Abstract

Mechanistic steady state simulation models, rather than empirical correlations, are used for a design of multiphase production system including well, pipeline and downstream installations. Among the available models, PEPITE, WELLSIM, OLGA, TACITE and TUFFP are widely used for this purpose and consequently, a critical evaluation of these models is needed. An extensive validation methodology is proposed which consists of two distinct steps : first to validate the hydrodynamic point model using the test loop data and, then to validate the over-all simulation model using the real pipelines and wells data. The test loop databank used in this analysis contains about 5952 data sets originated from four different test loops and a majority of these data are obtained at high pressures (up to 90 bars) with real hydrocarbon fluids. Before performing the model evaluation, physical analysis of the test loops data is required to eliminate non-coherent data. The evaluation of these point models demonstrates that the TACITE and OLGA models can be applied to any configuration of pipes. The TACITE model performs better than the OLGA model because it uses the most appropriate closure laws from the literature validated on a large number of data. The comparison of predicted and measured pressure drop for various real pipelines and wells demonstrates that the TACITE model is a reliable tool.

Introduction

Nowadays, the multiphase production technique has become a standard method for the development of marginal fields due to significant reduction in CAPEX and OPEX compared to those obtained in the conventional production method. The multiphase production technology consists in the transport of untreated wellhead effluent through a single pipeline either to an existing processing platform or to on-shore processing facilities. In the beginning of the production, the wellhead pressure can be sufficient, depending on the transportation distance, to achieve transportation up to the treatment installations. If the wellhead pressure is not sufficient, multiphase pumps can be used to boost the inlet pressure. The correct predictions of pressure drop, liquid hold-up and flow regime are vital to design a multiphase transport system i.e. pipeline, multiphase pump and downstream process facilities.

The old approach to design a multiphase pipeline was to use empirical correlations developed from the test loop experimental data. Nowadays, these correlations are hardly used as they do not take the physical phenomena into consideration and the accuracy of their predictions can be very poor.

Since the mid 1970's, significant progress has been made to understand the physics of two-phase flow in pipes and wells. This progress has resulted in several two-phase flow mechanistic models to simulate the pipelines and wells under steady state as well as transient conditions. The use of these steady state mechanistic models allows to correctly predict the pressure drop, liquid hold-up and flow regime for a pipeline and well operating under any condition.

The main purpose of this paper is first to define a proper methodology to carry out the validation of a mechanistic two-phase flow pipeline and well simulation model. Then, in accordance with the established methodology, a critical evaluation of the existing pipeline and well simulation models has been carried out using a large number of test loop, real pipeline and well data. This study identifies the domain of application, limitations and weaknesses of each model.

Brief description of different mechanistic models As mentioned earlier, most oil companies use mechanistic models rather than empirical correlations to simulate two-phase flow pipelines and wells. Several mechanistic models exist, but only the following models have been investigated in this paper either because they are commercially available or reputed as being good tools:

  • PEPITE

  • WELLSIM

P. 393

This content is only available via PDF.
You can access this article if you purchase or spend a download.