Models and Estimators for Flow of Topside Drilling Fluid

Abstract

The reduction of risk and non-productive time in oil drilling is a key research interest in the oil and gas industry. The early detection of kick and loss is a crucial part in safe well control operations, thus, it plays a major role in this regard. Early kick and loss detection is done by incorporating the available pressure data of the bottom side of the well with the available data at the surface on the topside. The data on the topside is mainly the return flow rate and the mud pit level. There are advanced flow measurement techniques available for the clean flow going into the well, which is comparatively easy to measure. On the contrary, the return flow consists of drill cuttings and gases which makes flow measurement difficult and inaccurate. Although there exist many flow meters that can measure the return flow rate, most of the on-shore and off-shore oil rigs still use conventional drilling systems. These conventional drilling processes use intermittent or online return flow rate and density measurements together with mud pit levels for kick and loss detection. There are various flow meters used in these processes, but most of the time paddle flow sensors are used. These have comparatively less accuracy as well as repeatability. In most of the conventional oil rigs, this is just an indicator rather than a realtime flowmeter, thus early kick and loss detection cannot be expected. Advances in flow metering technology will provide accurate differential flow measurements. Therefore, the development of cost-effective, accurate and online sensors for early kick and loss detection is vital.

The development of an efficient model based real-time estimator of the flow rate of the return flow using an open Venturi channel is studied in this research work, such that it can be used as a return flowmeter for early kick detection in conventional drilling. Different mathematical models are investigated for this purpose, and a suitable numerical solver for the models are developed based on the orthogonal collocation for real-time implementation. The effect of different types of drilling fluids and different geometries of channels are studied. The flow rate and various parameters like the friction factors are estimated in real-time using different estimators. The models and estimators are tested against a well-known numerical scheme and verified using experimental results from a test flow loop.

Further, the combination of two kick detection indicators, the return flow rate and the active mud pit level, are investigated in a modeling environment. For this, a combined model which includes both the bottomside and the topside of an oil well drilling process is developed and simulated to study the behavior of these indicators for different drilling operation scenarios. The model is able to show the bottomside pressure dynamics and the corresponding topside flow dynamics at once. This gives rise to a complete closed-loop model of an oil well drilling. The drilling fluid losses that can occur during the removal of drill cuttings using the solid removal equipment are estimated from these models. With the availability of real-time estimation of drill fluid losses at the top side, the replenishing of the lost mud could potentially be automated.