Performance-based fractal fracture model for complex fracture network simulation

The paper presents a novel hydraulic fracturing model for the characterization and simulation of the complex fracture network in shale gas reservoirs. We go beyond the existing method that uses planar or orthogonal conjugate fractures for representing the ＇＇complexity＇＇ of the network. Bifurcation of fractures is performed utilizing the Lindenmayer system based on fractal geometry to describe the fracture propagation pattern, density and network connectivity. Four controlling parameters are proposed to describe the details of complex fractures and stimulated reservoir volume（SRV）. The results show that due to the multilevel feature of fractal fractures, the model could provide a simple method for contributing reservoir volume calibration. The primary-and second-stage fracture networks across the overall SRV are the main contributions to the production, while the induced fracture network just contributes another 20% in the late producing period. We also conduct simulation with respect to different refracturing cases and find that increasing the complexity of the fracture network provides better performance than only enhancing the fracture conductivity.

机电传感器的特征化及其特征参数的确定

本文指出用Z参数和ABCD参数分别表示机电传感器在接收状态和发生状态下的网络参数是方便且合理的。然后本文规定接收器灵敏度和发生器发射电压和电流响应作为其相应的特征化参数。进而,本文提出一个三参数模型描述传感器的行为,并发展一个复数最小二乘拟合Fortran计算方法求出数学介。这样,机电传感器的网络参数和特征化参数可从它的一组电阻抗和负载阻抗数据对确定。

Jordan神经网络的改进研究

针对Jordan神经网络的反馈网络的反馈信息表征能力不强的缺点,提出了一种新的反馈网络模型,对Jordan神经网络的缺点进行了改进,并且对原来的训练学习算法进行了改进,提出了一种提取绝对值最大权的训练学习算法来降低计算复杂性,最终给出了实验结果证明。

复杂电磁环境量化表征模型

复杂电磁环境量化表征是构建逼真复杂电磁环境的基础。综合考虑环境总体和对装备影响两个层面,构建了一套复杂电磁环境量化表征模型体系,该体系从电磁信号和电磁网络两个方面对复杂电磁环境的量化表征问题进行研究,给出了相应的函数表征模型。