首页> 外国专利> Fremgangsmåte og system for sanntids-estimering av strömningsmodus for en flerfasefluidström i alle punkter i et rör

Fremgangsmåte og system for sanntids-estimering av strömningsmodus for en flerfasefluidström i alle punkter i et rör

机译:用于实时估计管道中所有点的多相流体流流动模式的方法和系统

摘要

- m method and system for estimating in time r olc the method of the discharge at all points of a pipe of the structure of efinie by a certain number of parameters of the structure of a multiphase fluid stream of efinie by a plurality of magnitudes physical. - it is in the form of a neural r network not lin eaire with a layer of between ee with as many between ed as the parameters of the structure and physical magnitudes, an output layer with as many outputs as parameters n ecessaires to the estimation of the method of the discharge and at least one layer interm ediate. It constitutes a learning base with the tables pr ed efinies connecting diff erent values obtained for the data ed of the output to the corresponding values of data ed of between ee, and it erations etermination by factors eration corresponds to the activation function making it possible to connect the values in the tables of data ed of between ee and output. In order to avoid the singularity data ed the output of the r network, which can distort the etermination corresponds eration factors, there is a sorting to dispose of data ed not relevant. The main advantages of the m method are: simplification of the mod ical modeling and gain period - applications, for example, to the r embodiment simplified ee of the hydrodynamic modules that can be int egrer, for example, to the tools of mod ical modeling.
机译:-m方法和系统,用于通过多个物理量的多态流体流的结构的一定数量的参数来及时估计在该结构的管道的所有点处的排放的方法。 -它的形式是神经网络,而不是线性的,在ee和ed之间的层与结构和物理量级的参数一样多,在输出层上的输出与参数n一样多,以估计e放电方法和至少一层中间层。它构成了一个学习基础,其中的表pr定义将对输出数据ed所获得的不同值连接到ee之间的数据ed的对应值,并且因数确定er对应于激活函数,从而使得在ee和输出之间连接数据ed表中的值。为了避免奇异的数据ed的r网络的输出,这可能会使确定的对应因子失真,对不相关的数据ed进行分类处理。 m方法的主要优点是:简化了模型建模和增益周期-例如,将其应用到流体动力学模块的简化实施例中,该模型可以集成到例如模型建模工具中。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号