A NEW METHOD FOR PREDICTION OF PRODUCTIVITY OF FRACTURED HORIZONTAL WELLS BASED ON NON-STEADY FLOW

Hydraulic fracturing technologies of horizontal well are important ways to develop oil-gas field with low permeability. Productivity forecast of fractured horizontal wells is a difficult problem of hydraulic fracturing technologies. Basing on non-steady flow of fractures fluid during production, applying potential function principles, superposition principle and mathematical method for solving, coupling of seepage flow in the formation and pipe flow in the well bore, a new model on multi-fracture interference productivity forecast of fractured horizontal well is established in this article. The results indicate the coincidence rate between this model and practice is high. The pressure loss in the horizontal well bore has definite influence on the production status of fractured horizontal wells. The productions of different fractures in horizontal well bore are unequal, the productions of outer fractures are higher than middle fractures; the pressure in the well bore shows an uneven distribution, the pressure declines gradually from finger tip to heel end. Asymmetry of fractures may make productivity of fractured horizontal wells decline. The conclusions are instructive in designing fractured horizontal well for low permeability reservoir.     

Liquid drop impact on solid surface with application to water drop erosion on turbine blades,Part I: Nonlinear wave model and solution of one-dimensional impact

Water drop erosion is regarded as one of the most serious reliability concerns in the wet steam stage of a steam turbine. The most challenging aspect of this problem involves the fundamental solution of the transient pressure field in the liquid drop and stress field in the metal substrate, which are coupled with each other. We solve the fundamental problem of high-speed liquid–solid impact both analytically and numerically. In Part I of this paper, the governing equations based on a nonlinear wave model for liquid are derived. Analytical and approximate solutions of one-dimensional liquid–solid impact are given for both linear and nonlinear models, which provide critical insights into the water drop erosion problem. Both continuous and pulsant impacts on rigid and elastic substrates are analyzed in detail. During continuous impact, the maximum impact pressure is always higher than the water hammer pressure. Upon pulsant impact and at a particular instant related with the impact duration, the maximum tensile stress appears at a certain depth below the solid surface, which can be readily related with the erosion rate. In Part II of this paper, two-dimensional (axisymmetric) liquid–solid impact is solved numerically, from which the most dangerous impact load/duration time and the most likely crack positions are deduced. Based on our recent solution of the water drop impact statistics (associated with the fluid flow in the blade channel), a comprehensive numerical study of the water drop erosion (fatigue) on a turbine blade is carried out.     

油气田CO2/H2S共存腐蚀与缓蚀技术研究进展

CO₂/H₂S是油气田采集、运输、处理过程中主要的腐蚀介质，由其引起的管道设备的腐蚀问题变得越来越严重，腐蚀和防腐已经成为研究热点。分别对近年来国内外开展的有关CO₂和H₂S共存腐蚀及缓蚀技术的研究进行综述。CO₂/H₂S共存腐蚀研究主要依靠试验手段，但目前的研究结果有很大的离散性，根据不同的试验条件会产生不同的研究结果。分压比是国内外大多数学者研究CO₂/H₂S腐蚀规律的切入点，但关于两者主导腐蚀的分压比界限划分现有研究存有争议。缓蚀技术研究讨论了缓蚀剂作用机理，评述了抑制CO₂/H₂S共存腐蚀常用的酰胺类、咪唑啉衍生物类、季铵盐类和Schiff碱类缓性剂在国内外的研究与应用现状，展望了这一领域的研究前景及发展方向。     

Flame synthesis of carbon nanotubes and few-layer graphene on metal-oxide spinel powders

Multi-walled and single-walled carbon nanotubes (CNTs) and few-layer graphene (FLG) are grown directly on spinel powders using flame synthesis. CNT and FLG growth occurs via the decomposition of flame-generated carbon precursors (e.g., CO, C₂H₂, and CH₄) over nanoparticles (i.e., Ni, Co, Fe, and Cu) reduced from the solid oxide. The growth of CNTs is investigated on NiAl₂O₄, CoAl₂O_4, and ZnFe₂O₄, using counterflow diffusion flame and multiple inverse-diffusion flames (m-IDFs), while the growth of FLG is investigated on CuFe₂O₄ using m-IDFs. As shown by analytical electron microscopy techniques, Raman spectroscopy, and X-ray diffraction, substrate temperature and spinel composition play critical roles in the growth of both CNTs and FLG.     

Boundary layer solution for the turbulent swirling decay flow through a fixed pipe: SBR at the inlet

In this study the developing turbulent swirling pipe flow is investigated both numerically and analytically. Governing equations are derived accompanying the boundary layer assumptions. Uniform and solid body rotation (SBR) distributions are taken into account for the axial and tangential velocities at the inlet of the pipe, respectively. Beyond the boundary layers, the flow pattern is considered to be the potential flow. Making use of the fourth-order Runge–Kutta scheme, the numerical solution of the differential equations is obtained. Further more, by simplifying the governing equations for large Rossby number, the analytical solution is performed. The results of numerical and analytical swirl intensity have been compared which shows reasonable agreement. As an alternative solution, a CFD analysis has been done, having applied FLUENT software to support the ability of our methodology.     

A novel MPPT controller based PEMFC system for electric vehicle applications with interleaved SEPIC converter

The Proton Exchange Membrane Fuel Cells (PEMFCs) are one of the most effective and optimistic renewable energy source and they are extensively used in automotive applications. In the past, many researchers focused on solving the issues of extracting maximum power from fuel cell, controlling the speed and reducing the torque ripple of Brushless DC (BLDC) motor for fuel cell based Electric Vehicle (EV) systems. However, it is challenging to fine-tuning the gain parameters in the existing works MPPT approach and extracting maximum amount of energy. Additionally, it has limitations like unregulated voltage, problems of large overshoot, slow tracking speed, output power fluctuation, computational complexity and intricate modeling. Thus, the proposed work aims to create a revolutionary methodology called Unified Firefly Ersatz Neural Network (UFENN) - Maximum Power Point Tracking (MPPT). The UFENN is a kind of optimization-based machine learning technique that was created for efficiently optimizing the parameters to extract the maximum energy from the fuel cells. Furthermore, in order to control the output voltage with the least amount of power loss, an Interleaved SEPIC converter is also used in this work. During performance analysis, an extensive simulation results have been taken for validating the results of the proposed scheme by using various evaluation indicators.     

Combination of thermodynamic analysis and regression analysis for steam and dry methane reforming

Hydrogen is predominantly produced via methane reforming. In this study, thermodynamic analysis and regression analysis of the steam reforming of methane (SRM) as well as dry methane reforming (DRM) are conducted. The method of Gibbs free energy minimization is applied for investigating the effect of factors, such as temperature, pressure, and inlet composition, on the performance of hydrogen production. Notably, this study is not restricted to the effect of a single factor, but to the combined results of all independent variables. Then, regression analysis is adopted for examining the quantitative relationship between response observed and conditions. As a result, different mathematical expressions are attempted, such as linear regression, second-order polynomial, and logarithmic form, for finding the optimal form to preferably illustrate the manner in which factors affect performance. In this process, the forms are compared in several ways from the perspective of not only regression parameters but also error bars on graphs curve images. Finally, a three-pieces-logarithmic model is proposed as the final form to explain the relation between factors and response with maximum error 7% and the most deviations range between 0% and 2%.     

无监督网络对抗生成模型的研究

一般的学习模型都是基于一个假设的随机分布,然后通过训练真实数据来拟合出模型。网络模型复杂并且数据集规模也不小,这种方法简直就是凭借天生蛮力解决问题。Goodfellow认为正确使用数据的方式,是先对数据集的特征信息有insight之后,再干活。无监督学习是当下较为流行的话题,但也是困难较为繁多的话题。目前无监督学习可以分成以下两类,分别是确定型的自编码方法以及概率型的受限波尔兹曼机,其目标主要是使受限玻尔兹曼机达到稳定状态时原数据出现的概率最大。如何更快速更有效地地搭建模型以及如何做实验并有效地获得相关的实验结论是人们讨论的重点。在研究中,在判别模型中增加正则化,用卷积层代替池化层,在生成模型中输出层使用tanh激活函数激活,这样使得最终运算的准确率和损失率大大下降,并减少了冗余成分。     

Steam methane reforming over structured reactors under concentrated solar irradiation

Intermittent nature of solar energy and solution strategies for steam methane reforming reaction powered by concentrated solar energy over Ni/mullite and Pd/CeO₂/mullite catalysts were demonstrated. The solar concentration was achieved using a parabolic mirror with a 70 cm, delivering concentrated solar flux onto a focal area that is approximately 3 cm in diameter. The solar field tests conducted on monolithic catalyst support structures were compared with the laboratory scale measurements on powdered catalysts. Despite the fluctuations in solar irradiation, CH₄ conversions higher than 90% could be obtained. Coke deposition was observed over the 15%Ni/Mullite monolith. On the 1%Pd/20%CeO₂/Mullite monolith, the oxidative nature of the catalyst resulted in oxidation reactions with local temperatures exceeding 1700 °C, inferred through the melting point of mullite. Numerical simulations revealed temperature gradients as large as 500 °C, over the refractory monoliths.     

Strain energy release rate in shaft-loaded blister tests for composite repairs on steel

The design of composite repairs of corroded oil and gas pipelines must take into account the strength of the interface adhesion between composite and metal. A shaft-loaded blister test is a common method to measure interface fracture toughness and energy release rate. The study aimed on evaluating shaft-loaded blister tests as replacements for more complex pressure blister tests. Specimens investigated were thick fibre-reinforced plates bonded on metal disks as substrates containing a circular through-hole defect. This paper presents the influence of different punch head geometries on the resulting energy release rates and compares the results with blister tests using fluid pressure. Test and simulation results are presented and analytical solutions were derived and evaluated to establish best fitting formulations. It was shown, that significant variations between the different means of loading exist.