difference can be avoided when assessing strata with complex geological structures. This method can be directly applied in single wells using only conventional longitudinal wave time difference log data. After converting the wave velocity of the high experimental frequency into the wave velocity of the logging frequency using the acoustic dispersion technique, the analysis results were applicable to the log evaluations. The relationship between the Hoek-Brown criterion parameters and the mechanical and acoustic parameters of the rock were systematically analyzed. By distinguishing the lithologies, the equations for the Young's moduli of tight sandstones and shales were established based on the Hoek-Brown criterion. The Young's modulus can be expressed using only the longitudinal wave velocity (VP). A significant negative correlation exists between the Poisson's ratio and Young's modulus; therefore, the Poisson's ratio can be evaluated from the Young's modulus. The errors of the Young's modulus and Poisson's ratio predicted by the proposed assessment method are small, which meets engineering requirements and validates the accuracy of the method.

Determination of Shale Volume and Distribution Patterns and Effective Porosity from Well Log Data Based On Cross-Plot Approach for A Shaly Carbonate Gas Reservoir

Determination of shale volume distribution is one of the most important factors that has to be considered in formation evaluation, since existence of shale reduces effective porosity and permeability of the reservoir. In this paper, shale volume and distribution (dispersed, laminar and structural) and formation effective porosity are estimated from well log data and cross-plots. Results show that distribution of shale is mainly dispersed with few of laminar ones, and the quality of reservoir (effective porosity) decreases with depth resulting in low productivity of gas wells drilled in lower zones. Good agreement of estimated shale volumes and effective porosities from neutron-density cross-plot with the values determined from gamma ray log (CGR) and core analysis demonstrates the accuracy and applicability of these plots in determination of petrophysical parameters from conventional log data.

Brittleness index prediction in shale gas reservoirs based on efficient network models

Brittleness index is one of the critical geomechanical properties of unconventional reservoir rocks to screen effective hydraulic fracturing candidates. In petroleum engineering, brittleness index can be generally calculated from the mineralogical composition by X-ray diffraction (XRD) test or rock mechanical parameters by tri-axial experiments and well logs. However, mineral composition analysis or tri-axial experiments cannot produce continuous brittleness profile. Well log-based brittleness index prediction conventionally relies on Young's modulus and Poisson's ratio, but sometimes shear compressional velocity is not available to derive elastic inputs for the brittleness index calculation. This study proposes some data-driven practical brittleness prediction approaches based on back-propagation artificial neural network (BP-ANN), extreme learning machine (ELM) and linear regression using commonly available conventional logging data and lab mineralogical-derived brittleness. A dataset of 71 mineralogical-derived brittleness measurements from Silurian Longmaxi marine shale, Jiaoshiba Shale Gas Field, Sichuan Basin, China were established. The model comparisons and error analysis reveal that the application of artificial intelligence models can be more effectively applied to brittleness prediction compared with simple regression correlations. Both BP-ANN and ELM models are competent for brittleness prediction while BP-ANN model can produce slightly better brittleness prediction results with same inputs and ELM model require less running time. Thus, more choices can be made according to accuracy and computational speed demand. Moreover, an overall ranking of sensitivity degree is then provided to show the impacts of different well logs as inputs on the BP-ANN and ELM model, which is helpful to find optimal inputs in given case. Comparing to traditional well-log based brittleness approaches, data-based approaches show its wider applications because the integration of mineralogical composition and well log information can provide continuous brittleness profile in terms of high accuracy while acoustic full waveform velocities are no longer necessary inputs in brittleness evaluation.