Seismic inversion is a widely used method for exploring and characterizing subsurface geological structures, especially in the context of oil and gas exploration. The data-driven approach exemplified by deep learning (DL) circumvents the need for pre-defined physical system and is used to solve seismic inversion problems. However, the DL-based inversion method is sensitive to noise and is typically performed trace-by-trace. When the inversion results are aggregated into a 2D image, the lateral continuity of the final output is often inadequate, impacting subsequent interpretation and evaluation. In contrast, conventional DL-based multi-trace seismic inversion typically treats multi-trace seismic data as input without fully accounting for the coupling relationships between neighboring traces. Therefore, we propose a lateral constrained multi-trace seismic inversion method based on DL to enhance the continuity and geological reliability of the inversion results. Given the similarities among neighboring traces, the method employs adjacent multi-trace seismic data as input to the network and designs multi-trace coupling constraints to ensure the lateral consistency of the prediction outcomes. Moreover, physical laws and low-frequency prior information are incorporated into the network training process to mitigate the dependence of data-driven methods on large amounts of training data. The effectiveness of the proposed method in enhancing both the lateral continuity and accuracy of inversion results is demonstrated by applying it to synthetic and real datasets, and comparing the results with those of conventional DL-based single-trace and multi-trace inversion methods.

This study investigates reservoir changes resulting from increased pore pressure during fracturing and develops a petrophysical model that incorporates pore pressure. We establish relationships between pore parameters, fracture parameters, and pore pressure using the pore compression coefficient, simulating the deformation of both pores and fractures. And we classify the effects of increased pore pressure on fractures into two phases: the fracture deformation phase and the fracture propagation phase. Additionally, we simulate the vertical expansion of primary pores within hydraulic fractures, considering secondary fracture formation mechanisms. Finally, we perform a forward hydraulic fracturing analysis using shale reservoir parameters from the Ordos Basin, China, accounting for various influencing factors. This research provides a comprehensive framework for modeling hydraulic fracturing and offers theoretical insights for evaluating fracture growth.

ABSTRACT Understanding the effect of reservoir petrophysical properties on wave-induced fluid pressure is crucial for fluid discrimination in seismic exploration. However, subsurface rocks typically contain multiple pore types filled with different fluids, and the resulting complexity of reservoir petrophysical properties limits effective pore-fluid identification. To elucidate pore-fluid pressure variations, a porous rock physics model was developed, introducing a relative fluid pressure parameter to quantify wave-induced pressure changes. This parameter reliably captured fluid pressure responses across diverse pore structures and provided a crucial foundation for accurate fluid identification. Our analysis showed that fluid mobility, along with heterogeneity in pore structure and fluid distribution, governed pore-fluid pressure. The relative fluid pressure was generally higher in cracks than in stiff pores, with both exhibiting pronounced sensitivity to gas-bearing reservoirs. The proposed relative fluid pressure acted as an intermediate parameter linking complex reservoir petrophysical properties to seismic responses, providing a fluid-pressure-based approach for precise characterization of fluids in hydrocarbon-bearing reservoirs.

Petroleum science stands at a historic turning point: exploration extends to ultra-deep and deepwater frontiers, unconventional resources become dominant, and the dual imperatives of carbon neutrality and artificial intelligence are reshaping the industrial landscape and disciplinary boundaries. Building upon the “100 Grand Challenges in Petroleum Science” initiative launched by Petroleum Science in 2021, this paper presents a curated list of 100 fundamental scientific problems through systematic literature synthesis, expert deliberation, and multi-criteria selection. The challenges span six domains: exploration and development, transport and pipelines, refining and petrochemicals, materials science, carbon capture, utilization and storage (CCUS)-geothermal-hydrogen, and energy economics and digital transformation. In-depth analysis reveals three archetypes: mechanism-oriented problems that seek unified theories of multi-scale coupling, technology-oriented problems that push engineering limits under extreme conditions, and system-level problems that address full-chain integration and energy transition. A key insight is that artificial intelligence has evolved from an enabler to a core driver, and together with the low-carbon transition it is redefining the very paradigm of petroleum science. This roadmap is intended to guide research planning, investment, and talent development over the coming decades and to foster interdisciplinary collaboration on a global scale.

The effective pore-fluid bulk modulus, as a crucial fluid identification factor, has garnered significant attention from geophysical inversion researchers in recent years. Currently, existing methods targeting this parameter are adopt linear approximations of the exact Zoeppritz equations as forward operators. However, such linear approximations are known to have limited computational accuracy, and their derivation processes involve numerous assumptions. These intrinsic constraints significantly compromise the estimation accuracy and applicability of existing methods. To address these challenges, we first derive a new form of the Zoeppritz equations that explicitly incorporates the effective pore-fluid bulk modulus, to establish a high-quality forward modeling framework for inverting this key parameter. Subsequently, we analyze the computational accuracy of the new derived equations and the key factors influencing it using the layered numerical models. Numerical analysis indicates that the accuracy of the new equations is affected by the critical porosity and the dry-rock VP/ VS ratio squared, implying that these two parameters need to be reasonably estimated in advance. In response, we propose the corresponding estimation methods for these two parameters based on well logging data and validated their effectiveness. Finally, we construct a Bayesian nonlinear inversion objective function and employed a generalized nonlinear inversion (GNI) approach to achieve fast and stable solutions, thereby establishing a novel nonlinear inversion method specifically for the effective pore-fluid bulk modulus. Tests on both synthetic and field data demonstrate that the proposed method can deliver high-accuracy estimates of target parameter, significantly outperforming existing method based on the approximation formula. These results clearly demonstrate the feasibility and effectiveness of the proposed approach.

Conversion from distributed acoustic sensing (DAS) measurements to geophone-equivalent data is important for integrating DAS into conventional seismic workflows. This is because most established seismic-processing algorithms are designed for particle-velocity or acceleration data, whereas DAS measures strain or strain rate. Recovering geophone-equivalent particle velocity from DAS strain-rate measurements requires inversion of a gauge-length-dependent spatial-difference operator, which can amplify measurement noise, particularly in field data with low signal-to-noise ratios (SNRs). Existing single-regularization methods often trade noise attenuation against waveform fidelity and the preservation of weak coherent events. To address these limitations, we propose an inverse reconstruction framework combining high-order Lp (HOLp) and overlapping group sparsity (OGS) regularizations. HOLp promotes a compact representation of second-order differences and suppresses incoherent fluctuations, whereas OGS exploits local coherence to reduce isolated artifacts and preserve weak continuous events. The resulting objective function is solved using the alternating direction method of multipliers, with iteratively reweighted L1 minimization for the HOLp subproblem and a majorization-minimization strategy for the OGS subproblem. Numerical and field experiments confirm that the method restores amplitude and waveform fidelity under low SNR conditions, demonstrating robust and reliable DAS-to-geophone conversion.

Traditional inversion methods rely on prior information to obtain a stable and unique solution to ill-posed problems. This stability is typically enforced through regularization terms derived from statistical priors, combined with geophysical data using Bayes' rule. Regularization leads to an optimization problem where the goal is to simultaneously fit the data while minimizing a prior-induced constraint that imposes certain characteristics on the estimated model. However, these constraints can introduce solution styles that may oversimplify geological features. For instance, quadratic regularization produces overly smooth solutions, potentially omitting critical details, while edge-preserving regularization can lead to unrealistic models. Recently, diffusion models have emerged as a flexible and robust approach for approximating complex, high-dimensional probability distributions, offering more expressive priors for solving inverse problems. Their ability to learn and reverse a noisy forward process makes them particularly useful for generating realistic geological models when trained appropriately. Inspired by this, we propose a diffusion model-based method for acoustic impedance (AI) inversion. The approach is not limited to impedance inversion but can be applied to any geophysical inverse problem where enough earth-representative synthetic models are available. Our method first trains a diffusion model on synthetic AI images to approximate their distribution. We then apply diffusion posterior sampling, conditioned on seismic data, to estimate an impedance model that aligns with the observed data. Specifically, the diffusion posterior sampling integrates the gradient of the error between observed and reconstructed data, guiding the reverse diffusion process toward solutions that are consistent with observed seismic traces. We incorporate 3D lateral constraints and gradient momentum during the reverse sampling process, further improving the continuity and stability of 3D AI inversion. We demonstrate the feasibility of using diffusion models as an alternative to traditional regularization techniques for AI inversion through synthetic and real data examples.

Shale reservoirs often exhibit low porosity and low permeability. Accurate characterization of brittleness, particularly the anisotropic brittleness index, is essential for effective hydraulic fracturing. However, it is difficult to obtain the stiffness coefficient c12 from seismic data in transversely isotropic (VTI) media, which limits the direct extraction of the transversely isotropic brittleness index (EBIVTI). We propose an approximate formula for c12 and a method to extract EBIVTI directly from seismic data. Initially, we derive an approximate formula for EBIVTI using petrophysical analysis. Its accuracy is verified with logging data. Subsequently, we construct a Jacobian matrix that includes horizontal slowness, based on the exact reflection coefficient equation. This supports EBIVTI extraction from pre-stack seismic data. Finally, the proposed inversion strategy is applied to both a synthetic well and field data from a shale reservoir. The results demonstrate that the method is accurate and robust against noise, suggesting its potential for brittleness evaluation in shale reservoirs.

Structure-guided inversion has been increasingly adopted to enhance structural features and improve the interpretability of seismic impedance imaging. However, the limited bandwidth of seismic wavelets leads to low-resolution poststack records, which inevitably compromises the accuracy of structural estimates in discontinuous regions. Consequently, the success of this approach critically depends on the appropriate use of structural information as geometric constraints. To address this limitation, we propose an adaptive structure-guided approach that maintains large-scale structural continuity while preserving fine-scale variability. The average structural similarity (ASSIM) index and structural constraint sensitivity (SCS) index are introduced to adjust the structural penalty intensity across adjacent traces. The former metric determines the upper bound of penalty intensity for uniform structural constraints, while the latter one characterizes regions prone to distortion where a local reduction in lateral smoothing is required. In addition, the sparse constraint is separated from the objective function and implicitly applied via an iteratively refined reference model, which is generated by an adaptive edge-preserving filter (AEPF) followed by L2-TV denoising. This combined strategy promotes blocky impedance reconstruction and effectively preserves thin-layer details. Numerical and field examples demonstrate the superiority of the proposed approach. These improvements establish a more rational and geologically consistent framework for incorporating prior structural information, thereby enhancing the practicality and imaging performance of the structure-guided inversion in complex scenarios.

Fractures represent a critical structural feature in unconventional reservoirs, as they create essential pathways for the migration and accumulation of oil and gas. Therefore, fracture characterization is a fundamental task in the exploration of unconventional hydrocarbon resources. Conventional fracture characterization methods typically do not account for the inherent anisotropy of the formation, which arises from the sedimentary environment and fluid distribution, often leading to inaccurate fracture predictions. To address this challenge, we propose a petrophysical model that incorporates inherent anisotropy, employing rock physics modelling to accurately characterize fracture distribution. Furthermore, to reduce the substantial workload involved in manually calibrating the petrophysical model, we introduce a one-dimensional convolutional neural network combined with an attention mechanism. By leveraging the advanced nonlinear learning capabilities of the convolutional neural network, we aim to fit the petrophysical model and extend its application across all exploration wells and the entire field. The effectiveness and feasibility of the proposed method are demonstrated through experiments using actual borehole data from a fracture-dominated reservoir.

Time-lapse (TL) seismic technology has been widely adopted across various geophysical domains by quantifying reservoir dynamics through the analysis of discrepancies in time-lapse seismic data. Currently, established TL seismic inversion methods are predominantly in the time domain. However, owing to advancements in depth migration technology, depth-domain imaging is now recognized for its superior imaging accuracy. Relative to time-domain imaging data, extracting requisite inversion outcomes directly from depth-domain seismic data preserves more precise TL disparity information. In this article, we propose a method that directly performing pre-stack TL seismic inversion in the depth domain to obtain time- lapse interpretation results in depth domain. In the article, we use a 1D point spread function to calculate seismic wavelets in the depth domain, and use the precise Zoeppritz equation to calculate the reflection coefficient. To enhance the stability of multi-parameter inversion, we establish an inversion objective function based on Bayesian framework and add low-wavenumber constraints and block constraints. Through a synthetic test, we compare the separate inversion and differential inversion in the depth domain, and results show that TL seismic simultaneous inversion is more suitable for depth-domain applications. Finally, we validate the algorithm using real data and obtain high-quality application results.

Effective stress is a key indicator for guiding hydraulic fracturing, predicting pore pressure, and identifying reservoir fluids. Accurate prediction of effective stress holds substantial practical importance for reservoir characterization. Currently, two primary methods are employed for effective stress prediction: indirect calculation and inversion based on linear approximation equations. However, indirect calculation introduces cumulative errors and fails to fully use the information in seismic data, whereas inversion based on linear approximation equations suffers from low resolution and limited accuracy. To address these limitations and obtain high-quality reservoir characterization parameters, we propose a novel approach by integrating the rock physics model with the Zoeppritz equation under the assumption of isotropic media. This integration yields a new mathematical expression for effective stress prediction. An objective function is constructed based on Bayesian theory, incorporating the modified Cauchy constraint as a regularization term to enhance stability and accuracy. The Gauss–Newton method is then applied to optimize the objective function, yielding inversion results. Both synthetic and real data tests demonstrate that the proposed method significantly improves precision and resolution compared to traditional approaches while maintaining robust stability. This advancement provides more reliable and effective information for reservoir characterization, offering valuable insights for future exploration and development efforts.

Seismic acoustic impedance (AI) inversion is essential for reservoir prediction and characterization. In recent years, deep learning has shown immense potential as a data-driven approach in seismic data processing, inversion, and interpretation. As a data-driven method, deep learning-based seismic inversion results better when sufficient labeled data are provided. Overfitting and poor generalization often occur when labels are insufficient. Due to the lack of labeled data in seismic inversion problems, the difficulty of inversion increases, leading to unstable and poor generalization of prediction results. To partially address this issue, we propose a constrained seismic inversion strategy. Since seismic records are time series, we exploit the convolutional neural network (CNN) and bidirectional LSTM (Bi-LSTM) network structures that are more applicable to time series. We combine the physical model and the initial model as constraints to improve the network stability and generalization ability, and impose sparse constraints on the reflection coefficient to further improve the prediction accuracy. The network structure transformation improves the efficiency and stability of the training process. Through numerical experiments and real data tests, it is proved that the proposed method improves the vertical resolution and geological reliability, providing a more stable and efficient method for seismic inversion under conditions of limited labeled data. The overall performance improved by 2% through comparative analysis.

Predicting reservoir parameters with high accuracy is still a crucial work of oil reservoir exploration and development. Due to the limitation of computational efficiency, deterministic methods are primarily used in practical production applications for predicting reservoir parameters. When the nonlinear forward equations are exceptionally complex and the initial model constructed deviates significantly from the true reservoir parameters, deterministic methods may have difficulty obtaining reasonable predictions of reservoir parameters. Compared to deterministic methods, intelligent optimization methods based on nature-inspired metaheuristic algorithms have unique advantages because they do not require derivative information, can achieve global optimization, and have less reliance on initial model. Therefore, they perform better in solving complex nonlinear optimization problems. In this paper, a new intelligent optimization algorithm called Nutcracker Optimization Algorithm (NOA) with a high convergence speed is introduced. By utilizing this optimization algorithm to solve the nonlinear inversion problem constructed by the highly nonlinear exact Zoeppritz equations, we analyze the potential of nonlinear reservoir parameters prediction methods based on intelligent optimization algorithms in practical production applications. The synthetic data test shows that, compared to the classical quantum particle swarm optimization (QPSO) algorithm and the highly-cited whale optimization algorithm (WOA), the prestack nonlinear inversion method based on NOA proposed in this paper ensures high convergence accuracy and exhibits high computational efficiency. It significantly reduces computation time and holds great potential for practical production applications. The field data test shows that the proposed method can rapidly and accurately estimates reservoir parameters, validating the feasibility and effectiveness of the proposed method. This has important theoretical value and practical significance for advancing the application of intelligent optimization algorithms in field reservoir exploration and development.

The brittleness index is a crucial parameter for evaluating the brittleness of subsurface reservoirs. Accurate brittleness determination optimizes fracture design and guides oil and gas extraction, especially in shale formations. Traditionally, the brittleness index assumes isotropy, which fails to capture the anisotropic nature of shale reservoirs and often leads to prediction errors. To mitigate this challenge, this study introduces a stiffness coefficient matrix specifically designed for anisotropic media and proposes a brittleness index equation tailored for transverse isotropic (VTI) media. Experimental results show that the proposed anisotropic brittleness index provides a more accurate assessment of shale reservoir brittleness than the conventional isotropic brittleness index. Ultimately, the anisotropic brittleness index is applied to field logging data, thereby validating the effectiveness of the method in distinguishing between reservoirs of high and low brittleness.

When seismic waves propagate through horizontal transversely isotropic (HTI) media, the variation of acoustic impedance (AI) with azimuth can effectively reflect the fracture density and direction information. The key to high-precision fracture parameter prediction lies in improving the accuracy of azimuthal impedance inversion and elliptical analysis. This study proposes an integrated solution. First, the total variation (TV) multiplicative regularization multichannel impedance inversion technique is employed to obtain azimuthal impedance information. By incorporating formation dip constraints and an improved Polak-Ribi & egrave;re-Polyak and Hestenes-Stiefel (PRP-HS) hybrid conjugate gradient algorithm, the inversion accuracy and lateral resolution are significantly enhanced. Second, to address the small-sample problem in azimuthal impedance elliptical analysis, we developed a hybrid elliptical analysis method incorporating probability distribution and uncertainty estimation. The methodology involves: 1) applying the adaptive random sampling consensus (RANSAC) algorithm to screen high-quality data points; 2) utilizing Bayesian regression for probabilistic modeling of the data; and 3) implementing Bayesian inference through Markov Chain Monte Carlo (MCMC) methods. This combined strategy of random sampling and iterative optimization ensures the stability and reliability of fitting results. Furthermore, a planar transformation matrix is introduced to correct the rotation angle, thereby optimizing the prediction accuracy of fracture direction. Application results from actual well-logging and seismic data demonstrate that this method exhibits superior performance in predicting both fracture direction and density, providing a reliable technical approach for fractured reservoir characterization.

Cracks are a common rock microstructure and have a large effect on elastic properties during wave propagation. The fluid flow between a crack and its adjacent pore space can cause wave attenuation and dispersion. In this work, we introduce a crack connectivity parameter which is meant to improve the expression of local flow by weighting the contributions of fully connected and isolated cracks. We then update the analytical expression for frequency-dependent moduli by modifying the boundary conditions of the linearized Navier-Stokes equation and mass conservation equation. The proposed model contains the effect of cracks and stiff pores, in which the attenuation and dispersion are determined by squirt-flow and stiff-pore relaxations. The resulting model shows the squirt-flow relaxation frequency depends on not only the crack aspect ratio but also the crack connectivity. However, their contributions are different. The crack connectivity has little effect on the attenuation amplitude of shear modulus, but affects the attenuation amplitude of bulk modulus when multiple sets of cracks exist in the rock. The attenuation frequency band is also affected by the crack connectivity. As the crack connectivity deteriorates, the attenuation peak moves to low frequencies. In addition, by comparing the crack connectivity with the fluid viscosity coefficient, it is observed that the crack connectivity only affects the attenuation frequency band of cracks, whereas the fluid viscosity coefficient affects the attenuation frequency bands of cracks and stiff pores simultaneously. Thus, the introduction of crack connectivity is a supplement to the theoretical model of cracked fluid-saturated rocks. It helps understand the local fluid flow induced by seismic waves and provides a reasonable variation analysis of moduli and attenuation, especially for tight reservoirs.

Due to the lack of stiffness coefficient c12 information in seismic data of VTI (vertical axis symmetry transverse isotropy) media, it is still challenging to obtain the brittleness index of transversely isotropic (BIVTI) directly from seismic data. In this paper, we established the approximate relation c12 by using the rock physics model, and obtained BIVTI directly from seismic data based on the exact reflection coefficient equation. Firstly, we first derive an approximate expression for c12 through shale rock physics analysis. Secondly, we have proposed a method for obtaining BIVTI from seismic data using the Graebner exact reflection coefficient equation. This methoud considers the influence of horizontal slowness and deriving a more accurate Jacobian Matrix. Finally, we established the inversion method using model well tests, demonstrating the method’s superior precision and noise resistance. Compared with the traditional brittleness index, BIVTI obtained by the inversion method can better characterize the reservoir brittleness.

Pre-stack seismic inversion is an effective way to investigate the characteristics of hydrocarbon-bearing reservoirs. Multi-parameter application is the key to identifying reservoir lithology and fluid in pre-stack inversion. However, multi-parameter inversion may bring coupling effects on the parameters and destabilize the inversion. In addition, the lateral recognition accuracy of geological structures receives great attention. To address these challenges, a multi-task learning network considering the angle-gather difference is proposed in this work. The deep learning network is usually assumed as a black box and it is unclear what it can learn. However, the introduction of angle-gather difference can force the deep learning network to focus on the lateral differences, thus improving the lateral accuracy of the prediction profile. The proposed deep learning network includes input and output blocks. First, angle gathers and the angle-gather difference are fed into two separate input blocks with ResNet architecture and Unet architecture, respectively. Then, three elastic parameters, including P- and S-wave velocities and density, are simultaneously predicted based on the idea of multi-task learning by using three separate output blocks with the same convolutional network layers. Experimental and field data tests demonstrate the effectiveness of the proposed method in improving the prediction accuracy of seismic elastic parameters.