Wenlian Xiao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Wenlian Xiao has a total of 107 co-authors. Here are the top co-authors by number of shared publications:
Lingli Zheng
13 shared publicationsJitian Ren
6 shared publicationsJinzhou Zhao
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu, China
4 shared publicationsMin Li
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu, China
4 shared publicationsYubin Yang
Petro China Southwest Oil and Gas Field Branch, Chengdu, China
4 shared publications- Yves Bernabé
Department of Earth, Atmospheric and Planetary Sciences Massachusetts Institute of Technology Cambridge MA USA
4 shared publications
Latest Publications
- 1 November 2026
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- p. 1-20
Constrained Pressure Reconstruction and Stage-Specific Production-Decline Diagnosis for Shale Oil Horizontal Wells: A Chang 7 Field Application in the Ordos Basin
Summary: A key challenge in managing shale oil horizontal wells is the frequent absence of continuous shut-in pressure data, which impedes quantitative diagnosis of reservoir-energy depletion and production decline. Using 22 hydraulically fractured horizontal wells from the Chang 7 Member in the Ordos Basin as a field case, we first reconstruct the equivalent average reservoir pressure from routine production and flowing-pressure data using a coupled flowing material balance-dynamic material balance (FMB-DMB) procedure. The reconstructed pressure is used to calculate pressure drawdown and, together with stress-sensitivity data, to construct equivalent dimensionless flow capacity, whereas pressure/volume/temperature (PVT) data are used to reconstruct fluid mobility. Produced-fluid compositional data are used to interpret the associated fluid-property evolution. The stage-specific diagnostic contributions of pressure drawdown, fluid mobility, and equivalent dimensionless flow capacity to oil-rate decline are quantified using 30-day-lag logarithmic differences and Shapley decomposition. The FMB-DMB calculation satisfies the prescribed convergence criteria, and the reconstructed pressure trajectory remains within the physical pressure bounds and is consistent with pressure-buildup-derived reservoir pressures [root mean square error (RMSE): 0.823 MPa; mean absolute relative difference: 4.60%]. These pressure comparisons provide an external group-scale field-consistency check rather than independent calibration or evidence of a unique pressure solution. Results reveal that fluid-property evolution is accompanied by C11+ enrichment and a reduction in fluid mobility, while the equivalent dimensionless flow capacity decreases by 29.28% over the analysis period. Critically, the dominant diagnostic pattern shifts from joint positive contributions of flow capacity and fluid mobility in the early stage (69.13% and 57.20%, respectively, with pressure drawdown showing an offsetting contribution of −26.33%) to pressure drawdown as the dominant diagnostic factor in the middle and late stages (68.28% and 64.16%, respectively). For the representative Chang 7 well group analyzed in this study, these results support a stage-specific decline-management strategy: Early-stage management should limit rapid pressure depletion and preserve fracture and microfracture conductivity, whereas middle- and late-stage management should maintain effective drawdown while tracking fluid-mobility evolution.
- 1 November 2026
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- p. 1-20
Constrained Pressure Reconstruction and Stage-Specific Production-Decline Diagnosis for Shale Oil Horizontal Wells: A Chang 7 Field Application in the Ordos Basin
Summary: A key challenge in managing shale oil horizontal wells is the frequent absence of continuous shut-in pressure data, which impedes quantitative diagnosis of reservoir-energy depletion and production decline. Using 22 hydraulically fractured horizontal wells from the Chang 7 Member in the Ordos Basin as a field case, we first reconstruct the equivalent average reservoir pressure from routine production and flowing-pressure data using a coupled flowing material balance-dynamic material balance (FMB-DMB) procedure. The reconstructed pressure is used to calculate pressure drawdown and, together with stress-sensitivity data, to construct equivalent dimensionless flow capacity, whereas pressure/volume/temperature (PVT) data are used to reconstruct fluid mobility. Produced-fluid compositional data are used to interpret the associated fluid-property evolution. The stage-specific diagnostic contributions of pressure drawdown, fluid mobility, and equivalent dimensionless flow capacity to oil-rate decline are quantified using 30-day-lag logarithmic differences and Shapley decomposition. The FMB-DMB calculation satisfies the prescribed convergence criteria, and the reconstructed pressure trajectory remains within the physical pressure bounds and is consistent with pressure-buildup-derived reservoir pressures [root mean square error (RMSE): 0.823 MPa; mean absolute relative difference: 4.60%]. These pressure comparisons provide an external group-scale field-consistency check rather than independent calibration or evidence of a unique pressure solution. Results reveal that fluid-property evolution is accompanied by C11+ enrichment and a reduction in fluid mobility, while the equivalent dimensionless flow capacity decreases by 29.28% over the analysis period. Critically, the dominant diagnostic pattern shifts from joint positive contributions of flow capacity and fluid mobility in the early stage (69.13% and 57.20%, respectively, with pressure drawdown showing an offsetting contribution of −26.33%) to pressure drawdown as the dominant diagnostic factor in the middle and late stages (68.28% and 64.16%, respectively). For the representative Chang 7 well group analyzed in this study, these results support a stage-specific decline-management strategy: Early-stage management should limit rapid pressure depletion and preserve fracture and microfracture conductivity, whereas middle- and late-stage management should maintain effective drawdown while tracking fluid-mobility evolution.
- Open Access
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Geo-Attention PINN: A Physics-Constrained Hydrocarbon Well-Site Prospectivity Prediction Method with Darcy-Type Discrete Local Flux Residuals
Accurate prediction of hydrocarbon well-site prospectivity is essential for resource evaluation and development planning. However, purely data-driven deep learning methods, which rely on statistical correlations, often produce predictions with non-physical spatial fluctuations that violate subsurface flow principles and compromise geological plausibility. To address this issue, we propose a physics-constrained prediction method that integrates a discrete Darcy-type local flux residual. Multi-source geological data are transformed into multi-channel two-dimensional feature maps, and a channel-attention mechanism adaptively fuses these features to enhance the perception of key reservoir properties. Crucially, we reinterpret the predicted well-placement probability as a relative prospectivity potential field and construct a differentiable, transmissibility-weighted local flux residual using a five-point finite-difference stencil. This residual, calculated from interfacial transmissibilities between adjacent grid cells, serves as a physics-inspired regularization term that enforces spatial continuity consistent with reservoir connectivity. The model is trained with a joint loss function, balancing data fidelity and physical consistency. Experimental results demonstrate that our method achieves superior performance (coefficient of determination R^2=0.983) and, more importantly, produces spatially smoother and geologically more plausible prospectivity maps compared to purely data-driven baselines, as quantified by a significantly lower neighborhood gradient fluctuation index. These results indicate that the integration of Darcy-type discrete local flux residual constraints with deep learning provides an effective framework for hydrocarbon well-site prospectivity mapping and candidate well-site screening under complex heterogeneous reservoir conditions.
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessGeo-Attention PINN: A Physics-Constrained Hydrocarbon Well-Site Prospectivity Prediction Method with Darcy-Type Discrete Local Flux Residuals
Accurate prediction of hydrocarbon well-site prospectivity is essential for resource evaluation and development planning. However, purely data-driven deep learning methods, which rely on statistical correlations, often produce predictions with non-physical spatial fluctuations that violate subsurface flow principles and compromise geological plausibility. To address this issue, we propose a physics-constrained prediction method that integrates a discrete Darcy-type local flux residual. Multi-source geological data are transformed into multi-channel two-dimensional feature maps, and a channel-attention mechanism adaptively fuses these features to enhance the perception of key reservoir properties. Crucially, we reinterpret the predicted well-placement probability as a relative prospectivity potential field and construct a differentiable, transmissibility-weighted local flux residual using a five-point finite-difference stencil. This residual, calculated from interfacial transmissibilities between adjacent grid cells, serves as a physics-inspired regularization term that enforces spatial continuity consistent with reservoir connectivity. The model is trained with a joint loss function, balancing data fidelity and physical consistency. Experimental results demonstrate that our method achieves superior performance (coefficient of determination R^2=0.983) and, more importantly, produces spatially smoother and geologically more plausible prospectivity maps compared to purely data-driven baselines, as quantified by a significantly lower neighborhood gradient fluctuation index. These results indicate that the integration of Darcy-type discrete local flux residual constraints with deep learning provides an effective framework for hydrocarbon well-site prospectivity mapping and candidate well-site screening under complex heterogeneous reservoir conditions.
- 24 August 2026
- Published by Elsevier in Geoenergy Science and Engineering
- Vol. 267
Dynamic visualization and simulation of CO2 miscible-front instability in low-permeability heterogeneous reservoirs
- 24 August 2026
- Published by Elsevier in Geoenergy Science and Engineering
- Vol. 267
Dynamic visualization and simulation of CO2 miscible-front instability in low-permeability heterogeneous reservoirs
- Open Access
- 18 August 2026
- Published by Elsevier in Results in Engineering
- Vol. 32
Impact of pore-throat heterogeneity and pre-Darcy flow on two-phase flow in ultra-low permeability sandstones
- 18 August 2026
- Published by Elsevier in Results in Engineering
- Vol. 32
Open AccessImpact of pore-throat heterogeneity and pre-Darcy flow on two-phase flow in ultra-low permeability sandstones
- Open Access
- 8 July 2026
- Published by MDPI AG in Preprints.org
Stage-Dependent Production Behavior and Adaptive Strategy Optimization for Horizontal Wells in Shale Oil Reservoirs, Longdong Area, NW China
Shale oil reservoirs are inherently tight and are currently developed mainly by natural energy depletion through horizontal wells with volumetric fracturing. However, the production characteristics during the post-fracturing process remain unclear, and the contributions of various energy sources have not been quantified, leading to a lack of targeted measures for maintaining stable production in horizontal wells. In this study, we combined microscopic visualization experiments with nuclear magnetic resonance (NMR) to characterize oil production in shale oil reservoirs across different development stages and to determine the contribution of each energy source to the overall recovery factor. Using reservoir engineering methods, we further evaluated the contributions of various energy sources in Xi-233 and Z-183 blocks. The results show that during the post-fracture shut-in (PFSI) stage, imbibition mobilizes oil from pores of all sizes, whereas the subsequent displacement phase primarily targets oil in meso and macropores. The PFSI not only enhances oil recovery factor but also improves oil-water flow during later production. The total recovery factor contributed by the shut-in phase, elastic energy of the fracturing fluid, elastic energy of the reservoir fluid, and solution gas drive is approximately 10%, with individual contributions of 7.40%, 6.67%, 66.74%, and 19.19%, respectively. Among these, the elastic energy of the reservoir fluid is the dominant contributor to the total recovery factor. These findings provide a theoretical foundation for optimizing production strategies in shale oil wells, support the efficient development of shale oil in Longdong area, and offer valuable insights for the development of shale oil resources across the world.
- 8 July 2026
- Published by MDPI AG in Preprints.org
Open AccessStage-Dependent Production Behavior and Adaptive Strategy Optimization for Horizontal Wells in Shale Oil Reservoirs, Longdong Area, NW China
Shale oil reservoirs are inherently tight and are currently developed mainly by natural energy depletion through horizontal wells with volumetric fracturing. However, the production characteristics during the post-fracturing process remain unclear, and the contributions of various energy sources have not been quantified, leading to a lack of targeted measures for maintaining stable production in horizontal wells. In this study, we combined microscopic visualization experiments with nuclear magnetic resonance (NMR) to characterize oil production in shale oil reservoirs across different development stages and to determine the contribution of each energy source to the overall recovery factor. Using reservoir engineering methods, we further evaluated the contributions of various energy sources in Xi-233 and Z-183 blocks. The results show that during the post-fracture shut-in (PFSI) stage, imbibition mobilizes oil from pores of all sizes, whereas the subsequent displacement phase primarily targets oil in meso and macropores. The PFSI not only enhances oil recovery factor but also improves oil-water flow during later production. The total recovery factor contributed by the shut-in phase, elastic energy of the fracturing fluid, elastic energy of the reservoir fluid, and solution gas drive is approximately 10%, with individual contributions of 7.40%, 6.67%, 66.74%, and 19.19%, respectively. Among these, the elastic energy of the reservoir fluid is the dominant contributor to the total recovery factor. These findings provide a theoretical foundation for optimizing production strategies in shale oil wells, support the efficient development of shale oil in Longdong area, and offer valuable insights for the development of shale oil resources across the world.
- Open Access
- 1 June 2026
- Published by KEAI Publishing in Energy Geoscience
- Vol. 7 (3)
Full scale pore throat characterization incorporating a dynamic shape factor and permeability prediction in tight rocks
- 1 June 2026
- Published by KEAI Publishing in Energy Geoscience
- Vol. 7 (3)
Open AccessFull scale pore throat characterization incorporating a dynamic shape factor and permeability prediction in tight rocks
- 1 May 2026
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- Vol. 31 (08) , 1-19
Automated Algorithms and Workflow for Production Data Analysis of Hydraulically Fractured Wells with Field-Case Validations
Summary: Production data analysis for hydraulically fractured wells is critical for optimizing unconventional reservoirs and choke management. However, traditional analysis is often subjective and labor-intensive and fails to handle the high noise and operational complexity of field data, hindering accurate parameter estimation. In this study, we present a comprehensive, fully automated workflow designed to transition this analysis from a manual art to a data-driven science. The core innovation is a holistic, intelligent pipeline built on a data-driven, self-calibrating architecture that eliminates manual parameter tuning. This system integrates physics-based bottomhole pressure (BHP) reconstruction, adaptive denoising, and automated segmentation. It then applies the unified transient analysis (UTA) framework, coupled with a robust regression and multigate validation module, to objectively quantify the dynamic effective fracture surface area (Af). The workflow’s robustness was demonstrated through application to three distinct and challenging field cases: a gas well with a complex operational history, an oil well with extremely noisy surface data, and a long-term oil production well exhibiting stress-dependent fracture degradation. This work provides the essential “monitoring-inversion-analysis” engine for a paradigm shift from static management to dynamic, automated single-well production optimization, enabling real-time, data-driven optimization and maximizing ultimate recovery.
- 1 May 2026
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- Vol. 31 (08) , 1-19
Automated Algorithms and Workflow for Production Data Analysis of Hydraulically Fractured Wells with Field-Case Validations
Summary: Production data analysis for hydraulically fractured wells is critical for optimizing unconventional reservoirs and choke management. However, traditional analysis is often subjective and labor-intensive and fails to handle the high noise and operational complexity of field data, hindering accurate parameter estimation. In this study, we present a comprehensive, fully automated workflow designed to transition this analysis from a manual art to a data-driven science. The core innovation is a holistic, intelligent pipeline built on a data-driven, self-calibrating architecture that eliminates manual parameter tuning. This system integrates physics-based bottomhole pressure (BHP) reconstruction, adaptive denoising, and automated segmentation. It then applies the unified transient analysis (UTA) framework, coupled with a robust regression and multigate validation module, to objectively quantify the dynamic effective fracture surface area (Af). The workflow’s robustness was demonstrated through application to three distinct and challenging field cases: a gas well with a complex operational history, an oil well with extremely noisy surface data, and a long-term oil production well exhibiting stress-dependent fracture degradation. This work provides the essential “monitoring-inversion-analysis” engine for a paradigm shift from static management to dynamic, automated single-well production optimization, enabling real-time, data-driven optimization and maximizing ultimate recovery.