摘要
基于油气水多相流混合物在管道内流动时的多传感器信号———压力和压差信号 ,建立特定流型的识别规则 ,采用学习矢量量化模式分类器作为未知流型的分类器 ,根据数据融合的技术思路获得了油气水多相流的流型在线识别技术 .通过系统集成的手段 ,利用微处理装置研制出了流型识别系统 .该识别技术具有置信度高、容错性好、性能稳定、降低了对单个传感器的性能要求等优点 .该识别系统结构简单、无运动部件、不改变管道的截面结构 ,可以处理非快速变化的瞬态流动 ,测量误差小于 10 % ,并具有连续工作、定时打印。
The recognition of oil-gas-water multiphase flow regime is investigated. Based on multi-sensor data-pressure signal and differential pressure signal, containing sufficient information about the multiphase piping flow, the recognition rule for special flow regime is established by the aid of the signal feature analysis experimental flow parameters. A classifying approach is secondly induced using the learn quantification vector arithmetic to recognize the unknown flow regime. And then the recognition technology is proposed which is a fusion method combining the rule of flow regimes with the classifying approach. The technology has the advantage of good reliability, error-capability, stable performance, and is lack of the strict request for each sensor. The recognition system is developed by means of microprocessor integrating technologies so it has a simple structure without moving device, and can be applied to a not fast transient flow with a measuring error lower than 10 percent. The system also functions such as remaining the original pipe structure, continuous utilizing, real-time print and signal transport.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2003年第3期306-309,共4页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目 (5 0 0 0 60 10
5 9995 460 2 )
