Virtual inflow estimation with simplified tuning using cascaded and Kalman-like least squares observers

Abstract

Multiphase metering is often unavailable, unreliable or expensive, which makes model-based, virtual monitoring a cheap and reliable alternative in order to obtain information about e.g. the inflows to a device. This paper presents an approach for inflow estimation to a gravity separator based on a standard Extended Kalman Filter. Potential improvements include a cascaded structure, a Kalman-like least squares observer with forgetting factor, and a combination of these.