声电双模态阵列传感器油气水三相流相含率融合估计方法

The phase-fraction multi-level fusion estimation model based on acoustic-electric dual modality array sensor will be studied for the horizontal oil-gas-water three-phase flow which widely exists in petroleum, chemical industry and other industrial processes. The coupled numerical simulation method of the three-phase flow field and the acoustic field formed by acoustic array sensor will be used to explore the theory of multiple propagation modes and ultrasonic multi-sensitivity effects in non-uniformly distributed fluids. The coupled numerical simulation method of the three-phase flow field and the electric field form by conductance/capacitance array sensor will be used to explore the variation characteristics and analyze the measurement mechanism when the flow state changes. According to the physical model of oil-gas-water three-phase flow, the measurement model and state model will be constructed based on the acoustic-electric array sensor structure in order to design the phase fraction estimation model of acoustic modality and conductance/capacitance mode, and the electrical modality phase fraction fusion model will also be designed based on the associated information between the modes of the electrical modality to realize the initial fusion process. The correlation of the information acquired by dual modality sensors will be analyzed. The phase fraction information coupled with oil and water information obtained by the acoustic modality will be used to realize the adaptive cancellation with the coupling phase fraction information obtained by the electrical modality after the multi-scale decomposition. The adaptive information cancellation can be used to reduce the influence of the coupling information of each modality on the phase fraction calculation. Acoustic-electric dual modality array sensor phase fraction fusion estimation model will be constructed to realize the decoupling of the three-phase phase fraction. The robust performance of the model will be optimized to realize the online real-time phase fraction estimation of oil-gas-water three-phase flow in horizontal pipelines.

针对石油、化工等工业过程中广泛存在的水平油气水三相流，研究基于超声、电学双模态阵列传感器的相含率多层融合估计模型。通过三相流场与声场的耦合数值仿真，探索在非均匀分布流体中超声多种传播模式规律和超声多敏感效应特性；通过三相流场与电导/电容电场的耦合数值仿真，探索流态变化时的电场变化规律，分析测量机理；以三相流物理模型为核心，结合声、电阵列传感器结构，构建各模态/模式的测量模型和状态模型，设计声模态、电导/电容模式相含率估计方法，并基于电模态各模式间的关联信息，设计电模态相含率融合模型实现初步融合；开展双模态耦合相含率信息关联性分析，将声模态测量得到的含有油水耦合信息的气相含率与电模态融合得到的耦合相含率信息进行多尺度分解后自适应对消，降低相间耦合信息对相含率计算的影响，构建声电双模态阵列传感器相含率融合估计模型，实现三相含率解耦，优化鲁棒性能，实现水平管道油气水三相流相含率在线实时估计测量。

多相流动广泛存在于工业过程中，是现代能源与环境领域多相流与工程热物理研究的常见对象。相含率作为多相流动特有的重要参数，对探索流动状态转换和相间相互作用规律，准确监控多相流动状态，保证生产和输运安全具有重要意义。但是多相流流动状态复杂多变，单一模态传感器因其测量原理限制，只能获得被测流体某一方面的特性，在多特征信息的获取和参数测量上难以实现全面的观测描述，需要采用多种模态传感器进行测量：.1）通过设计基于电导六环结构、电容双对极板结构和超声结构组成的多模态传感器，并设计、搭建、实现了一种基于CPCI总线，以FPGA为控制核心的电学-超声多模态测量系统，通过可靠的板间通讯，可实现不同模态灵活组合测量，从而满足不同的测量需求；.2）分别针对气水两相流和水基油气水三相流进行流型识别，进行时域、频域和时频域分析，对得到的多域特征信息进行量化，得到不同流型的高维特征向量，采用等距特征映射Isomap方法综合特征结构关系，并将其映射到低维空间。通过对低维特征向量进行可视化分析得到流型识别模型，可实现气水两相流和水基油气水三相流可视化流型识别。.3）针对低流速下的油水分散流，利用所提出的分形修正超声衰减模型，建立所计算的Sauter平均滴径与分形维数之间的关系，表征分散相非均匀分布对衰减的影响。以分形修正超声衰减模型作为正向求解器且以RL-CMA-ES算法作为反向求解器的扫频超声衰减方法可以有效实现不同流速下非均匀流动的油水分散流相含率的测量。所测油水分散流相含率的最大误差为2.45%，平均误差为0.93%。.利用非侵入、非扰动手段，对多相流动相含率进行安全、实时、快速、稳定、准确的测量，不仅对多相流动机理的发展具有推动作用，而且深入探索基于测量模型的信息融合方法，并将其广泛应用于工程实际中，具有重要的科学和工程价值。