Parameter estimation for estimation of bottom hole pressure during drilling.

Abstract

In this thesis we examine four bottom hole pressure estimators based on adaptive estimation of the friction pressure for the drill string and the annulus. Knowledge about the bottom hole pressure is crucial to achieve security and commercial objectives. Bottom hole pressure measurements transmitted by mud pulse telemetry have limited bandwidth and it is common to use additional models to estimate the bottom hole pressure when measurements are unavailable. The motivation for an adaptive approach is that several parameters are encumbered with high uncertainty and are continuously, but slowly changing. The suggested online parameter estimation scheme is implementable in a wellsite computer system and may be a valuable tool in drilling operations. The frictional pressure loss estimates use the relation between the frictional pressure loss and the flow through the system. It is shown that the pump flow is not a good estimate for the flow in this context. The best results were obtained both for mean flow estimates and for an adaptive bit-flow observer. The parameter estimation scheme has been tested using the drilling simulatorWemod developed by IRIS. Simulations shows that all four bottom hole pressure estimates are within a drilling window of ±2.5 bar. To further develop this solution we recommend that the system is tested on data from a real well.