Geophysical Well-Log Evaluation in the Era of Unconventional Hydrocarbon Resources: A Review on Current Status and Prospects

Abstract

Geophysical well-log evaluation in the era of unconventional hydrocarbon resources (mainly tight oil and gas, shale oil and gas) is complicated and challenging. This review aims to fill this gap between well-log evaluation and unconventional hydrocarbon resources by characterizing the source rock property, reservoir property and engineering property using petrophysical well logs. The advanced well-log series used for unconventional oil and gas evaluation include nuclear magnetic resonance (NMR) log, image logs, array acoustic logs, elemental capture spectroscopy (ECS) and LithoScanner logs. The source rock property in terms of total organic carbon content is predicted using conventional logs and LithoScanner log. Then petrophysical parameters including porosity, permeability and oil saturation are calculated, and the appearance of natural fracture is predicted from conventional, sonic logs, image logs and NMR logs. Additionally, the reservoir property is evaluated to optimize the favorable layers with high hydrocarbon bearing property and productivity. Brittleness index as well as in situ stress direction and magnitudes are characterized by the comprehensive use of density, sonic log, ECS log and image logs. Then, the engineering property (high brittleness index but low horizontal stress difference) is evaluated to screen out the prospected layers for hydraulic fracturing. The internal relationships between the three types of properties are unraveled, and the geological and engineering sweet spots are optimized by integrating lithology, reservoir quality, hydrocarbon bearing property, source rock property, brittleness and in situ stress magnitude and direction. This multidisciplinary approach provides a comprehensive method for optimizing sweet spots in unconventional play, and will support petroleum geoscientists’ and engineers’ decisions in exploration and exploitation of unconventional hydrocarbon resources.

FormalPara Article Highlights

• This paper surveys the current status and prospects of well log evaluation of unconventional oil and gas

• Geophysical well log evaluation of unconventional hydrocarbon resources characterizes source rock property, reservoir property and engineering property

• Geological and engineering sweet spots can be optimized by petrophysical and geomechanical properties determined from well logs

1 Introduction

Unconventional hydrocarbon resources, which mainly include tight and shale oil and gas, etc., play more and more important roles in the world energy structure (Zou et al. 2019; Nikolaev and Kazak 2019; Wu et al. 2019; Amosu et al. 2021; Mukhametdinova et al. 2021). The increasing market demand and technological advances in directional geosteering, horizontal drilling and multi-stage hydraulic fracturing have made unconventional plays a major focus for the global petroleum industry (Qiu et al. 2016; Curtis et al. 2012; Rybacki et al. 2016; Iqbal et al. 2018; Sun et al. 2021). However, unconventional tight to shale reservoirs have varied lithologies (structure and composition) (Chen et al. 2017; Cao et al. 2017; Li Maowen et al. 2019), ultra-low permeability (Gale et al. 2007; Josh et al. 2012; Avanzini et al. 2016; Bai et al. 2017) and the pore systems are dominated by highly heterogeneous nano- to microscale pore assemblages (Curtis et al. 2012; Loucks et al. 2012; Manjunath and Jha 2019; Chandra and Vishal 2021). Therefore, unconventional reservoirs commonly have remarkably different well-log responses compared with conventional reservoirs (Iqbal et al. 2018), and formation evaluation for unconventional plays using petrophysical well logs remains complicated and challenging (Du et al. 2021; Amosu et al. 2021; Liu 2021).

Source rock property needs to be evaluated via well logs for the self-sourced and self-retained unconventional resources (Zhao et al. 2019). Reservoir quality is also one of the key risk factors for productivity (Zhang et al. 2015; Mukhametdinova et al. 2021). In addition, unconventional reservoirs have no natural productivity, and therefore, hydraulic fracturing is required for the economic production (Gale et al. 2007; Josh et al. 2012; Avanzini et al. 2016; Dong et al. 2018; Manjunath and Jha 2019). Consequently, the engineering property (brittleness and in situ stress states) has become critical petrophysical parameters for screening prospected layers for hydraulic fracturing (Lai et al. 2015; Iqbal et al. 2018). The optimization of “sweet spots” (hydrocarbon-rich zones with matured high-quality source rocks, favorable reservoir quality and prospected layers for stimulation) integrating source rock property, reservoir property and engineering property is essential for exploitation of unconventional resources (Zou et al. 2013; Lu et al. 2019). Petrophysical well logs, when calibrated with core analysis data, have the advantages for continuous evaluation of source rock property, reservoir property and engineering property of unconventional plays with low cost (Clarkson et al. 2012; Avanzini et al. 2016; Iqbal et al. 2018).

Due to the fundamental difference in petrophysical and geomechanical properties of unconventional plays, a systematic workflow is required to characterize the seven kinds of parameters (lithology, reservoir quality, hydrocarbon bearing property, electronic well-log responses, source rock property, brittleness and in situ stress magnitude and direction) and the three kinds of properties (source rock property, reservoir property and engineering property) (Zou et al. 2019; Mukhametdinova et al. 2021). This review aims to fill the gap between geological and petrophysical characterization of unconventional resources, and the different types of logs from the Ordos Basin, the Subei Basin, the Junngar Basin and the Tarim Basin are used. Firstly, the commonly used geophysical well-log series are reviewed. Secondly, the source rock property in terms of TOC abundance is evaluated using conventional, spectrum gamma-ray and LithoScanner logs. Thirdly, the reservoir property in terms of lithology, porosity, permeability, oil saturation and the presences of fractures is predicted by conventional, sonic logs, image logs and NMR logs. Fourthly, the engineering property in terms of in situ stress direction and magnitudes as well as brittleness index is characterized by sonic logs and image logs. Lastly, the geological and engineering sweet spots optimization is performed by unraveling the relationships between three types of properties. This study critically reviews the petrophysical well-log evaluation of unconventional resources, as assessed from peer reviewed papers and from the authors’ personal experiences, with the aim that readers can use these results for their future work.

2 Geophysical Well-Log Series

A series of geophysical well-log suits spanning from conventional well logs to advanced well-log suits, which have varied vertical resolution and depth of investigation, can be used for unconventional oil and gas resource evaluation (Fig. 1) (Yarmohammadi et al. 2020; Mukhametdinova et al. 2021).

Fig. 1

Conventional geophysical well-log data and advanced well-log suits with various vertical resolutions used in well-log evaluation

Conventional well-log series include caliper (CAL), spontaneous potential (SP), natural gamma-ray (GR), litho-density (Pe), sonic interval transit time or acoustic log (AC), compensated neutron log (CNL) and bulk density (DEN) (Fig. 1). Three resistivity logs can be divided according to their depth of investigation, and micro log (MSFL, MIL) measures the resistivity of mudcake with a depth of investigation of 2.5–10 cm. Lateral logs (LLS, LLD) and induction logs (ILM, ILD) can measure the resistivity of flushed zone (Rxo) and uninvaded zone (Rt) (Fig. 1). In addition, in order to improve the vertical resolution and depth of investigation, high definition induction logs HDIL (M2Rx, M2R9, M2R6, M2R3, M2R2, M2R1), which have a vertical resolution of 1–2 ft, and depths of investigation of 10–120 ft, are developed (Fig. 1).

Electrical or ultrasonic borehole image logs, which measure the electrical resistivity or acoustic impedance of borehole wall (Khoshbakht et al. 2009), provide very high-resolution (5 mm) borehole pictures (Prioul et al. 2007; Folkestad et al. 2012). FMI (fullbore formation microimager) imaging tool, which have 8 pads and each contains 24 buttons, electrically scan the borehole wall, and a total of 192 micro-resistivity curves are collected (Khoshbakht et al. 2009; Rajabi et al. 2010). Then, the micro-resistivity curves are used to build up a “pseudo-picture” of the wellbore through speed correction, eccentering correction and normalization (Goodall et al. 1998; Rajabi et al. 2010; Wilson et al. 2013; Keeton et al. 2015). Geological features in terms of fracture, fault, bedding and vug as well as cavity can be picked out from the high-resolution images (Fig. 1) (Xu et al. 2009; Khoshbakht et al. 2012; Lai et al. 2018a; Wang et al. 2020).

Nuclear magnetic resonance (NMR) logging tool measures the longitudinal relaxation time (T₁), transversal relaxation time (T₂) spectrum and diffusion coefficient (D) and has a vertical resolution of 0.2 m (Kleinberg et al. 2005; Tan et al. 2014; Bauer et al. 2015; Liu et al. 2019; Wang et al. 2020). The NMR T₂ spectrum is widely used for fluid property discrimination (Wang et al. 2020), and can be used for determination of porosity, estimation of permeability, evaluating irreducible water saturation and hydrocarbon saturation (Fig. 1) (Dunn et al. 2002; Kleinberg et al. 2005; Hübner 2014; Bauer et al. 2015; Olatinsu et al. 2017). Additionally, plot of T₁ versus T₂ as well as plot of T₂ versus diffusion coefficient can be used for fluid property determination since the hydrocarbon and water signals can be clearly separated due to the large contrast between diffusion coefficients of hydrocarbon and water (Sun and Dunn 2005).

The elemental capture spectroscopy (ECS) and LithoScanner logs with a vertical resolution of 0.457 m (1.5 ft) can directly provide the rock compositions (clay, quartz, feldspar, and mica (Q-F-M), carbonate and pyrite, etc.) (Maliva et al. 2009; Collett et al. 2011), and are widely used for mineral content determination, lithology recognition and even TOC calculation (Fig. 1) (Guo et al. 2019).

Array acoustic logging tools or sonic scanner (MSIP—modular sonic imaging platform) logs, which have a vertical resolution of 3.0 m, measure the full wave-forms including compressional wave slownesses, shear-wave slownesses, Stoneley wave and pseudo-Rayleigh wave (Collett et al. 2011; Zaree et al. 2016). Therefore, sonic logs are widely used in the fields of engineering geology in calculating geomechanical parameters, determining in situ stress and rock anisotropy (Fig. 1) (Liu et al. 2018).

Conventional well logs can be collected in most of the wells, while the advanced well logs (image logs, NMR log, LithoScanner log, array sonic logs) are too expensive, and are not available for all the wells drilled.

3 Source Rock Property

Quality (types), quantity (abundance) and thermal maturity are three important geochemical parameters for source rock evaluation (Zhao et al. 2019). Organic matter abundance (quantity) can be characterized by total organic carbon content (TOC), which is defined as the organic richness or amount of organics within source rocks (Jarvie et al. 2007; Iqbal et al. 2018). Thermal maturity is the degree or stage of organic matters transformation into hydrocarbon at adequate pressure and temperature with increase in burial depth (Zhao et al. 2019). Vitrinite reflectance (Ro, %) is often used as a measure of thermal maturity of source rocks (Zhao et al. 2019). Zhao et al. (2019) integrates resistivity, neutron and density logs to estimate thermal maturity index for Barnett shale.

3.1 Well-log Responses of Source Rocks

Shales have distinct responses in well logs due to the unique physical properties of the organic matters (Aziz et al. 2020). Source rocks are organic matter-rich mudstones/shales, which can be qualitatively recognized by conventional well logs due to their distinct petrophysical properties compared with reservoir rocks. Mudstones and shales are inherently different from sandstones or carbonate rocks due to their high gamma-ray; additionally, the presences of organic matters will amplify this effect (Aziz et al. 2020). Conventional well-log series which are sensitive to organic matters include gamma-ray (GR), sonic transit interval time (AC), neutron porosity (CNL), bulk density (DEN) and resistivity (RT) (Wang et al. 2019; Aziz et al. 2020).

GR tool measures the radioactivity, and the presence of organic matter-rich source rocks will cause intense radioactivity and consequently high GR values (Shalaby et al. 2019). For instance, the source rocks in Yanchang Formation Member 7 of Ordos Basin, West China, have high GR readings (Fig. 2). The anomalously high GR was mainly ascribed to uranium, which is related to the organic matter (Fig. 2) (Wang et al. 2019).

Fig. 2

source rocks in Yanchang Formation Member 7 of Ordos Basin, West China

Well-log responses of

Mudstones conventionally show high AC values than sandstones, and transit interval time of organic matters (about 500 μs/m) is much larger than that of rock matrix. Therefore, organic matter-rich source rocks will show much higher AC logging values (Fig. 2) (Wang et al. 2019).

Neutron log measures the hydrogen index in rocks (Shalaby et al. 2019). Organic matters commonly have high hydrogen index, which gives high neutron porosity logging (CNL) value (Fig. 2) (Wang et al. 2019).

Density log is the comprehensive reflection of fluids and matrix components (Shalaby et al. 2019). Organic matters (kerogen) have much lower bulk density (about 1.0 g/cm³) than the matrix rocks (2.5–2.7 g/cm³); consequently, the density log value will significantly decrease when encountered with source rocks (Fig. 2) (Wang et al. 2019; Shalaby et al. 2019).

The mudstone intervals generally exhibit low resistivity because of the good conductivity of clay minerals and pore water. However, the organic matter-rich mudstones (shales) are non-conductive; additionally, in mature source rocks, the non-conductive hydrocarbon will further cause a high anomaly in the resistivity logs (Passey et al. 1990; Wang et al. 2019; Shalaby et al. 2019; Wood 2020a). The high definition resistivity log (AT10-AT90) shows abnormally high readings and is deviated evidently in the source rock intervals (Fig. 2) (Shalaby et al. 2019).

In Fig. 2, the source rocks are therefore evidently recognized in the 1710–1718 m depth intervals and are characteristic of abnormal high GR readings, low bulk density, but high sonic transit time, high neutron porosity and high resistivity (Fig. 2). The image logs are white due to the high resistivity and abundant lamina is recognized. The NMR logs and array acoustic logs show no evident responses in the source rock intervals (Fig. 2).

3.2 TOC Prediction via Well Logs

TOC, volatile hydrocarbon (S₁) and remaining hydrocarbon (S₂), which reflect richness and hydrocarbon generation potential of organic matters, are three significant factors for source rock property evaluation (Aziz et al. 2020). The content of S₁ and S₂ estimations can be processed through regression analysis between TOC and (S₁ + S₂) content (Wang et al. 2019). TOC is the organic richness or amount of organics within rock (Jarvie et al. 2007; Iqbal et al. 2018). TOC value can be accurately measured using rock pyrolysis (Iqbal et al 2018); however, core data are not available in all intervals or wells due to high cost and low recovery rate (Aghli et al. 2016; Lai et al. 2017). Therefore, TOC estimation using well logs is vital for continuous evaluation of source rock property. There are various methods proposed to calculate TOC using well logs, including (1) ΔlogR method, (2) spectral gamma-ray log, (3) multivariate fitting method and (4) advanced log method (LithoScanner log), (5) machine Learning (learning) method, etc.

3.2.1 ΔlogR Method

The ΔlogR method was initially proposed by Passey et al. (1990), and had been widely used for TOC estimation using well logs in carbonates and clastic rocks. ΔlogR, i.e., the sonic-resistivity overlay plot AC and true formation resistivity Rt logs in one track, additionally AC and RT are scaled as a ratio of 50 μs/ft to one resistivity cycle. The separation between two curves (AC to the left and RT to the right) is defined as ΔlogR (Passey et al. 1990; Shalaby et al. 2019). The AC curve will reflect low density/low velocity kerogens, while the Rt curve will respond to formation fluid (Tenaglia et al. 2020).

The method of sonic-resistivity overlay (ΔlogR) is used to calculate TOC content (Eqs. 1, 2).

$$\Delta \log R = \log (R/R_{{B{\text{aseline}}}} ) + 0.02(\Delta t - \Delta t_{{_{{B{\text{aseline}}}} }} )$$

(1)

$${\text{TOC}} = \Delta \log R \times 10^{{_{{^{(2.297 - 0.1688LOM)} }} }}$$

(2)

R is deep resistivity log (Ω m⁻¹), and the RT and LLD logs can be used; Δt (AC) is the sonic transit time (μs/ft or μs/m), almost all suits of logs give the AC curve; R_Baseline and Δt_Baseline are the resistivity and sonic transit time values at the base line; LOM (local level of organic metamorphism) is a constant related to thermal maturity. In some cases, sonic transit time curve may not be unavailable, and then, density or neutron log can be used instead (Shalaby et al. 2019; Tenaglia et al. 2020).

The scaling of sonic and resistivity well logs during overlaying is set as a ratio of − 164 μs/m (− 50 μs/ft) sonic transit time (Δt) to one logarithm resistivity cycle Ω m⁻¹ (for instance, 1–10 Ω m⁻¹) (Passey et al. 1990; Iqbal et al. 2018; Godfray and Seetharamaiah 2019; Wang et al. 2019; Shalaby et al. 2019). The TOC can be calculated using the above ΔlogR method providing that the baseline of AC and RT is determined. The TOC predicted from ΔlogR methods is in accordance with the core-measured TOC content in Lucaogou Formation of Jimusar Sag, Junggar Basin, West China (Fig. 3). Then, the TOC content is estimated using the above formula, and the results are in accordance with core-measured TOC content (Fig. 3).

Fig. 3

TOC estimation using ΔlogR method in Lucaogou Formation of Jimusar Sag, Junggar Basin, West China

Source rock intervals can be firstly qualitatively identified using GR, CNL, DEN and resistivity log curves. Additionally, plotted AC and RT in one track with scaling of 100 μs/ft (for instance 140–40 μs/ft) transit time to two logarithm resistivity cycle (10–1000 Ω m⁻¹) also give the source rock intervals. As is known, AC and RT logs will overlap with each other in non-source rock interval, while the two logs will deviate evidently at the source rock intervals. There are many ΔLogR-based methods widely used to estimate TOC content using well logs (Zhao et al. 2016).

The limitation of ΔlogR is that the baselines vary from formation to formation and well to well (Wang et al. 2019). Additionally, the presence of pyrite will mask resistivity profile and show low resistivity (Passey et al. 1990; Iqbal et al. 2018).

3.2.2 Spectral Gamma-Ray Log

GR logging tool measures the total intensity of radioactivity of formation which come from K, Th and U elements. GR spectrometer allows to detect the individual concentrations of K (%), U (ppm) and Th (ppm) (Sérgio et al. 2018). Spectral gamma-ray logs are therefore widely used for estimating clay content, paleoclimate reconstruction grain size evaluation (Sérgio et al. 2018). Organic matters can absorb abundant U elements, and therefore, U log curve or KTH (GR log without U contribution) can be used for TOC estimation. As can be observed in Fig. 4, in Funing Formation in Subei Basin, East China, the TOC value shows negative relationship with KTH log with high correlation coefficient (Fig. 4).

Fig. 4

Crossplot of KTH versus core-measured TOC content of Funing Formation in Subei Basin, East China. KTH is the GR log without U contribution derived from spectral gamma-ray log

3.2.3 Multivariate Fitting Method

The multivariate fitting method is integrating two or more log curves which are sensitive to the source rocks to establish a model to calculate TOC (Aziz et al. 2020). For instance, the model used to estimate TOC using well logs can be written as the following Eq. (3) through multivariate fitting method. Consequently, the TOC content in well intervals can be predicted, and the results are in good accordance with core-measured TOC (Fig. 5).

$${\text{TOC}} = 0.0194{\text{GR}} + 3.582{\text{AC}} - 0.0051{\text{LLD}} - 8.124{\text{DEN}} - 175.352$$

(3)

Fig. 5

TOC content calculated from multivariate fitting method in Tarim Basin, West China

3.2.4 LithoScanner Logs

The LithoScanner logging technology, which was proposed by Schlumberger Company, is the improvement of elemental capture spectroscopy (ECS) log (Guo et al. 2019). LithoScanner logs can measure the content of common elements including carbon, potassium, magnesium, aluminum and sodium (Guo et al. 2019). Through data processing, the element content can be transformed into the mineralogy content, including clay, Q–F–M (quartz–feldspar–mica), carbonate (calcite and dolomite) and pyrite (Fig. 1). Consequently, the ECS and LithoScanner logs are widely used for mineral composition determination, lithology recognition or even brittleness index evaluation (Maliva et al. 2009; Collett et al. 2011; Lai et al. 2015).

LithoScanner provides the total carbon element (TC) content; however, there is also inorganic carbon (TIC) in the formation (Fig. 6). In hydrocarbon reservoirs, the inorganic carbon is mainly associated with carbonate rocks (Ca₂CO₃, Ca(Mg)CO₃, etc.). Consequently, the TOC content can be obtained, providing that the inorganic carbon content is eliminated (Fig. 6). The TOC content calculated from LithoScanner log in Yanchang Formation Member 7 of Ordos Basin is consistent with the core analysis data (Fig. 6).

Fig. 6

TOC content calculated from LithoScanner log in Yanchang Formation Member 7 of Ordos Basin, West China

3.2.5 Machine Learning Methods

The relationships between TOC and logs (GR, AC, DEN, RT, etc.) are complicated and nonlinear; therefore, machine learning methods should be integrated to predict TOC content (Mahmoud et al. 2017; Wang et al. 2019). Wang et al. (2019) uses the artificial neural network (ANN) method to predict TOC content using AC and DEN logs, which improve the efficiency. The support vector machine (SVM) method can also be adopted in TOC prediction via GR, AC and DEN logs (Amosu and Sun 2021). The machine learning methods have the advantages of high accuracy (Mahmoud et al. 2017; Amosu and Sun 2021).

4 Reservoir Property

Reservoir property evaluation aims at characterizing the lithology, porosity (φ), permeability (K) and hydrocarbon saturation (Sh). Conventional (full suite) logs can be used for calculation of these petrophysical parameters for conventional hydrocarbon reservoirs (Yarmohammadi et al. 2020). However, advanced well-log suits including LithoScanner logs, image logs, NMR logs and sonic array logs are required to evaluate the reservoir property (lithology, reservoir quality, fracture, as well as oil bearing property) for unconventional reservoirs due to the complex pore assemblage and petrophysical log responses (Rybacki et al. 2016; Avanzini et al. 2016; Iqbal et al. 2018; Zhao et al. 2019; Yarmohammadi et al. 2020; Liu 2021).

4.1 Well-Log Responses of Oil and Water Bearing Layers

The dry layers (non-reservoir intervals) of Yanchang Formation Member 7 in Ordos Basin have low porosity as can be evidenced by the three porosity logs, and no evident deviations of the deep (AT60–AT90) and shallow (AT10–AT20) induction logs (Fig. 7). Reservoir layers are characterized by high reservoir quality. As can be observed in Layer 1 and Layer 2 in Fig. 7, the high sonic transit time, high CNL but low bulk density gives signatures of high reservoir quality (Fig. 7). However, the reservoirs can be water saturated and oil saturated, and the water bearing layers commonly have no deviations of resistivity logs with various depth of investigation (Fig. 7). Conversely, the oil bearing layers are recognized by evident deviations in deep (AT60–AT90) and shallow (AT10–AT20) resistivity logs (Liu et al. 2020) (Fig. 7).

Fig. 7

Well-log responses of typical reservoirs (water and oil bearing layer) in Yanchang Formation Member 7 of Ordos Basin, West China

Additionally, NMR logs can also be used for of the fluid type identification and estimation of fluid volumes (Anand 2017). Typical oil bearing layers have high T₂ amplitudes and wide T₂ spectrum or even contain tail distribution (Fig. 7) (Liu et al. 2020). Water bearing layers are characterized by low T₂ amplitudes and have narrow T₂ spectrum, containing no tail distributions (Fig. 7) (Liu et al. 2020). The T₂ spectrum of dry layer is very narrow and the T₂ amplitudes are low (Fig. 7).

4.2 Lithology

The unconventional hydrocarbon resources are mainly reserved in the fine-grained sedimentary rocks, which consist of carbonate, silt and clay (Zhao et al. 2019; Yang et al. 2019). In shale reservoirs, the various mineral compositions including felsic, clay, carbonate or even organic matters can form a complex lamina assemblage (Zhao et al. 2019; Wang et al. 2021). The complexity of unconventional resources requires an accurate evaluation of a petrophysical model for lithology prediction (Stadtmuller et al. 2018). Mudstone/shales mainly constitute the source rock intervals, while the interbedded thin layer of siltstone and/or carbonate rocks with a wide range of pore spaces from microscale to nanoscale will act as the reservoir rocks (Gao et al. 2016; Li et al. 2019; Liu et al. 2020).

Lithology can be identified by core observation, and the discontinuous core can be translated to the continuous petrophysical logs (He et al. 2019; Su et al. 2019). Wireline logs that can be used for lithology identification and prediction include GR, bulk DEN, CNL, AC, resistivity and image logs (Hsieh et al. 2005; He et al. 2019; Nhabanga et al. 2021; Venieri et al. 2021). The organic matter-rich shales and siltstones of Yanchang Formation Member 7 in Ordos Basin have distinct responses on the GR, AC and RT logs (Fig. 8). The black shales recognized on the core have very high GR, high resistivity and high sonic transit time (Venieri et al. 2021) (Fig. 8). For siltstones saturated with oil, the resistivity is also high, but has low GR and low sonic transit time compared with shales (Fig. 8). Additionally, the image logs reveal the internal laminated structure (bedding planes) of shales, while at the siltstone-shale contact surface, a scour surface can be observed (Ran et al. 2016) (Fig. 8). In addition, petrophysical inversion methods can also be used for mineral composition and lithology prediction (Doveton 2014; Ran et al. 2016).

Fig. 8

Well-log expressions of sandstones and shales in Yanchang Formation Member 7 of Ordos Basin, West China

Besides conventional logs, the advanced logs including LithoScanner logs and log processing methods of Quanti elemental log analysis (ELAN) are also useful for lithology identification (Stadtmuller et al. 2018). The prediction of lithology from wireline logs will help extend observations from core scale (centimeters to meters) to the well scale (meters or tens of meters) (He et al. 2019).

4.3 Reservoir Properties and Pore Systems

A series of geological and petrophysical measurements including thin section, scanning electron microscopy (SEM), mercury injection capillary pressure (MICP), nuclear magnetic resonance spectroscopy (NMR) and computed tomography (CT) can be used to characterize the various types of spaces spanning a wide range from nanometer scale to microscales (Josh et al. 2012; Lai et al. 2018b; Zhao et al. 2019; Liu et al. 2019; Du et al. 2021).

In the unconventional hydrocarbon resources, the interbedded siltstone or sandstone as well as carbonate rocks (dolomite, etc.) have anomalously high porosity (Zhao et al. 2019). The siltstone or sandstone intervals contain abundant intergranular pores and intragranular dissolution pores (Fig. 9a, b) (Lai et al. 2018b), while the dolomites (mainly dolomicrite) are dominated by intercrystalline pores as well as intercrystalline dissolution pores (Fig. 9c, d). Fracture and microfracture (aperture < 0.1 mm) also constitute the important reservoir pore spaces in unconventional hydrocarbon reservoirs (Fig. 9e, f).

Fig. 9

Thin section images showing the pore spaces of unconventional reservoirs. a Intergranular pores in siltstones, Yanchang Formation, Zhuang 234, 1297.07 m, b Intragranular dissolution pores in siltstones, Yanchang Formation, Zhuang 233, 1399.32 m, c Intercrystal and dissolution pores in dolomicrite, Lucaogou Formation, Ji 174, d Intercrystal pores and dissolution pores in dolomicrite, Lucaogou Formation, Ji 174, e Microfracture in tight sandstones, DB 14, 6349.34 m, f Note that the microfracture is the only pore spaces, DB 17, 6149.24 m

Shales have ultra-low porosity and the pore spaces (nanopores) are commonly below the resolution of optical microscope (10 μm), but can be easily detected by SEM images (Loucks et al. 2012; Josh et al. 2012; Zhao et al. 2019). The shale reservoirs mainly contain interparticle pores (Fig. 10a), intraparticle pores (Fig. 10b, c), organic matter pores (Fig. 10d) as well as microfractures (Fig. 10e, f) (Loucks et al. 2012; Josh et al. 2012; Su et al. 2018; Chandra and Vishal 2021).

Fig. 10

SEM images showing the pore spaces of oil shale reservoirs. a Interparticle pores, b Intraparticle pores, c Intraparticle pores, d Nanopores in organic matters, e Microfracture, f Microfracture

Reservoir quality evaluation using well logs aims at calculating porosity and permeability of the unconventional petroleum reservoir (Schmid et al. 2004). Reservoirs with relatively higher porosity and permeability are called “sweet spots” (Huang et al. 2017). Predicting porosity (including effective porosity), permeability and oil saturation (discussed below) from well-log data is a challenging task because core data are not available for all intervals or wells (Wood 2020b).

Porosity can be computed using the density log (Iqbal et al. 2018). The effective porosity, which is total porosity without the clay-bound water, can be calculated using the density neutron crossplot method, and also the NMR logs (Stadtmuller et al. 2018). For instance, the ratio of NMR T₂ components > 1.7 ms to the total T₂ components is calculated as effective porosity in Lucaogou Formation in Jimusar Sag of Junggar Basin, and the predicted results are in accordance with the core porosity (Fig. 11) (Wang et al. 2019).

Fig. 11

Calculation of porosity, permeability and oil saturation using NMR log in oil shale reservoirs of Lucaogou Formation in Jimusar Sag, Junggar Basin, West China

There are no direct well logs for permeability, but permeability can be derived from NMR logs, and there are two classical model models: the SDR model (Shlumberger Doll research center) (Eq. 4) and the Timur–Coates model (Eq. 5) (Fig. 11) (Coates et al. 1999; Yarmohammadi et al. 2020):

$$K_{{{\text{SDR}}}} = D\phi^{4} \left( {T_{{2{\text{gm}}}} } \right)^{2}$$

(4)

$$K_{{T{\text{imu}}}} = \left[ {\left( {\frac{\phi }{C}} \right)^{2} \left( {\frac{{{\text{FFI}}}}{{{\text{BVI}}}}} \right)} \right]^{2}$$

(5)

K_SDR (mD) is permeability from SDR model, and D is a constant (Glover et al. 2006; Rezaee et al. 2012). Where K_Timu is permeability in mD, φ is fractional NMR porosity, and FFI free fluid index (FFI) as well as the bulk volume irreducible (BVI). C is also a constant (Rezaee et al. 2012; Yarmohammadi et al. 2020).

Additionally, reservoir quality index (RQI), which was proposed by Amaefule et al. (1993) as the ratio of permeability to porosity under the square root, links the microscopic pore structure with macroscopic reservoir quality (Lai et al. 2016; Henares et al. 2016).

4.4 Oiliness and Oil Bearing Property

Core observation can show the oil bearing grade from low to high grade as fluorescence (no visible oil), oil trace, oil patch, oil immersion, etc. (Wu et al. 2017). Core observation under the fluorescence light will evidently reveal the oil bearing property (Fig. 12). The fluorescence scanning of core shows the varied degree of oil bearing for various lithologies. The carbonate intervals show strong fluorescence intensity, and the siltstone as well as the dolomite lamina is also fluorescent (Fig. 12).

Fig. 12

Core photo taken under the normal and fluorescence light to show the oil bearing property

Interparticle pores are favorable pore spaces for unconventional reservoirs, and almost all the edges of the particles emit strong fluorescences (Fig. 13a, b) (Liu et al. 2020). The intraparticle pores, especially those within carbonate particles, are fluorescent since the carbonate minerals (mainly microcrystalline dolomite) are oil-wet (Fig. 13c, d) (Xi et al. 2019). Organic matter pores as well as micropore associated with clay minerals emit scattered strong blue fluorescence (Fig. 13e, f) (Liu et al. 2020). Microfractures, especially those remain open status, emit strong fluorescences (Fig. 13g–h). The fluorescence thin sections reveal that almost all the entire pore systems in shales are fluorescent (Liu et al. 2021).

Fig. 13

Thin section images under plane polarized light and fluorescence light showing the microscopic oil bearing property. a Intraparticle dissolution pores, b Intraparticle dissolution pores are fluorescent, c Microcrystalline dolomite, the dark areas are organic matters, Ji19, 3820.83 m, d The dolomite particles are fluorescent, the same field view under fluorescence light of c, e Organic matter pores and clay minerals, f The organic matter pores and micropores within clays are fluorescent, g Filled microfracture, dark organic matters, h The microfracture emits blue fluorescence, the same field view under fluorescence light of G

Hydrocarbon saturation is also a vital petrophysical parameter, but is difficult to predict via well logs (Zhao et al. 2020). Archie’s equation, which is commonly used for fluid saturation calculation, may not be applicable for unconventional reservoirs (Clarkson et al. 2012; Li et al. 2021). The nondestructive NMR log has a distinctive advantage over conventional well logs in unconventional hydrocarbon reservoirs since it can provide petrophysical parameters of porosity, permeability and NMR T₂ amplitude and distribution (Deng et al. 2014; Guo et al. 2020; Li et al. 2020, 2021; Du et al. 2021; Zhang et al. 2021). Therefore, NMR logs are required to provide an accurate estimation of oil saturation (Wang et al. 2020). As is known, hydrocarbon is mainly associated with long T₂ components. Therefore, the NMR T₂ signal amplitudes longer than certain T₂ values (for instance, the threshold value of T₂ is set as 7.0 ms are for Lucaogou Formation in Jimusar Sag) can be treated as hydrocarbon signals, and consequently, the ratio of T₂ components larger than the threshold values to the total effective porosity is calculated as oil saturation (Wang et al. 2019) (Fig. 11). The calculated oil saturation is also in accordance with the core-measured oil saturation (Fig. 11).

4.5 Fracture

Unconventional oil and gas reservoirs (especially shales) have no natural productivity due to their complex pore structure and inherent strong heterogeneity (Zhang et al. 2021). Natural fractures are widespread in unconventional reservoirs (Gale et al. 2014; Lee et al. 2015; Li et al. 2018; Xu et al. 2020). Fractures not only provide pore space for fluid storage, but also greatly improve the reservoir performance, fluid flow, gas enrichment and hydrocarbon productivity (Curtis 2002; Zeng and Li 2009; Zeng et al. 2016; Li et al. 2018; Ladevèze et al. 2018; Basa et al. 2019; Zhang et al. 2021). Additionally, unconventional reservoir relies on hydraulic fracture stimulation (Curtis et al. 2012), and the preexisting (opening-mode) natural fractures will be further reactivated during stimulation and therefore enhance hydrocarbon productivity (Gale et al. 2007, 2014).

The fracture effectiveness of fracture is determined by the attitudes of fractures (high-angle, low-angle, horizontal and network fractures), scale of fractures (macroscopic and microfractures) and status fracture surface (open and closed fracture) (Gale et al. 2014; Hooker et al. 2017; Zhang et al. 2021). Consequently, the prediction and evaluation of subsurface fracture are important for unconventional hydrocarbon reservoir assessment (McGinnis et al. 2017; Lai et al. 2018a; Ladevèze et al. 2018).

Borehole image logs, which have a vertical resolution of 5 mm, are sensitive for the rock composition, structure and fluids in the formation (Ameen 2014; Zhang Shaolong et al. 2021). The presence of natural fractures will cause a rapid decrease in resistivity, and the fracture planes will appear as a sinusoidal wave on the image logs (Lai et al. 2021). Fractures in the Xujiahe Formation Member 2 of Sichuan Basin have caused rapid decrease in resistivity (Fig. 14). The fracture attitudes (dip and dip angles), fracture status (open, sealed, partly sealed) and fracture parameters (fracture length, aperture, porosity and density) can be picked out from the sinusoidal curves on image logs (Fig. 14) (Lai et al. 2019). Fractures can be divided into conductive and resistive types, and interpreted as partially or fully open and sealed fractures in terms of image log interpretation (Hooker et al. 2017). High-angle fractures will trace as sinusoids, whereas the planar or horizontal fractures intersect the circular wellbore (Fig. 14) (Hooker et al. 2017).

Fig. 14

Fracture responses on conventional and image logs in tight gas sandstones (Xujiahe Formation Member 2 of Sichuan Basin, West China)

The conventional well-log suits sensitive for the natural fractures include deep and shallow resistivity logs (Rt, Rxo), CAL as well as three porosity logs, especially AC, are sensitive for the presences of natural fractures (Khoshbakht et al. 2012; Zazoun 2013; Aghli et al. 2016; Lai et al. 2021). There is an evident decrease in lateral resistivity logs and density logs, while the AC values are significantly increased in the fractured intervals (Fig. 14).

Bed-parallel (horizontal) fractures are more prevalent in unconventional reservoirs (primarily fine-grained sedimentary rocks) than in sandstone or carbonate rocks (Fig. 15) (Gale et al. 2014). Layered structures are common in the fine-grained sedimentary rocks, which act as reservoirs for unconventional resources (Gale et al. 2014; Yawar and Schieber 2017). The multiple lamina or weak bedding interfaces, which are easily opened due to changing of in situ stress status, are favorable for formation of bedding parallel fractures (Zhang et al. 2017a, b). These bedding parallel fractures, which occur as horizontal fractures (Fig. 15), strongly influence fluid flow, and therefore hydrocarbon storage and productivity (McGinnis et al. 2017; Basa et al. 2019; Liang et al. 2021).

Fig. 15

Well-log responses of horizontal fractures in shale oil reservoirs (Yanchang Formation Member 7 of Ordos Basin, West China)

The bedding parallel fractures in Yanchang Formation Member 7 of Ordos Basin result in the reduction of resistivity and bulk density, but the sonic transit time increases (Fig. 15). Besides conventional logs and image logs, sonic scanner logs are also sensitive for the fractures (Zaree et al. 2016). The sonic transit time will increase, and the amplitudes of full wave forms will be attenuated, showing V-shape interferometric fringe in the fractured zones (Collett et al. 2011; Assousa and Elkington 2014; Zaree et al. 2016; Lai et al. 2017) (Fig. 15).

5 Engineering Property

Horizontal well drilling and multi-stage hydraulic (volume) fracturing are required for efficient exploitation of unconventional hydrocarbon resources due to their ultra-low matrix permeability (Curtis et al. 2012; Clarkson 2013; Fuentes-Cruz et al. 2014; Avanzini et al. 2016; Dong et al. 2018; Liu 2021). The implementation of advanced drilling and completion techniques significantly improve the successful production of hydrocarbons (Curtis et al. 2012; Clarkson 2013). Engineering property evaluates the brittleness, fracability, in situ stress anisotropy and magnitudes for unconventional hydrocarbon resources (Rybacki et al. 2016; Avanzini et al. 2016; Iqbal et al. 2018; Zhao et al. 2019; Yarmohammadi et al. 2020). Therefore, brittleness and in situ stress states and magnitudes are critical parameters for optimizing engineering sweet spots during hydraulic fracturing in unconventional reservoirs (Gale et al. 2007; Iqbal et al. 2018).

Brittleness evaluates the rock behavior of fracability during hydraulic fracturing (Verma et al. 2016; Iqbal et al. 2018). Brittle layers are easier to form fracture network than ductile layers (Soliman and Kabir 2012; Iqbal et al. 2018; Sun et al. 2021). The present-day maximum horizontal stress controls the geometry of the natural fracture system and direction of hydraulic fracture propagation, and therefore is important for design of hydraulic fracture treatment (Gale et al. 2007). Therefore, engineering property evaluation mainly focuses on the brittleness index and in situ stress fields (Avanzini et al. 2016; Iqbal et al. 2018).

5.1 Brittleness Index

Brittle layers with high brittleness index, which are easier to be fractured (brittle enough to initiate fractures), and to keep the fractures open, will be optimized for hydraulic fracturing in unconventional resources (Rickman et al. 2008; Sondergeld et al. 2010; Josh et al. 2012; Lai et al. 2015; Gholami et al. 2016; Iqbal et al. 2018; Nhabanga et al. 2021). In terms of geomechanical evaluation, brittleness is closely associated with elastic parameters Young’s modulus (E) and Poisson’s ratio (v) (Iqbal et al. 2018; Mews et al. 2019). Poisson’s ratio is the ratio of transverse to axial strain, and measures the rock’s ability to form fractures under stress. Young’s modulus is a ratio of stress to strain, and measures the rock’s ability to maintain fracture after treatment (Iqbal et al. 2018). Ductile rock may require more energy/fracturing pressure to break, and fractures formed by hydraulic fracturing in ductile rocks may easily be healed (Iqbal et al. 2018).

From a petrophysical point of view, there are two commonly used methods to calculate brittleness index. The first is the brittle mineral content, while the second is the elastic rock parameters (Young’s modulus and Poisson’s ratio) (Guo et al. 2015; Lai et al. 2015; Fan et al. 2019; Zhao et al. 2019; Wood 2021).

The brittleness index (%) derived from Young’s modulus and Poisson’s ratio can be defined as the average of the BI_E and BI_ν (Eqs. 6–8) (Lai et al. 2015; Iqbal et al. 2018; Fan et al. 2019). Brittle rocks have higher Young’s modulus and lower Poisson’s ratio (Rybacki et al. 2016; Zhang et al. 2016; Iqbal et al. 2018; Kumar et al. 2018). High Young’s modulus is associated with layers of low porosities but high brittle minerals (quartz, carbonate and feldspar) (Liu et al. 2018).

$${\text{BI}} = \frac{{{\text{BI}}_{E} + {\text{BI}}_{\nu } }}{2} \times 100\%$$

(6)

$${\text{BI}}_{E} = \frac{{E - E_{\min } }}{{E_{\max } - E_{\min } }}$$

(7)

$${\text{BI}}_{\nu } = \frac{{\nu - \nu_{\max } }}{{\nu_{\min } - \nu_{\max } }}$$

(8)

where E (GPa) is Young’s modulus and v (dimensionless) is Poisson’s ratio. v_min and v_max are the minimum and maximum Poisson’s ratio, whereas E_min and E_max are the minimum and maximum Young’s modulus.

The dynamic Young’s modulus and Poisson’s ratio can be calculated from V_p, V_s and bulk density logs (Eqs. 9–10) (Lai et al. 2015).

$$E = \frac{\rho }{{V_{{\text{s}}}^{2} }}\frac{{3V_{{\text{p}}}^{2} - 4V_{{\text{s}}}^{2} }}{{V_{{\text{p}}}^{2} - V_{{\text{s}}}^{2} }}$$

(9)

$$\nu = \frac{{V_{{\text{p}}}^{2} - 2V_{{\text{s}}}^{2} }}{{2(V_{{\text{p}}}^{2} - V_{{\text{s}}}^{2} )}}$$

(10)

where ρ is the bulk density log (kg/m³), while V_p is the P (compressive) wave velocity (m/s), and V_s is S (shear) wave velocity (m/s) (Lai et al. 2015).

Dynamic Young’s modulus and Poisson’s ratio calculated using sonic and density logs should be calibrated with the static elastic parameters (core-measured) to improve accuracy (Iqbal et al. 2018).

Jarvie et al. (2007) defined the mass ratio of quartz to all minerals as index of brittleness. However, besides quartz, the carbonates are also brittle minerals (Jarvie et al. 2007; Rybacki et al. 2016; Fan et al. 2019; Qian et al. 2020).

The method to calculate brittleness index using brittle mineral content is written as (Eq. 11).

$${\text{BI}} = \left( {{\text{Qz}} + {\text{Car}}} \right)/\left( {{\text{Qz}} + {\text{Car}} + {\text{Fels}} + {\text{Clay}}} \right) \times 100\%$$

(11)

Qz is quartz content, %, Car is the carbonate content, %, Fels is the feldspar content, %, and Clay is the total clay content by weight, %. Rocks with brittleness index > 40% are treated as brittle rocks (Guo et al. 2015; Lai et al. 2015; Iqbal et al. 2018).

The elastic parameters (v and E) can be calculated from the well logs of V_p, V_s and density logs. Then, the Poisson’s ratio–Young’s modulus method can be adapted to calculate brittleness index (Kumar et al. 2018). Layers with low Poisson’s ratio and high Young’s modulus contribute to a high brittleness index, and they are favorable for hydraulic fracturing, and maintaining as well as propagating fractures (Fig. 16) (Rybacki et al. 2015; Kumar et al. 2018).

Fig. 16

Comparison of brittleness index calculated by Young’s modulus–Poisson’s method and by ECS logs

Additionally, the ECS logs, which can derive the rock compositions of clay, Q–F–M (quartz, feldspar and mica), carbonate, etc., can be used to calculate the brittleness index using the method of brittle mineral ratio (Fig. 16) (Maliva et al. 2009; Lai et al. 2015; Kumar et al. 2018). The brittleness index calculated by the two methods (mineralogy method and elastic parameter method) is generally in accordance with each other (Fig. 16).

5.2 In Situ Stress Status and Direction

The in situ stress fields commonly include vertical stress (S_v), direction and magnitudes of maximum horizontal stress (SH_max) and minimum horizontal stress (Sh_min), as well as formation pressure (P_p) (Zoback et al. 2003; Verweij et al. 2016; Dixit et al. 2017; Lai et al. 2019).

Besides brittleness index, in situ stress fields also play critical roles in hydraulic fracturing and horizontal well trajectory design (Qian et al. 2020). For engineering “sweet spot” evaluation, brittleness measures the ability to form fracture and keep fracture open, while the in situ stress field (direction and magnitude) evaluates the horizontal well trajectory design and propagation of hydraulic fractures (Josh et al. 2012). A large amount of high-pressure fluids are injected into the formation to reopen the natural fracture system and create new hydraulic fractures during hydraulic fracturing (Rybacki et al. 2016; Zhao et al. 2019). Consequently, a complex and new pore–fracture network system will be formed and hydrocarbon will migrate toward the wellbore for production (Zhao et al. 2019).

Therefore, the magnitude and direction of in situ stresses are also required to optimize the prospected layers for hydraulic fracturing (Josh et al. 2012; Iqbal et al. 2018). The direction of in situ stress fields determines the drilling direction of horizontal wells, and the propagation of hydraulic fractures. Brittleness alone is not sufficient to optimize prospected layers, and the magnitudes of in situ stress are also an important index for optimizing prospected layers for engineering sweet spots. Complex and new pore–fracture systems will be formed in layers with low horizontal stress differences.

5.2.1 Direction of In Situ Stress

The brittleness affects the formation and preservation of natural fractures, while the in situ stress fields control the initiation and propagation of hydraulic fractures (Fig. 17) (Josh et al. 2012; Rybacki et al. 2016). Brittle rocks are expected to contain more natural fractures and are more easily to be fractured by hydraulic stimulation (Gale et al. 2007; Rybacki et al. 2016; Zhang et al. 2016). In addition, the hydraulic fractures will propagate along the direction of maximum horizontal stress (SH_max) (Kingdon et al. 2016; Iqbal et al. 2018). Hydraulic fractures follow SH_max, until they encounter the preexisting natural fractures, and then, the hydraulic fractures will be blocked from further propagation (Gale et al. 2007). Consequently, the horizontal wells are drilled along the minimum horizontal stress (Sh_min), and hydraulic stimulation will be toward the SH_max direction (Josh et al. 2012; Iqbal et al. 2018). In this situation, a large number of hydraulic fractures will be formed along the SH_max direction, and these induced fractures will intersect with the preexisting fractures to form complex fracture networks (Fig. 17).

Fig. 17

Horizontal drilling direction, hydraulic fracture propagation and in situ stress field in unconventional hydrocarbon resources

Image logs, which can pick out the borehole breakouts and induced fractures, are widely used for the determination of SH_max and Sh_min (Lai et al. 2018a; Stadtmuller et al. 2018). The borehole breakouts, which appear as broad, parallel, dark bands with 180° apart on image logs, indicate the orientations of Sh_min (Massiot et al. 2015; Nian et al. 2016) (Fig. 18a). The drilling induced fractures are recognized as two vertical fractures (“two ways”) with 180° offset at the borehole surfaces on the image logs (Fig. 18b), and they show the orientations of SH_max (Ameen et al. 2012; Khair et al. 2013; Nian et al. 2016; Lai et al. 2019). Consequently, both the induced fractures and borehole breakouts can unravel the in situ orientation. Natural fractures can be distinguished from drilling induced fractures and borehole breakouts on image logs by their continuous sinusoid nature (Fig. 14) (Khair et al. 2015).

Fig. 18

In situ stress direction determined from image logs

Besides image logs, the sonic logs, which provide the shear-wave velocities and direction, can reveal the in situ stress fields (Stadtmuller et al. 2018). As is known, in anisotropic rocks, the shear wave will be split into fast and slow waves, i.e., shear-wave birefringence (Liu et al. 2018). Therefore, fast S-wave azimuth indicates the SH_max direction (Liu et al. 2018; Stadtmuller et al. 2018) (Fig. 19). The SH_max direction can be determined from the fast S-wave azimuth as near NW–SE to S–E direction (Fig. 19).

Fig. 19

Maximum horizontal stress (SHmax) direction from sonic fast shear-wave orientation

5.2.2 In Situ Stress Magnitudes

In situ stress magnitudes play vital roles in different aspects of hydraulic fracturing treatment (Iqbal et al. 2018). The difference between SH_max and Sh_min (SH_max–Sh_min) is important for directional drilling, hydraulic fracturing design, and optimization of engineering “sweet spots” (Stadtmuller et al. 2018). Intervals with low differences between SH_max and Sh_min are suggested to be easily fractured and therefore will be optimized for hydraulic fracturing.

The vertical stress (Sv) is commonly calculated by integrating the weight of the overburden rocks by well logs (Eq. 12) (Fig. 20) (Maleki et al. 2014; Verweij et al. 2016; Iqbal et al. 2018; Lai et al. 2022).

$$S_{{\text{v}}} = \int_{{0}}^{{\text{Z}}} {\rho {\text{g}}dz}$$

(12)

In this formula, Z is the burial depth, m, g is the gravitational acceleration (m/s²), and ρ is the bulk density, kg/m³ (Maleki et al. 2014; Verweij et al. 2016).

Fig. 20

The in situ stress fields (direction and magnitudes) and brittleness index calculated from well logs

The two components of horizontal stress (SH_max, Sh_min) are closely associated with elastic modulus (Young’s modulus and Poisson’s ratio) (Du et al. 2021). The one-dimensional mechanical earth model is commonly adopted to calculate the horizontal stresses using poroelastic theory (Eqs. 13, 14) (Fig. 20) (Engelder 1993; Zoback et al. 2003; Stadtmuller et al. 2018; Lai et al. 2022).

$${\text{SH}}_{{{\text{max}}}} = \frac{\nu }{1 - \nu }(Sv - \alpha P_{{\text{p}}} ) + \alpha P_{{\text{p}}} + \frac{E\varepsilon }{{1 - \nu^{2} }}$$

(13)

$${\text{Sh}}_{\min } = \frac{\nu }{1 - \nu }({\text{Sv}} - \alpha P_{{\text{p}}} ) + \alpha P_{{\text{p}}} + \frac{E\nu \varepsilon }{{1 - \nu^{2} }}$$

(14)

In Eqs. (13) and (14), α is the Biot’s coefficient. P_p is pore pressure, and can be derived from the Eaton’s method. The ε is strain factor (Stadtmuller et al. 2018; Iqbal et al. 2018). Consequently, the horizontal stresses can be calculated from Sv, E, v, Biot’s coefficient, strain factor and pore pressure (Iqbal et al. 2018).

The tectonic regimes proposed by Anderson (1951) can be determined from the relative amplitudes of the three stress components, and they include normal (Sv > SH_max > Sh_min), strike-slip (SH_max > Sv > Sh_min) and thrust faulting stress regime (SH_max > Sh_min > Sv) (Zoback et al. 2003; Verweij et al. 2016; Dixit et al. 2017; Stadtmuller et al. 2018; Lai et al. 2019). The strike-slip faulting stress regime is encountered in Fig. 20 (Lai et al. 2022).

Engineering property evaluation should comprehensively take brittleness index as well as in situ stress direction and magnitude into consideration. The horizontal wells are designed to drill parallel to Shmin, with the aim to maximize the volume stimulated by induced fractures (Gale et al. 2007). The dominant SH_max direction (fast shear wave) is NE–SW direction (Fig. 20) (Lai et al. 2022). Then, the layers with high brittleness index but low horizontal principle stress differences will be optimized for hydraulic fracturing in order to create the most abundant hydraulic fracture networks (Stadtmuller et al. 2018).

6 Optimization of Sweet Spots

Unconventional hydrocarbon resources have no natural productivity; therefore, the identification of sweet spots is important (Zou et al. 2019). Sweet spots in unconventional resources refer to the best zones or intervals for hydrocarbon exploration and exploitation, and there are mainly geological and engineering sweet spots (Lu et al. 2019; Zhao et al. 2019; Zou et al. 2019). Geological sweet spots are the zone or intervals with the best reservoir quality and oil bearing property, and they can be optimized considering reservoir property (lithology, porosity, oil saturation), and presence of natural fractures (Zhao et al. 2019; Zou et al. 2019). Engineering sweet spots refer to the zone or intervals prospected for hydraulic fracturing stimulation, and therefore, brittleness and in situ stress anisotropy are the critical parameters to be evaluated (Rybacki et al. 2016; Iqbal et al. 2018; Zhao et al. 2019).

Insights into the seven kinds of parameters (lithology, reservoir quality, hydrocarbon bearing property, well-log responses, source rock property, brittleness and in situ stress field) and three types of properties (source rock property, reservoir property and engineering property) lay the foundation for sweet spot optimization using well logs (Zou et al. 2019). Geological sweet spot evaluation aims at selecting the favorable hydrocarbon bearing reservoirs; therefore, the lithology, porosity, hydrocarbon saturation, fracture and source–reservoir assemblage should be evaluated. Engineering sweet spot evaluation focuses on optimizing the prospected layers for hydraulic stimulation, and therefore, evaluation of geomechanical property (brittleness index, in situ stress) is particularly important (Zhang et al. 2017a, b; Iqbal et al. 2018; Zhao et al. 2019). Both the petrophysical attributes and geomechanical properties need to be fully understood to identify the prospected sweet spots (Iqbal et al. 2018).

6.1 Relationships Between the Three Types of Properties

Relationships between the source rock property, reservoir property and engineering property should be unraveled when selecting the prospected layers for production. TOC is an excellent indicator of organic matrix and hydrocarbon potential as well as source rock property, and TOC can be continuously evaluated via well logs (Amosu et al. 2021). Reservoir property in terms of porosity, permeability and oil saturation can be predicted by integrating conventional logs and NMR log (Wang et al. 2019). Brittleness and in situ stresses, which describe the engineering property, can be estimated according to the elastic parameters using sonic logs (Iqbal et al. 2018; Zhao et al. 2019).

Source rock intervals in Funing Formation in Subei Basin (East China), which have high TOC content, have low reservoir property as can be observed from the crossplot of porosity versus TOC content (Fig. 21). In addition, TOC shows negative correlation relationships with horizontal stress difference (SH_max–Sh_min) (Fig. 22a), indicating that source rock intervals can also be fractured due to the low horizontal stress differences. The brittleness index shows complex relationship with TOC and rocks with brittleness index 40–60% have the highest TOC values, indicating brittleness is complex reflection of lithology, composition, TOC and diagenesis (Clarkson et al. 2012; Iqbal et al. 2018) (Fig. 22b).

Fig. 21

Crossplot of porosity versus TOC of Funing Formation in Subei Basin, East China

Fig. 22

Crossplots of TOC versus horizontal stress difference and brittleness index in Funing Formation in Subei Basin, East China

Petrophysical properties affect the geomechanical properties of unconventional reservoirs (Iqbal et al. 2018; Zhao et al. 2019). Engineering property shows complex relationships with reservoir property as can be indicated by the crossplots of brittleness index and horizontal stress difference with porosity (Fig. 23a, b). Therefore, the reservoir property, which describes the geological sweet spots, is not matching the engineering property characterizing the engineering sweet spots (Fig. 23).

Fig. 23

Crossplots of horizontal stress difference and brittleness index versus porosity in Funing Formation in Subei Basin, East China

6.2 Optimization of Geological and Engineering Sweet Spots

The three types of properties determine the distribution of sweet spots (Zhao et al. 2019). The source rock property is evaluated in terms of TOC content as well as the conventional well-log responses (Fig. 24). Lithology is predicted by ECS log profile. Then, porosity, permeability and oil saturation are determined from the NMR logs, and NMR T₂ spectrum is also presented to show the fluid bearing property. Additionally, the image logs are used to pick out the fracture traces and derive the in situ stress direction using induced fractures. Consequently, the reservoir property can be evaluated in terms of reservoir quality (porosity, permeability), oil bearing property (oil saturation and T₂ spectrum) and the presences of natural fractures (Fig. 24). The in situ stress profiles including vertical/overburden stress and maximum/minimum horizontal stress are calculated by density and sonic logs. Brittleness index is determined using Poisson’s ratio and Young’s modulus (Fig. 24). Additionally, the in situ stress anisotropy profile is generated by picking out the fast shear-wave direction (Fig. 24).

Fig. 24

Comprehensive diagram of the seven types of relationships and three types of properties in Well Cheng 96 in Ordos Basin, West China

The best source rock intervals are easily recognized by the conventional well logs as well as the calculated TOC content (Fig. 24). The best reservoir property intervals can be distinguished by the wide and high amplitudes of NMR T₂ spectrum. Additionally, the high values of calculated petrophysical parameters of porosity, permeability and oil saturation prove the presences of best reservoir property. Furthermore, the appearances of natural fractures will improve the permeability and form favorable geological sweet spots (Fig. 24). The interconnectivity of the source rock and reservoir is also important for sweet spots in unconventional resources (Kumar et al. 2018; Zhao et al. 2019; Radwan et al. 2021). Therefore, the reservoirs adjacent with the source rock intervals have the best potential for accumulating oil and gas resources (Fig. 24).

The economic production of unconventional reservoirs strongly relies on the hydraulic induced fracture system (Fuentes-Cruz et al. 2014; Iqbal et al. 2018). The direction of stimulation treatments is toward the SH_max direction in order to create complex fracture network for oil and gas to flow into the borehole since the hydraulic fractures propagate along SH_max (Fig. 17). In addition, the prospected layers optimizing for hydraulic fracturing are those layers with high brittleness index but low horizontal stress differences (Zhang et al. 2017a, b). Layer with high brittleness index (> 0.5) but low horizontal stress differences (< 15 MPa) are optimized for hydraulic fracturing layers (Fig. 24).

7 Prospects

Geophysical well logs play important roles in the exploration and development of unconventional hydrocarbon resources. However, the high cost of acquiring a comprehensive suite of the required well logs (especially advanced well logs including spectral gamma-ray, NMR, image logs, ECS, etc.) will hinder the application of geophysical well logs in the fields of unconventional hydrocarbon resources. Therefore, the basic suite of well logs should be optimized. Conventional well logs are necessary for source rock property evaluation, and image logs as well as NMR logs need to be logged with the aim for reservoir property evaluation. Array sonic logs and image logs should be optimized for engineering property determination. In addition, the calibration of well-log data with core analysis data will reduce uncertainty, and the optimum sampling of core data will help improve the accuracy of well-log data interpretation.

Geophysical well logs precisely evaluate the petrophysical properties (lithology, porosity, permeability and oil saturation) and geomechanical properties (Poisson’s ratio, Young’s modulus, brittleness index and in situ stress) of unconventional hydrocarbon resources (Avanzini et al. 2016; Kumar et al. 2018). Consequently, the well logs are widely adopted to answer the key questions in unconventional oil and gas geology and engineering: “Whether there contains abundant oil and gas resources?”, “Where are the hydrocarbon reserved?” and “How to optimize the prospected layers for stimulation?”. The comprehensive petrophysical approaches will support petroleum geoscientists’ and engineers’ decisions throughout whole life of unconventional hydrocarbon resources including horizontal drilling, resource assessment, reservoir characterization and hydraulic fracture simulation (Avanzini et al. 2016).

8 Summary and Conclusions

The advanced well-log series used in unconventional oil and gas play evaluation include high definition induction logs, image logs, array acoustic logs, nuclear magnetic resonance (NMR) log, elemental capture spectroscopy (ECS) as well as LithoScanner logs.

Source rock intervals are recognized on the conventional well logs as high GR, low density, high CNL, AC and resistivity values. TOC can be predicted by the ΔlogR method, spectral GR logs, multivariate fitting method and LithoScanner logs.

Lithology can be predicted by conventional and image logs. Porosity, permeability and oil saturation can be calculated from NMR logs combined with conventional well logs. Fracture can be picked out from the image logs. Reservoir property can be evaluated from the combination of lithology, reservoir quality and the presences of natural fractures. Intervals with high reservoir quality and hydrocarbon bearing property and adjacent with the source rock interval are screened out as the geological sweet spots.

Brittleness can be determined not only from Young’s modulus also from rock composition. Brittleness should be well evaluated in terms of fracture initiation and propagation, as well as keeping fracture reopening. Additionally, in situ stress direction and magnitudes also need well understood in terms of horizontal well drilling direction and optimizing propertied hydraulic fracturing intervals. Layers with high brittleness index but low horizontal stress difference are optimized as engineering sweet spots.

Geophysical well logs can characterize the seven kinds of parameters (lithology, reservoir quality, hydrocarbon bearing property, electronic well-log responses, source rock property, brittleness, and in situ stress magnitude and direction) and three kinds of properties (source rock property, reservoir property and engineering property), and therefore will be widely used in optimizing sweet spots in unconventional play in the future.