Unconventional Resources最新文献_第5页

A hybrid numerical-analytical approach for solar cells and modules parameter extraction using the Regula-Falsi method 太阳能电池和组件参数提取的正则-法尔西混合数值分析方法

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-10-22 DOI: 10.1016/j.uncres.2025.100266

Fatima Ezzahrae Souaidi, Mustapha Elyaqouti, El Hanafi Arjdal, Driss Saadaoui, Imade Choulli, Abdelfattah Elhammoudy, Brahim Ydir, Ismail Abazine, Souad Lidaighbi, Dris Ben Hmamou, Chahir Omar, Lahboub Ayoub, Bendriouich Youssouf

The growing world interest in clean and sustainable energy has tremendously driven research on photovoltaic systems, with a strong emphasis on the need for accurate modeling and parameter extraction. Accurate PV modeling is vital for optimizing, simulating, and generally enhancing PV device performance. As much as there has been great development, available techniques have commonly faced limitations in finding an adequate balance between analytical accuracy and numerical reliability. In order to overcome these limitations, this paper proposes an innovative hybrid technique that integrates analytical solutions with numerical iterative strategies to accurately extract parameters of the single diode model using the Regula Falsi method. The new approach provides high accuracy in simulating photovoltaic modules by minimizing the Root Mean Square Error (RMSE) between experimental and simulated current-voltage data. It achieves an RMSE value of 7.746E-04 A, which is 1.35 % better than the best alternative approach for RTC France. For Photowatt-PWP 201, it achieves 2.148E-03 A with a 0.12 % improvement. It maintains comparable accuracy for PVM 752 GaAs, with an RMSE of 2.912E-04A, and reaches 1.731E-03A with an increase of 9.02 % for the STM6-40/36 module. Comparative analysis with state-of-the-art techniques highlights the superior efficiency and reliability of the proposed method.

世界对清洁和可持续能源日益增长的兴趣极大地推动了光伏系统的研究，强调了对准确建模和参数提取的需要。准确的PV建模对于优化、模拟和普遍提高PV设备性能至关重要。尽管已经有了很大的发展，但现有的技术在寻找分析精度和数值可靠性之间的适当平衡方面通常面临限制。为了克服这些局限性，本文提出了一种创新的混合技术，将解析解与数值迭代策略相结合，利用正则法精确提取单二极管模型的参数。新方法通过最小化实验和模拟电流电压数据之间的均方根误差（RMSE），提供了较高的光伏组件模拟精度。它的RMSE值为7.7460 e -04 A，比RTC France的最佳替代方法好1.35%。对于photowatt - pwp201，它实现了2.148E-03 A，提高了0.12%。对于PVM 752 GaAs，它保持了相当的精度，RMSE为2.912E-04A，对于STM6-40/36模块，RMSE达到1.73e - 03a，增加了9.02%。与最先进的技术进行对比分析，突出了该方法的效率和可靠性。

{"title":"A hybrid numerical-analytical approach for solar cells and modules parameter extraction using the Regula-Falsi method","authors":"Fatima Ezzahrae Souaidi, Mustapha Elyaqouti, El Hanafi Arjdal, Driss Saadaoui, Imade Choulli, Abdelfattah Elhammoudy, Brahim Ydir, Ismail Abazine, Souad Lidaighbi, Dris Ben Hmamou, Chahir Omar, Lahboub Ayoub, Bendriouich Youssouf","doi":"10.1016/j.uncres.2025.100266","DOIUrl":"10.1016/j.uncres.2025.100266","url":null,"abstract":"<div><div>The growing world interest in clean and sustainable energy has tremendously driven research on photovoltaic systems, with a strong emphasis on the need for accurate modeling and parameter extraction. Accurate PV modeling is vital for optimizing, simulating, and generally enhancing PV device performance. As much as there has been great development, available techniques have commonly faced limitations in finding an adequate balance between analytical accuracy and numerical reliability. In order to overcome these limitations, this paper proposes an innovative hybrid technique that integrates analytical solutions with numerical iterative strategies to accurately extract parameters of the single diode model using the Regula Falsi method. The new approach provides high accuracy in simulating photovoltaic modules by minimizing the Root Mean Square Error (RMSE) between experimental and simulated current-voltage data. It achieves an RMSE value of 7.746E-04 A, which is 1.35 % better than the best alternative approach for RTC France. For Photowatt-PWP 201, it achieves 2.148E-03 A with a 0.12 % improvement. It maintains comparable accuracy for PVM 752 GaAs, with an RMSE of 2.912E-04A, and reaches 1.731E-03A with an increase of 9.02 % for the STM6-40/36 module. Comparative analysis with state-of-the-art techniques highlights the superior efficiency and reliability of the proposed method.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100266"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145365589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Thank you reviewers! 谢谢审稿人！

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2026-01-06 DOI: 10.1016/j.uncres.2026.100306

引用次数: 0

Quantifying mechanical stress effects on mono PERC and polycrystalline PV modules with EL imaging and statistical analysis 用光电成像和统计分析量化单PERC和多晶光伏组件的机械应力效应

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-11-07 DOI: 10.1016/j.uncres.2025.100276

Sumit Verma , D.K. Yadav , Birinchi Bora , Shiv Lal

Solar photovoltaic modules convert solar energy into electrical energy, but defects like cell cracks, cell breakage, inactive cell area, and other defects impact the long-term performance and reliability of photovoltaic modules. Therefore, developing effective techniques for identifying such defects is crucial for optimising photovoltaic power plants efficiency. Using electroluminescence imaging, a method is proposed to quantify defects in mono PERC and polycrystalline solar photovoltaic modules.

Various statistical measures, such as mean intensity, variance, kurtosis, and skewness of the electroluminescence intensity histogram, were compared. Healthy mono PERC modules exhibit higher values (for mean intensity: 0.530, variance: 0.867, kurtosis: 1.352, and skewness: 1.697) than defective modules (mean intensity: 0.441, variance: 0.396, kurtosis: 0.797, skewness: 1.466), indicating fewer defects with better performance. The polycrystalline photovoltaic module showed higher degradation with P_max and I_mp reductions of 15.13 % and 14.09 %, respectively, during static mechanical load testing. Results also show that modules with fewer defects exhibit better long-term performance, durability, and efficiency, particularly for mono PERC photovoltaic modules, which outperform polycrystalline modules in degradation analysis. Analysing electroluminescence images enables researchers and industry professionals to identify reliability issues in photovoltaic modules and to optimise the performance of solar photovoltaic power plants.

太阳能光伏组件将太阳能转化为电能，但电池裂纹、电池断裂、电池闲置面积等缺陷影响了光伏组件的长期性能和可靠性。因此，开发识别此类缺陷的有效技术对于优化光伏电站效率至关重要。利用电致发光成像技术，提出了一种对单PERC和多晶太阳能光伏组件缺陷进行量化的方法。比较了电致发光强度直方图的平均强度、方差、峰度和偏度等统计指标。健康的单PERC模块表现出更高的值（平均强度：0.530，方差：0.867，峰度：1.352，偏度：1.697）比缺陷模块（平均强度：0.441，方差：0.396，峰度：0.797，偏度：1.466），表明缺陷更少，性能更好。在静态机械负载测试中，多晶光伏组件表现出更高的退化，Pmax和Imp分别降低了15.13%和14.09%。结果还表明，缺陷较少的组件具有更好的长期性能，耐用性和效率，特别是单PERC光伏组件，在降解分析中优于多晶组件。分析电致发光图像使研究人员和行业专业人员能够识别光伏组件的可靠性问题，并优化太阳能光伏发电厂的性能。

{"title":"Quantifying mechanical stress effects on mono PERC and polycrystalline PV modules with EL imaging and statistical analysis","authors":"Sumit Verma , D.K. Yadav , Birinchi Bora , Shiv Lal","doi":"10.1016/j.uncres.2025.100276","DOIUrl":"10.1016/j.uncres.2025.100276","url":null,"abstract":"<div><div>Solar photovoltaic modules convert solar energy into electrical energy, but defects like cell cracks, cell breakage, inactive cell area, and other defects impact the long-term performance and reliability of photovoltaic modules. Therefore, developing effective techniques for identifying such defects is crucial for optimising photovoltaic power plants efficiency. Using electroluminescence imaging, a method is proposed to quantify defects in mono PERC and polycrystalline solar photovoltaic modules.</div><div>Various statistical measures, such as mean intensity, variance, kurtosis, and skewness of the electroluminescence intensity histogram, were compared. Healthy mono PERC modules exhibit higher values (for mean intensity: 0.530, variance: 0.867, kurtosis: 1.352, and skewness: 1.697) than defective modules (mean intensity: 0.441, variance: 0.396, kurtosis: 0.797, skewness: 1.466), indicating fewer defects with better performance. The polycrystalline photovoltaic module showed higher degradation with P<sub>max</sub> and I<sub>mp</sub> reductions of 15.13 % and 14.09 %, respectively, during static mechanical load testing. Results also show that modules with fewer defects exhibit better long-term performance, durability, and efficiency, particularly for mono PERC photovoltaic modules, which outperform polycrystalline modules in degradation analysis. Analysing electroluminescence images enables researchers and industry professionals to identify reliability issues in photovoltaic modules and to optimise the performance of solar photovoltaic power plants.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100276"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145571416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Machine learning method for lacustrine shale oil reservoirs: Improving movable fluid porosity prediction 湖相页岩油藏机器学习方法：改进可动流体孔隙度预测

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-10-17 DOI: 10.1016/j.uncres.2025.100263

Xinyu Liu , Sandong Zhou , Weixin Zhang , Qiaoyun Cheng , Dameng Liu , Detian Yan , Hua Wang

Movable fluid porosity (Φ_mf) is a key parameter to evaluate the seepage capacity and movable oil volume of shale reservoirs. Accurate measurement of the Φ_mf relies on expensive and time-consuming experimental methods, and little work has been done on intelligent prediction. In this study, the Upper Cretaceous Qingshankou Formation shale in the Changling Depression of the Songliao Basin is selected as the research object. The 635 data points are divided into three datasets: training set, validation set, and testing set, allocated in a ratio of 0.7:0.15:0.15. Using five logging parameters including Gamma ray log (GR), Deep lateral resistivity log (RLLD), Acoustic log (AC), Density log (DEN), and Neutron log (CNL), four machine learning models are selected to predict Φ_mf, including Decision tree (DT), Random forest (RF), Gradient boosting decision tree (GBDT), and Artificial neural network (ANN). Multiple evaluation indicators are adopted to compare the accuracy and applicability of different algorithms. The results show the performance of different algorithms in predicting Φ_mf, from highest to lowest, as follows: GBDT > RF > DT > ANN. The sensitivity analysis based on the Shapley Additive exPlanations (SHAP) indicates that AC has the greatest positive impact on predicting the Φ_mf. The applicability analysis shows that compared with the traditional Multiple regression analysis (MRA), the use of machine learning algorithms can effectively improve the prediction accuracy, with a maximum increase of 24 %. This study holds that the GBDT can predict the Φ_mf of shale reservoirs efficiently and accurately, providing valuable insights for the global evaluation and development of lacustrine shale oil resources.

可动流体孔隙度（Φmf）是评价页岩储层渗流能力和可动油体积的关键参数。Φmf的精确测量依赖于昂贵且耗时的实验方法，而在智能预测方面的工作很少。本研究以松辽盆地长岭坳陷上白垩统青山口组页岩为研究对象。635个数据点分为三个数据集：训练集、验证集和测试集，按0.7:0.15:0.15的比例分配。利用伽马测井（GR）、深部电阻率测井（RLLD）、声波测井（AC）、密度测井（DEN）和中子测井（CNL） 5种测井参数，选择决策树（DT）、随机森林（RF）、梯度增强决策树（GBDT）和人工神经网络（ANN） 4种机器学习模型进行Φmf预测。采用多个评价指标比较不同算法的准确性和适用性。结果表明，不同算法对Φmf的预测性能从高到低依次为：GBDT >； RF > DT >； ANN。基于Shapley加性解释（SHAP）的敏感性分析表明，AC对Φmf的预测有最大的正向影响。适用性分析表明，与传统的多元回归分析（MRA）相比，使用机器学习算法可以有效提高预测精度，最大提高24%。研究认为，GBDT可以高效、准确地预测页岩储层Φmf，为全球湖相页岩油资源评价与开发提供有价值的见解。

{"title":"Machine learning method for lacustrine shale oil reservoirs: Improving movable fluid porosity prediction","authors":"Xinyu Liu , Sandong Zhou , Weixin Zhang , Qiaoyun Cheng , Dameng Liu , Detian Yan , Hua Wang","doi":"10.1016/j.uncres.2025.100263","DOIUrl":"10.1016/j.uncres.2025.100263","url":null,"abstract":"<div><div>Movable fluid porosity (Φ<sub>mf</sub>) is a key parameter to evaluate the seepage capacity and movable oil volume of shale reservoirs. Accurate measurement of the Φ<sub>mf</sub> relies on expensive and time-consuming experimental methods, and little work has been done on intelligent prediction. In this study, the Upper Cretaceous Qingshankou Formation shale in the Changling Depression of the Songliao Basin is selected as the research object. The 635 data points are divided into three datasets: training set, validation set, and testing set, allocated in a ratio of 0.7:0.15:0.15. Using five logging parameters including Gamma ray log (GR), Deep lateral resistivity log (RLLD), Acoustic log (AC), Density log (DEN), and Neutron log (CNL), four machine learning models are selected to predict Φ<sub>mf</sub>, including Decision tree (DT), Random forest (RF), Gradient boosting decision tree (GBDT), and Artificial neural network (ANN). Multiple evaluation indicators are adopted to compare the accuracy and applicability of different algorithms. The results show the performance of different algorithms in predicting Φ<sub>mf</sub>, from highest to lowest, as follows: GBDT > RF > DT > ANN. The sensitivity analysis based on the Shapley Additive exPlanations (SHAP) indicates that AC has the greatest positive impact on predicting the Φ<sub>mf</sub>. The applicability analysis shows that compared with the traditional Multiple regression analysis (MRA), the use of machine learning algorithms can effectively improve the prediction accuracy, with a maximum increase of 24 %. This study holds that the GBDT can predict the Φ<sub>mf</sub> of shale reservoirs efficiently and accurately, providing valuable insights for the global evaluation and development of lacustrine shale oil resources.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100263"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145365442","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Comparative impact of cold water and thermochemical cooling methods on breakdown pressure for improved stimulation in unconventional formations 冷水和热化学冷却方法对非常规地层破裂压力的影响比较，以提高增产效果

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2026-01-08 DOI: 10.1016/j.uncres.2026.100309

Fahad Khan , Arshad Raza , Mohamed Mahmoud , Murtadha J. AlTammar , Talal Al Shafloot

High breakdown pressure poses a significant challenges in terms of pumping pressure and associated costs during stimulation operations in unconventional reservoirs. These reservoirs are characterized with low porosity, low permeability, and high in-situ temperature due to greater depths. To address this challenge, the current study investigates the effectiveness of various cooling strategies and their comparative impact on the breakdown pressure and strength of unconventional rocks (Kentucky sandstone and Eagleford shale). These rocks were heated up to 150 °C and cooled down by following three different strategies: ⅰ) spontaneous cooling i.e. without any external aid ⅱ) cooling with cold water and ⅲ) cooling with endothermic chemicals involving NH₄Cl and NaOH. With endothermic cooling, the temperature of Kentucky sandstone decreases from 150 °C to 56.7 °C in 14 s, while Eagleford shale cools from 150 °C to 43.8 °C in 10 s. The endothermic cooling was followed by the Cold water and spontaneous cooling which showed slower and less pronounced temperature drops in both rocks as compared to the endothermic cooling. The endothermic cooling also leads to highest reduction in rock strength and breakdown pressure. The strength shows a reduction of 21.9 % in Kentucky sandstone and 25.4 % in Eagleford shale while the breakdown pressure reduces by 38.6 % and 37.3 % for the Kentucky sandstone and Eagleford shale respectively. The study also shows the structural changes in the rocks, particularly rock morphology and pore volume. FIB-SEM analysis shows the development of multiple micro-cracks in the rocks which plays an important role in reducing the breakdown pressure. The outcomes of this study indicate that pre-fracturing cooling treatment using endothermic fluids can enhance the effectiveness of hydraulic fracturing operations by reducing the formation breakdown pressure.

在非常规油藏增产作业中，高破裂压力给泵送压力和相关成本带来了重大挑战。这些储层具有低孔隙度、低渗透率、深层温度高的特点。为了应对这一挑战，目前的研究调查了各种冷却策略的有效性，以及它们对非常规岩石（肯塔基砂岩和Eagleford页岩）破裂压力和强度的比较影响。将这些岩石加热至150℃后，采用三种不同的冷却策略进行冷却：ⅰ)自然冷却，即没有任何外部帮助；ⅱ)冷水冷却；ⅲ)含NH4Cl和NaOH的吸热化学物质冷却。通过吸热冷却，Kentucky砂岩的温度在14 s内从150°C降至56.7°C， Eagleford页岩在10 s内从150°C降至43.8°C。吸热冷却之后是冷水和自发冷却，与吸热冷却相比，这两种岩石的温度下降速度更慢，也更不明显。吸热冷却还能最大程度地降低岩石强度和破裂压力。肯塔基砂岩强度降低21.9%，Eagleford页岩强度降低25.4%，破裂压力降低38.6%，Eagleford页岩强度降低37.3%。研究还揭示了岩石的结构变化，特别是岩石形态和孔隙体积的变化。FIB-SEM分析表明，岩石中存在多个微裂纹，这对降低破裂压力起着重要作用。研究结果表明，采用吸热流体进行压裂前冷却处理可以通过降低地层破裂压力来提高水力压裂作业的有效性。

{"title":"Comparative impact of cold water and thermochemical cooling methods on breakdown pressure for improved stimulation in unconventional formations","authors":"Fahad Khan , Arshad Raza , Mohamed Mahmoud , Murtadha J. AlTammar , Talal Al Shafloot","doi":"10.1016/j.uncres.2026.100309","DOIUrl":"10.1016/j.uncres.2026.100309","url":null,"abstract":"<div><div>High breakdown pressure poses a significant challenges in terms of pumping pressure and associated costs during stimulation operations in unconventional reservoirs. These reservoirs are characterized with low porosity, low permeability, and high in-situ temperature due to greater depths. To address this challenge, the current study investigates the effectiveness of various cooling strategies and their comparative impact on the breakdown pressure and strength of unconventional rocks (Kentucky sandstone and Eagleford shale). These rocks were heated up to 150 °C and cooled down by following three different strategies: ⅰ) spontaneous cooling i.e. without any external aid ⅱ) cooling with cold water and ⅲ) cooling with endothermic chemicals involving NH<sub>4</sub>Cl and NaOH. With endothermic cooling, the temperature of Kentucky sandstone decreases from 150 °C to 56.7 °C in 14 s, while Eagleford shale cools from 150 °C to 43.8 °C in 10 s. The endothermic cooling was followed by the Cold water and spontaneous cooling which showed slower and less pronounced temperature drops in both rocks as compared to the endothermic cooling. The endothermic cooling also leads to highest reduction in rock strength and breakdown pressure. The strength shows a reduction of 21.9 % in Kentucky sandstone and 25.4 % in Eagleford shale while the breakdown pressure reduces by 38.6 % and 37.3 % for the Kentucky sandstone and Eagleford shale respectively. The study also shows the structural changes in the rocks, particularly rock morphology and pore volume. FIB-SEM analysis shows the development of multiple micro-cracks in the rocks which plays an important role in reducing the breakdown pressure. The outcomes of this study indicate that pre-fracturing cooling treatment using endothermic fluids can enhance the effectiveness of hydraulic fracturing operations by reducing the formation breakdown pressure.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100309"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924085","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Electric vehicles with renewables integration in electrical power systems: A review of technologies, uncertainties and optimization allocations 在电力系统中集成可再生能源的电动汽车：技术、不确定性和优化分配的回顾

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-10-09 DOI: 10.1016/j.uncres.2025.100252

Abdullah M. Shaheen , Aya R. Ellien , Adel A. El-Ela , Ali M. El-Rifaie

-Electric vehicles (EVs) are becoming a key part of future transportation and energy systems as the world moves towards decarbonization and sustainable mobility. This review explores the evolution, classification, and technical architecture of EVs, emphasizing their integration within renewable energy-powered electrical networks. It examines the main EV technologies, such as battery systems, drivetrains, and charging infrastructure, highlighting recent advancements such as fast-charging capabilities and bidirectional energy flows (e.g., V2G, V2B). A critical analysis of energy storage technologies and battery management systems (BMS) is presented, addressing their influence on vehicle performance and grid interaction. To deal with the unpredictability that comes with EVs and renewables, the study examines many ways to simulate uncertainty, such as Monte Carlo simulation, Markov chains, copula functions, point estimate methods, and neural networks. These tools are essential for forecasting load demand, charging behavior, and battery performance in dynamic grid conditions. Furthermore, the paper surveys recent optimization frameworks developed for planning and operation of EV infrastructure, focusing on objectives such as loss minimization, voltage profile improvement, cost reduction, and environmental impact mitigation. This study observes common research gaps by considering a number of different studies, including limited treatment of unbalanced distribution networks, insufficient real-time control strategies, and the underutilization of advanced optimization methods for large-scale deployment. The study reveals that EVs can enhance electrical systems by integrating with renewable energy sources, and suggests future research to overcome technical hurdles and expedite their adoption in modern power grids.

随着世界向脱碳和可持续移动的方向发展，电动汽车（ev）正在成为未来交通和能源系统的关键组成部分。本文探讨了电动汽车的发展、分类和技术架构，强调了它们在可再生能源供电网络中的集成。报告介绍了主要的电动汽车技术，如电池系统、动力传动系统和充电基础设施，重点介绍了快速充电能力和双向能量流（如V2G、V2B）等最新进展。对储能技术和电池管理系统（BMS）进行了批判性分析，解决了它们对车辆性能和电网交互的影响。为了应对电动汽车和可再生能源带来的不可预测性，该研究研究了许多模拟不确定性的方法，如蒙特卡罗模拟、马尔可夫链、copula函数、点估计方法和神经网络。这些工具对于预测动态电网条件下的负载需求、充电行为和电池性能至关重要。此外，本文还调查了最近为电动汽车基础设施规划和运营开发的优化框架，重点关注损耗最小化、电压分布改善、成本降低和环境影响缓解等目标。本研究通过考虑一些不同的研究，包括对不平衡配电网络的有限处理，实时控制策略的不足，以及大规模部署的先进优化方法的利用不足，观察到常见的研究差距。该研究表明，电动汽车可以通过与可再生能源相结合来增强电力系统，并建议未来的研究克服技术障碍，加快其在现代电网中的应用。

{"title":"Electric vehicles with renewables integration in electrical power systems: A review of technologies, uncertainties and optimization allocations","authors":"Abdullah M. Shaheen , Aya R. Ellien , Adel A. El-Ela , Ali M. El-Rifaie","doi":"10.1016/j.uncres.2025.100252","DOIUrl":"10.1016/j.uncres.2025.100252","url":null,"abstract":"<div><div>-Electric vehicles (EVs) are becoming a key part of future transportation and energy systems as the world moves towards decarbonization and sustainable mobility. This review explores the evolution, classification, and technical architecture of EVs, emphasizing their integration within renewable energy-powered electrical networks. It examines the main EV technologies, such as battery systems, drivetrains, and charging infrastructure, highlighting recent advancements such as fast-charging capabilities and bidirectional energy flows (e.g., V2G, V2B). A critical analysis of energy storage technologies and battery management systems (BMS) is presented, addressing their influence on vehicle performance and grid interaction. To deal with the unpredictability that comes with EVs and renewables, the study examines many ways to simulate uncertainty, such as Monte Carlo simulation, Markov chains, copula functions, point estimate methods, and neural networks. These tools are essential for forecasting load demand, charging behavior, and battery performance in dynamic grid conditions. Furthermore, the paper surveys recent optimization frameworks developed for planning and operation of EV infrastructure, focusing on objectives such as loss minimization, voltage profile improvement, cost reduction, and environmental impact mitigation. This study observes common research gaps by considering a number of different studies, including limited treatment of unbalanced distribution networks, insufficient real-time control strategies, and the underutilization of advanced optimization methods for large-scale deployment. The study reveals that EVs can enhance electrical systems by integrating with renewable energy sources, and suggests future research to overcome technical hurdles and expedite their adoption in modern power grids.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100252"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145327421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Feasibility of PEM electrolysis using seawater for on-site hydrogen production at the port of New York and New Jersey 在纽约和新泽西港利用海水电解PEM现场制氢的可行性

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-12-10 DOI: 10.1016/j.uncres.2025.100292

Eliseo Curcio

This study evaluates the real-world feasibility of deploying on-site hydrogen production at the Port of New York and New Jersey using seawater-fed proton exchange membrane (PEM) electrolysis under hard infrastructure constraints. Unlike greenfield assessments that assume unconstrained access to land, electricity, or freshwater, this analysis adopts a constraint-based systems model. It incorporates validated electrolyzer performance metrics, land footprint requirements, seawater treatment demands, interconnection limits, and actual port fuel consumption to simulate realistic deployment scenarios. The model identifies configurations capable of displacing 10–40 % of the port's current diesel and natural gas use. A system designed to meet 20 % of fuel demand (1000 kg/day of hydrogen) requires 60 MW of PEM capacity, 6000 m² of land, and 120 m³/day of treated seawater, while avoiding over 3700 tonnes of CO₂ annually when benchmarked against GREET diesel emissions. Applying the U.S. §45V clean hydrogen tax credit reduces system costs by more than 60 %, bringing the effective hydrogen price near diesel parity for yard tractors and cargo-handling equipment (CHE). Spatial screening confirms that such systems can be integrated at three terminal sites without disrupting core port operations. These findings validate that phased hydrogen deployment starting at 500–1000 kg/day is technically, economically, and spatially feasible using only existing port-controlled infrastructure. As such, hydrogen is positioned not as a long-term contingency but as an immediate and actionable decarbonization solution for high-throughput, urbanized ports.

本研究评估了在硬基础设施限制下，在纽约港和新泽西港使用海水质子交换膜（PEM）电解进行现场制氢的可行性。与假定不受限制地获取土地、电力或淡水的绿地评估不同，该分析采用了基于约束的系统模型。它结合了经过验证的电解槽性能指标、土地足迹要求、海水处理要求、互连限制和实际港口燃料消耗，以模拟现实的部署场景。该模型确定了能够取代港口目前柴油和天然气使用量10 - 40%的配置。一个旨在满足20%燃料需求（1000千克/天氢气）的系统需要60兆瓦的PEM容量，6000平方米的土地和120立方米/天的处理海水，同时以GREET柴油排放为基准，每年避免超过3700吨的二氧化碳。采用美国§45V清洁氢税收抵免可使系统成本降低60%以上，使氢的有效价格接近庭院拖拉机和货物装卸设备（CHE）的柴油价格。空间筛选证实，这些系统可以在三个码头站点集成，而不会中断核心港口业务。这些发现证实，仅使用现有的港口控制基础设施，从500-1000公斤/天开始分阶段部署氢气在技术上、经济上和空间上都是可行的。因此，氢不是作为一种长期的应急方案，而是作为高吞吐量、城市化港口的一种即时可行的脱碳解决方案。

{"title":"Feasibility of PEM electrolysis using seawater for on-site hydrogen production at the port of New York and New Jersey","authors":"Eliseo Curcio","doi":"10.1016/j.uncres.2025.100292","DOIUrl":"10.1016/j.uncres.2025.100292","url":null,"abstract":"<div><div>This study evaluates the real-world feasibility of deploying on-site hydrogen production at the Port of New York and New Jersey using seawater-fed proton exchange membrane (PEM) electrolysis under hard infrastructure constraints. Unlike greenfield assessments that assume unconstrained access to land, electricity, or freshwater, this analysis adopts a constraint-based systems model. It incorporates validated electrolyzer performance metrics, land footprint requirements, seawater treatment demands, interconnection limits, and actual port fuel consumption to simulate realistic deployment scenarios. The model identifies configurations capable of displacing 10–40 % of the port's current diesel and natural gas use. A system designed to meet 20 % of fuel demand (1000 kg/day of hydrogen) requires 60 MW of PEM capacity, 6000 m<sup>2</sup> of land, and 120 m<sup>3</sup>/day of treated seawater, while avoiding over 3700 tonnes of CO<sub>2</sub> annually when benchmarked against GREET diesel emissions. Applying the U.S. §45V clean hydrogen tax credit reduces system costs by more than 60 %, bringing the effective hydrogen price near diesel parity for yard tractors and cargo-handling equipment (CHE). Spatial screening confirms that such systems can be integrated at three terminal sites without disrupting core port operations. These findings validate that phased hydrogen deployment starting at 500–1000 kg/day is technically, economically, and spatially feasible using only existing port-controlled infrastructure. As such, hydrogen is positioned not as a long-term contingency but as an immediate and actionable decarbonization solution for high-throughput, urbanized ports.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100292"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145736298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Ensemble learning for well logging evaluation of the hybrid shale brittleness index: A case from the Gaoyou sag, Subei basin 混合页岩脆性指数测井评价中的集成学习——以苏北盆地高邮凹陷为例

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-12-11 DOI: 10.1016/j.uncres.2025.100294

Jiayi He , Taohua He , Haotian Liu , Huijun Wang , Can Huang , Jiuhong Qi , Jin Xu , Yuanzhen Zhou , Juan Teng , Yaohui Xu , Changjun Ji , Zhigang Wen

Accurate evaluation of the brittleness index (BI) is crucial for optimizing hydraulic fracturing during the extraction of hydrocarbon fluids in hybrid shale reservoirs, yet conventional petrophysical methods face limitations in scalability and generalizability. This study presents an integrated evaluation framework utilizing four ensemble learning algorithms—Random Forest (RF), AdaBoost, XGBoost, and CatBoost—to predict BI from well logging data in the second member of Funing Formation (E₁f₂) shale from the Gaoyou Sag, Subei Basin, eastern China. A dataset comprising 1295 well logging data points and mineralogical compositions from 174 core samples was used to train and validate the models. These models were optimized by Particle Swarm Optimization (PSO) to resolve nonlinear interdependencies between logging responses and mechanical brittleness. Comparative analysis demonstrates that RF achieves the highest prediction accuracy (R² = 0.84 on the test set), outperforming CatBoost (R² = 0.79), AdaBoost (R² = 0.71), and XGBoost (R² = 0.65). The superior performance of RF is attributed to its robustness against overfitting and its ability to effectively capture complex nonlinear relationships in logging responses. SHapley Additive exPlanations (SHAP) analysis identifies acoustic (AC) and resistivity (Rt) logs as the most influential predictors, reinforcing their strong physical correlations with mineralogical brittleness. This study represents the application of ensemble learning for BI evaluation in the Funing Formation shale, providing a cost-effective alternative to laboratory-based methods and demonstrating the viability of data-driven approaches for fracturability assessment. The proposed framework offers significant potential for extension to other unconventional reservoirs, contributing to enhanced hydraulic fracturing design and improved reservoir development strategies for unconventional hydrocarbon fluid development.

在混合页岩储层中，脆性指数（BI）的准确评价对于优化水力压裂过程至关重要，但传统的岩石物理方法在可扩展性和通用性方面存在局限性。利用随机森林（random Forest， RF）、AdaBoost、XGBoost和catboost四种集成学习算法对苏北盆地高邮凹陷阜宁组二段（E1f2）页岩测井数据进行BI预测，提出了一种综合评价框架。该数据集包括来自174个岩心样品的1295个测井数据点和矿物成分，用于训练和验证模型。利用粒子群算法（PSO）对模型进行优化，求解测井响应与机械脆性之间的非线性关系。对比分析表明，RF达到了最高的预测精度（R2 = 0.84），优于CatBoost （R2 = 0.79）、AdaBoost （R2 = 0.71）和XGBoost （R2 = 0.65）。RF的优越性能归功于其对过拟合的鲁棒性以及有效捕获测井响应中复杂非线性关系的能力。SHapley加性解释（SHAP）分析确定声波（AC）和电阻率（Rt）测井是最具影响力的预测指标，强化了它们与矿物脆性的强物理相关性。该研究代表了集成学习在阜宁组页岩BI评价中的应用，为基于实验室的方法提供了一种具有成本效益的替代方案，并证明了数据驱动方法在可压裂性评价中的可行性。所提出的框架具有推广到其他非常规油藏的巨大潜力，有助于提高水力压裂设计和改进非常规油气流体开发的油藏开发策略。

{"title":"Ensemble learning for well logging evaluation of the hybrid shale brittleness index: A case from the Gaoyou sag, Subei basin","authors":"Jiayi He , Taohua He , Haotian Liu , Huijun Wang , Can Huang , Jiuhong Qi , Jin Xu , Yuanzhen Zhou , Juan Teng , Yaohui Xu , Changjun Ji , Zhigang Wen","doi":"10.1016/j.uncres.2025.100294","DOIUrl":"10.1016/j.uncres.2025.100294","url":null,"abstract":"<div><div>Accurate evaluation of the brittleness index (BI) is crucial for optimizing hydraulic fracturing during the extraction of hydrocarbon fluids in hybrid shale reservoirs, yet conventional petrophysical methods face limitations in scalability and generalizability. This study presents an integrated evaluation framework utilizing four ensemble learning algorithms—Random Forest (RF), AdaBoost, XGBoost, and CatBoost—to predict BI from well logging data in the second member of Funing Formation (E<sub>1</sub>f<sub>2</sub>) shale from the Gaoyou Sag, Subei Basin, eastern China. A dataset comprising 1295 well logging data points and mineralogical compositions from 174 core samples was used to train and validate the models. These models were optimized by Particle Swarm Optimization (PSO) to resolve nonlinear interdependencies between logging responses and mechanical brittleness. Comparative analysis demonstrates that RF achieves the highest prediction accuracy (R<sup>2</sup> = 0.84 on the test set), outperforming CatBoost (R<sup>2</sup> = 0.79), AdaBoost (R<sup>2</sup> = 0.71), and XGBoost (R<sup>2</sup> = 0.65). The superior performance of RF is attributed to its robustness against overfitting and its ability to effectively capture complex nonlinear relationships in logging responses. SHapley Additive exPlanations (SHAP) analysis identifies acoustic (AC) and resistivity (Rt) logs as the most influential predictors, reinforcing their strong physical correlations with mineralogical brittleness. This study represents the application of ensemble learning for BI evaluation in the Funing Formation shale, providing a cost-effective alternative to laboratory-based methods and demonstrating the viability of data-driven approaches for fracturability assessment. The proposed framework offers significant potential for extension to other unconventional reservoirs, contributing to enhanced hydraulic fracturing design and improved reservoir development strategies for unconventional hydrocarbon fluid development.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100294"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145736297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Accurate parameter identification of solar cells and photovoltaic modules under real conditions using a new Differential Evolution approach 应用差分演化方法对太阳能电池和光伏组件在实际条件下的精确参数辨识

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-11-24 DOI: 10.1016/j.uncres.2025.100269

Driss Saadaoui, Mustapha Elyaqouti, Imade Choulli, Ismail Abazine, Abdelfattah Elhammoudy, Souad Lidighbi, Dris Ben Hmamou, Brahim Ydir, El Hanafi Arjdal, Khalid Assalaou

The global shift towards solar energy has intensified the need for advanced research on high-efficiency photovoltaic (PV) systems. Accurate extraction of PV parameters from current-voltage curves is critical for precise simulation, evaluation, and optimization of PV systems. Although numerous methods exist, many suffer from instability, premature convergence, or limited applicability to complex models. In this paper Progressive Differential Evolution with Adaptive Mutation (Pro-DEAM) is presented which eliminates the aforementioned limitations and introduces a new optimization algorithm generation. Pro-DEAM integrates two novel mutation operators: a hybrid DE/rand-to-best/1/bin to strengthen local exploitation and diversity, and a modified DE/rand/2/bin equipped with random scaling factors to dynamically promote exploration. These two strategies are adaptively weighted with exponential increasing to achieve the best trade-off between exploration and exploitation. The efficiency of Pro-DEAM is proved by extensive experiments of parameter extraction of various PV models, including complex multijunction solar cells (15 parameters) as opposed to many more complex five parameters models in traditional. The results indicate that the proposed Pro-DEAM algorithm consistently attains the lowest root mean square error (RMSE) among existing state-of-the-art approaches, highlighting its remarkable accuracy and computational efficiency. Additionally, detailed statistical analysis results confirm that it is highly robust and reliable for various PV module complexities and climates. The proposed algorithm addresses key limitations of existing methods and establishes a new benchmark for stability and convergence in PV parameter extraction. Its adaptive and dynamic framework makes it a powerful tool for both standard and advanced PV technologies, paving the way for more accurate and efficient solar energy systems.

全球向太阳能的转变加强了对高效光伏系统的先进研究的需要。从电流-电压曲线中准确提取PV参数对于PV系统的精确仿真、评估和优化至关重要。尽管存在许多方法，但许多方法存在不稳定性、过早收敛或对复杂模型的适用性有限的问题。本文提出了一种基于自适应突变的渐进差分进化算法（Pro-DEAM），它消除了上述局限性，并引入了一种新的优化算法生成。Pro-DEAM集成了两种新的变异算子：一种是混合DE/rand-to-best/1/bin，以加强局部开发和多样性；另一种是改进的DE/rand/2/bin，配备随机比例因子，以动态促进探索。这两种策略以指数递增自适应加权，以实现勘探与开采的最佳权衡。Pro-DEAM的有效性通过各种PV模型的大量参数提取实验得到了证明，包括复杂的多结太阳能电池（15个参数），而不是传统的更复杂的5个参数模型。结果表明，本文提出的Pro-DEAM算法在现有最先进的算法中始终获得最低的均方根误差（RMSE），突出了其显著的精度和计算效率。此外，详细的统计分析结果证实，它是高度稳健和可靠的各种光伏组件复杂性和气候。该算法解决了现有方法的主要局限性，并为PV参数提取的稳定性和收敛性建立了新的基准。它的适应性和动态框架使其成为标准和先进光伏技术的强大工具，为更精确和高效的太阳能系统铺平了道路。

{"title":"Accurate parameter identification of solar cells and photovoltaic modules under real conditions using a new Differential Evolution approach","authors":"Driss Saadaoui, Mustapha Elyaqouti, Imade Choulli, Ismail Abazine, Abdelfattah Elhammoudy, Souad Lidighbi, Dris Ben Hmamou, Brahim Ydir, El Hanafi Arjdal, Khalid Assalaou","doi":"10.1016/j.uncres.2025.100269","DOIUrl":"10.1016/j.uncres.2025.100269","url":null,"abstract":"<div><div>The global shift towards solar energy has intensified the need for advanced research on high-efficiency photovoltaic (PV) systems. Accurate extraction of PV parameters from current-voltage curves is critical for precise simulation, evaluation, and optimization of PV systems. Although numerous methods exist, many suffer from instability, premature convergence, or limited applicability to complex models. In this paper Progressive Differential Evolution with Adaptive Mutation (Pro-DEAM) is presented which eliminates the aforementioned limitations and introduces a new optimization algorithm generation. Pro-DEAM integrates two novel mutation operators: a hybrid DE/rand-to-best/1/bin to strengthen local exploitation and diversity, and a modified DE/rand/2/bin equipped with random scaling factors to dynamically promote exploration. These two strategies are adaptively weighted with exponential increasing to achieve the best trade-off between exploration and exploitation. The efficiency of Pro-DEAM is proved by extensive experiments of parameter extraction of various PV models, including complex multijunction solar cells (15 parameters) as opposed to many more complex five parameters models in traditional. The results indicate that the proposed Pro-DEAM algorithm consistently attains the lowest root mean square error (RMSE) among existing state-of-the-art approaches, highlighting its remarkable accuracy and computational efficiency. Additionally, detailed statistical analysis results confirm that it is highly robust and reliable for various PV module complexities and climates. The proposed algorithm addresses key limitations of existing methods and establishes a new benchmark for stability and convergence in PV parameter extraction. Its adaptive and dynamic framework makes it a powerful tool for both standard and advanced PV technologies, paving the way for more accurate and efficient solar energy systems.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100269"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145789577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Prediction of pressure build-up distribution and geomechanical analysis in a CO2 sequestration reservoir using optimized artificial intelligence models 基于优化人工智能模型的CO2封存油藏压力累积分布预测与地质力学分析

IF 4.6

Unconventional Resources

Pub Date : 2026-01-01 Epub Date: 2025-12-17 DOI: 10.1016/j.uncres.2025.100290

Emmanuel Karikari Duodu , Eric Thompson Brantson , Binshan Ju , Richard Fiifi Annan , Eugene Jerry Adjei

Injecting carbon dioxide into geological formations is a common method, but it carries significant geomechanical risks due to pore pressure buildup. This pressure increase can cause caprock failure, fault reactivation, poroelastic responses, and compromise well integrity. In this study, we develop a predictive model for effective mean stress that directly links reservoir pressure buildup to geomechanical deformation. We introduce a hybrid Artificial Intelligence (AI) workflow that forecasts pressure buildup and its effects on effective stresses, eliminating the need for extensive compositional simulations. Additionally, we compare hybrid algorithms with the traditional ADAM optimizer and Physics-Informed Neural Networks (PINN). Our investigation further examines how pressure-induced effective stresses influence CO₂ injection in reservoirs. To improve predictive accuracy and computational efficiency, we utilize innovative hybrid models that combine Artificial Neural Networks (ANNs) with advanced optimization algorithms, including the Whale Optimization Algorithm (WOA), Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), and Gorilla Troops Optimization (GTO). Among these, the ANN-GTO model exhibits the highest predictive accuracy in our tests, with a correlation coefficient of 0.99970 and an R² of 0.99941. Microseismic analysis reveals temporal and spatial clustering of induced events, strongly linked to pressure-related stress changes. Event densities increase notably over the simulation period, from 0.1 to 1 event to 30 to 220 events, providing valuable indicators for leakage risk and enabling proactive mitigation. Our results confirm that hybrid ANN models effectively predict leakage risks caused by pressure buildup and microseismic activity, thereby enhancing the safety and efficiency of CO₂ sequestration.

向地层中注入二氧化碳是一种常用的方法，但由于孔隙压力的增加，这种方法存在很大的地质力学风险。这种压力的增加可能导致盖层破裂、断层重新激活、孔隙弹性响应，并损害井的完整性。在这项研究中，我们开发了一个有效平均应力的预测模型，该模型直接将储层压力累积与地质力学变形联系起来。我们引入了一种混合人工智能（AI）工作流程，可以预测压力累积及其对有效应力的影响，从而消除了大量成分模拟的需要。此外，我们将混合算法与传统的ADAM优化器和物理信息神经网络（PINN）进行了比较。我们的研究进一步研究了压力诱导的有效应力如何影响储层中的二氧化碳注入。为了提高预测精度和计算效率，我们利用创新的混合模型，将人工神经网络（ann）与先进的优化算法相结合，包括鲸鱼优化算法（WOA）、粒子群优化算法（PSO）、灰狼优化器（GWO）和大猩猩部队优化算法（GTO）。其中，ANN-GTO模型在我们的检验中表现出最高的预测准确率，相关系数为0.99970，R2为0.99941。微震分析揭示了诱发事件的时空聚类，与压力相关的应力变化密切相关。在模拟期间，事件密度显著增加，从0.1到1个事件增加到30到220个事件，为泄漏风险提供了有价值的指标，并实现了主动缓解。研究结果表明，混合人工神经网络模型能够有效预测压力累积和微地震活动引起的泄漏风险，从而提高CO2封存的安全性和效率。

{"title":"Prediction of pressure build-up distribution and geomechanical analysis in a CO2 sequestration reservoir using optimized artificial intelligence models","authors":"Emmanuel Karikari Duodu , Eric Thompson Brantson , Binshan Ju , Richard Fiifi Annan , Eugene Jerry Adjei","doi":"10.1016/j.uncres.2025.100290","DOIUrl":"10.1016/j.uncres.2025.100290","url":null,"abstract":"<div><div>Injecting carbon dioxide into geological formations is a common method, but it carries significant geomechanical risks due to pore pressure buildup. This pressure increase can cause caprock failure, fault reactivation, poroelastic responses, and compromise well integrity. In this study, we develop a predictive model for effective mean stress that directly links reservoir pressure buildup to geomechanical deformation. We introduce a hybrid Artificial Intelligence (AI) workflow that forecasts pressure buildup and its effects on effective stresses, eliminating the need for extensive compositional simulations. Additionally, we compare hybrid algorithms with the traditional ADAM optimizer and Physics-Informed Neural Networks (PINN). Our investigation further examines how pressure-induced effective stresses influence CO<sub>2</sub> injection in reservoirs. To improve predictive accuracy and computational efficiency, we utilize innovative hybrid models that combine Artificial Neural Networks (ANNs) with advanced optimization algorithms, including the Whale Optimization Algorithm (WOA), Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), and Gorilla Troops Optimization (GTO). Among these, the ANN-GTO model exhibits the highest predictive accuracy in our tests, with a correlation coefficient of 0.99970 and an R<sup>2</sup> of 0.99941. Microseismic analysis reveals temporal and spatial clustering of induced events, strongly linked to pressure-related stress changes. Event densities increase notably over the simulation period, from 0.1 to 1 event to 30 to 220 events, providing valuable indicators for leakage risk and enabling proactive mitigation. Our results confirm that hybrid ANN models effectively predict leakage risks caused by pressure buildup and microseismic activity, thereby enhancing the safety and efficiency of CO<sub>2</sub> sequestration.</div></div>","PeriodicalId":101263,"journal":{"name":"Unconventional Resources","volume":"9 ","pages":"Article 100290"},"PeriodicalIF":4.6,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145789666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0