Design and comparison of adaptive estimators for Under-balanced Drilling

Abstract

Abstract-Real time knowledge of total mass of gas and liquid in the annulus and geological properties of the reservoir is very useful in many active controllers, fault detection systems and safety applications in the well during petroleum exploration and production drilling. Sensors and instrumentation can not measure the total mass of gas and liquid in the well directly and they are computed by solving a series of nonlinear algebraic equations with measuring the choke pressure and the bottom-hole pressure. This paper presents different estimator algorithms for estimation of the annular mass of gas and liquid, and production constants of gas and liquid from the reservoir into the well during Under Balanced Drilling. The results show that all estimators are capable of identifying the production constants of gas and liquid from the reservoir into the well, while the Lyapunov based adaptive observer gives the best performance comparing with other methods when there is a significant amount of noise.

Evaluation of Lyapunov-based Adaptive Observer using Low-Order Lumped Model for Estimation of Production Index in Under-balanced Drilling

“…The simplified LOL model equations for mass of gas and liquid in the annulus are derived from mass and momentum balances as follows (Nikoofard et al (2014a))…”

“…Fast changing parameters are estimated online simultaneously with the states of the model, but other parameters are calibrated infrequently and offline. Nikoofard et al (2014a) designed Lyapunov-based adaptive observer, a recursive least squares estimator and a joint unscented Kalman filter based on a LOL model to estimate states and parameters during UBD operations by using the total mass of gas and liquid as measurements calculated from pressure measurements using a model. The performance of the adaptive estimators were compared and evaluated for pipe connection procedure using a simple simulation model.…”

“…The performance of the adaptive estimators were compared and evaluated for pipe connection procedure using a simple simulation model. In this paper, states and parameters are estimated by directly using real-time measurements of the choke and the bottom-hole pressures and using an extension of the observer in Nikoofard et al (2014a) The purpose of the paper is to evaluate the LOL model for estimation of PI in UBD employing an adaptive observer that uses the bottom hole and choke pressure measurements. This paper describes the design of a novel Lyapunov-based adaptive observer and the joint unscented Kalman filter for estimating the total mass of gas and liquid in the annulus and production constant of gas and liquid from the reservoir into the well based on a LOL model for UBD operation.…”

State and Parameter Estimation of a Drift-Flux Model for Underbalanced Drilling Operations

“…The Low-Order Lumped (LOL) models are simpler methods that can be used. However, these models are only able to capture the major effects in the well and for the general purpose it produces less accurate results [8], [9], [10]. Due to the complexity of the multi-phase flow dynamics of a UBD well coupled with a reservoir, the modeling, estimation and model based control of UBD operations are still considered emerging and challenging topics within drilling automation.…”

“…The unscented Kalman filter was used to estimate the states, and the friction and choke coefficients. Nikoofard et al [9] designed a Lyapunovbased adaptive observer, a recursive least squares estimation and a joint unscented Kalman filter based on a low-order lumped model to estimate states and parameters during UBD operations. This model was the extended version of an adaptive observer used in [9] for directly using real-time measurements of the choke and the bottom-hole pressures to estimate states and parameters [13].…”