For early recognition of flow regime in pipeline-riser system, use ofsignals from accelerometers are suggested instead of a commonly usedpressure gauge signal. A main factor that discriminates between stableflow and severe slugging is drastically decreased vibration in liquidaccumulation of the severe slugging. In this study, a support vectormachine is employed for binary classification to identify between stableflow and severe slugging. For multi-class classification, a neuralnetwork is applied to recognize four classes of stable flow, two types ofsevere slugging, and irregular transition. The performance is alsoanalyzed based on the signal length employed.
展开▼
