基于鲸鱼算法优化AM-BiLSTM模型的储层产气量预测

产气量是评估天然气井生产能力和开发工艺效果的重要指标。准确的预测产气量是保证高效生产的关键。为了准确预测储层产气量,本文提出了一种基于鲸鱼优化算法(whale optimization algorithm,WOA)、注意机制(attention mechanism,AM)和双向长短时记忆(bi-directional long short-term memory,BiLSTM)相结合的日产气量预测模型。首先,以Volve油田的排采数据为研究对象,分析了产气量与这些排采参数之间的相关性,并利用LightGBM算法进行重要性排序;然后使用有较强非线性处理能力的基于注意力机制的双向长短时记忆神经网络(bi-directional long short-term memory based on attention mechanism,AM-BiLSTM)构建日产气量预测模型;最后通过鲸鱼算法对AM-BiLSTM模型中的相关参数进行优化,并将参数优化后的模型(WOA-AMBiLSTM)应用于Volve油田的A井。实验结果表明,WOA-AMBiLSTM模型的综合预测性能优于传统的反馈神经网络模型(the back-propagation neural network model,BP)和其他提出的深度学习模型(LSTM、BiLSTM和AM-BiLSTM)。WOA-AM-BiLSTM模型预测曲线与实测测井曲线更加接近,具有更好的预测表现,为储层产能预测提供了一种新思路。

强电磁干扰区大地电磁远参考技术试验效果分析

铜川市调查区工业干扰强烈,地层结构复杂。为了研究在城市强电磁干扰环境下大地电磁远参考技术应用效果,在铜川市调查区布设了一条音频大地电磁法剖面,采用非线性共轭梯度(NLCG)法进行数据反演,并结合工作区地质资料进行分析。结果表明:应用远参考技术对大地电磁数据进行处理,可以有效改善受噪数据质量,反演成果与地质资料更为吻合。

附加列插值法在起伏地形条件下二维大地电磁正演中的研究与应用

基于双二次插值的有限单元法进行大地电磁二维正演数值模拟研究,在传统的以整个模型为对象进行插值计算的基础上,提出了附加列计算法并设计起伏地形模型进行正演计算,分别计算了TE模式和TM模式下的电阻率,计算结果表明,该插值方法计算精度高,抗干扰能力强,尤其在TM模式下能很好地模拟起伏地形对视电阻率的影响。

Groundwater resources survey of tongchuan city using the audio magnetotelluric method

The development of Tongchuan City,Shaanxi Province,located in the northwestern region of China,is restricted by water resources.The direct current resistivity and induced polarization sounding methods are typically applied in finding urban groundwater.These methods,however,are not effective due to their complicated topography and geological conditions.The application practice shows that the audio magnetotelluric(AMT)method has a large depth of exploration,high work effi ciency,and high lateral resolution.To investigate the distribution of groundwater resources,we deployed three audio-frequency magnetotelluric profiles in the city area.The impedance tensor information of AMT data is obtained using SSMT2000.AMT data dimension analysis reveals that the two-dimensional structural features of the observation area are obvious.The main structure of the observation area is about 45°northeast,as indicated by structural trend analysis.A shallow two-dimensional electrical profile of 1 km in Tongchuan City is obtained by two-dimensional nonlinear conjugate gradient inversion.Finally,combined with regional geological information,the geological structure characteristics reflected by the electrical profile were obtained along with the detailed characteristics of water-rich structures in the area.The infl uence of the structure on the groundwater distribution was analyzed,and the water-rich areas were identifi ed.This work contributes to the prospective development of Tongchuan City.

综合物探方法在广西大塘深部隐伏铅锌矿勘查中的应用效果

大塘铅锌矿位于博白—岑溪铅锌金银多金属成矿带上,结合该区域地质矿产特征,及前人在本区工作的地面高精度磁测和土壤地球化学测量成果,提出了采用激电和电磁测深等综合物探方法勘探的思路,利用激电扫面圈定了大塘矿区局部低阻高极化异常,利用激电对称四极测深和音频大地电磁测深查明异常的空间展布形态,在空间上圈定找矿靶区,开展钻孔验证,取得较好的找矿效果,表明了综合物探方法在本区寻找全隐伏铅锌矿的有效性。