Page 1 January 2015

FLOW ASSURANCE

CASE STUDY USING STEADY STATE & TRANSIENT TOOLS

Introduction

As operators start to explore more remote hydrocarbon plays, our production systems are exposed to more challenging
environments and fluid behaviour. In this contexts, any operational decision can have far reaching implications upon recovery. As
such, our decision making process must be rooted in understanding of the environment with respect to the physics. This is where
(in recent years) the flow assurance discipline has taken a central role in guiding operations.

In the field, turndown of production rate is often required so that routine maintenance operations can be performed on parts of
the production system, such as ESP maintenance, wellhead maintenance, and pigging operations. As the production system is
subjected to a lower throughput, the concept of turndown stability becomes important. Lower throughput, or even shut-in can lead
to operationally difficult scenarios such as wax blockages, hydrate formation, asphaltene deposition, riser slugging etc. all of which
are undesirable. Traditionally, this has been the domain of the flow assurance discipline, where modelling has been performed
exclusively by transient simulators which account for pressure/temperature changes in seconds/hours.

In the production context, production is usually considered in much larger time frequencies (recovery is considered over decades).
As such, using transient simulators for production forecasting is not practical due to the long run times, which limits their
applicability to forecasting. Thus in terms of field planning/forecasting, steady state tools are used, where the entire production
system is modelled over decades with reasonable run times, capturing the full reservoir, well and surface network response.

In reality, the long term is made up of the aggregated short term periods, and as such the two (steady state and transient)
responses must be considered together. However, this has seldom been the case – as the traditional approach of functional silos
mean these disciplines rarely interact, and use completely separate software. As such, consistency in assumptions and learning
was rarely captured, whilst the practice of multiple engineers working on the same thing in isolation in different time periods was
prevalent.

To address this shortfall, Petroleum Experts has been advocating for many years the combination of steady state and transient
simulation, and RESOLVE (vendor neutral platform)
is used to make this work. This article show how the
LEDAflow data object can be used to fulfil the
above objectives of integration between two
traditionally different approaches.

Context
The production system below has 3 producing gas
wells that deliver fluid to the production manifold,
before entering the subsea riser which delivers fluid
to the offshore production platform.

There is an opportunity to drill another well (Well 4)

once the gas production comes off production
plateau (and this is captured in GAP as a separator
constraint).

Overall Engineering Objective

The production platform is sensitive to production

surges, and as such any slugging that develops in
the riser results in consequences that vary from
being undesirable to damaging of production
facilities downstream.

© Text and images contained herein are subject to copyright laws, and should not be reproduced without permission of Petroleum Experts. Petroleum Experts is a registered trademark.
Page 2 January 2015

FLOW ASSURANCE

CASE STUDY USING STEADY STATE & TRANSIENT TOOLS

Traditionally, flow assurance limits were implicitly modelled in steady state simulations by use of constraints. In the new approach
(shown below) the limits are explicitly modelled in a transient simulator, and controlling logic applied in the steady state model.
The traditional approach is outlined below:

 The flow assurance discipline would perform isolated studies on the turndown stability of the production system, and the
propensity to result in riser slugging.
 The study would have to assume a PVT description, and approximations of pressures/rates. These would then be used as
boundary conditions within the transient simulator to calculate the slug volumes and velocities that may damage process
equipment. (Additionally, the transient simulator may not handle the PVT in the same way as the steady state tool).
 The next task would be to take this understanding and apply it to the steady state simulation, and this is usually captured
as a “proxy” constraint in GAP (i.e. a constraint that is not due to the physical limitation of the equipment, but rather an
operational limit).

With the introduction of the LEDAflow data object, the transient simulation is:
 Dynamically triggered depending on the conditions that manifest in the simulation
 The intake pressures, rates and PVT are not approximated, rather they come directly from the steady state simulation, and
the discharge pressures are calculated by the transient simulation.

Modelling Objective
 Detect slugging in the riser
 If slugging occurs, assess the severity based on slug volumes and velocities
 If required, mitigate the slug occurrence by changing the control strategy in the steady state model.

Model Setup
The below model setup was used to capture the response of the
Riser over a prediction forecast. The modules solve in the following
order:

1. GAP was used to perform the forecast (Steady state

multiphase network Optimiser)
2. PROSPER data Objects in RESOLVE were used to evaluate
the occurrence of slugs (Steady state nodal analysis)
3. The results of the data object are used by the Workflow, to
evaluate whether a slug exists or not. If a slug does not
exist, the next timestep is taken.
4. If a slug exists, then data from GAP is passed to LEDAFlow
(Transient simulator) and the simulation run
5. The final piece of the analysis, is to find the average slug
properties for volume and velocity (the transient simulator
solves the riser in time, and so different volumes exist at
different times within a specific section).
6. Feedback Loop: If the response of the system is not
acceptable, an additional control feedback loop can also be
introduced in step 1 and the analysis re-run for this
timestep.

© Text and images contained herein are subject to copyright laws, and should not be reproduced without permission of Petroleum Experts. Petroleum Experts is a registered trademark.
Page 3 January 2015

FLOW ASSURANCE

CASE STUDY USING STEADY STATE & TRANSIENT TOOLS

Steady state Analysis

The results of the workflow (step 5) show

“SlugVelocity” and “SlugVolume” variables,
which are user defined and are reported in
the RESOLVE results. As soon as WGR of
these wells increase, the slugs start to form.

In the below screen shot it can be seen that

after a particular WGR threshold is reached
(blue curve), the slugs immediately form
(green curve).

Transient Analysis (Results imported into

RESOLVE)  
The below graphic is from LEDAFlow, and shows a 3D Plot (time,
pipe length and pressure): it can be seen that slugs develop and
travel up the riser.

Rotating the above graphic makes it easier to interpret:

It can be seen that a slug forms and travels the entire length of
the Riser, and is immediately followed by smaller slugs that
dissipate as the travel the length of the riser.

In this case the slugging starts but appears to dissipate (rather

than grow in strength).

The model was modified to initiate stable slugs, and the transient
response can be seen in the liquid volume fraction overleaf.

© Text and images contained herein are subject to copyright laws, and should not be reproduced without permission of Petroleum Experts. Petroleum Experts is a registered trademark.
Page 4 January 2015

FLOW ASSURANCE

CASE STUDY USING STEADY STATE & TRANSIENT TOOLS

The below plot shows that a

large portion of the riser is
being exposed to slugging,
and that the slugs travel the
entire length of the pipeline.

Looking at the pipeline at one

moment in time (250seconds)
shows that there are more
than one section that have
slug regime manifesting.

Summary

Traditionally, flow assurance

limits were implicitly modelled
in steady state simulations by
use of constraints. In the
above approach the limits are
explicitly modelled in a
transient simulator, and
controlling logic applied in the
steady state model.

The above example of riser

slugging shows how the steady state and transient modelling approaches should be used together to understand the response of
the system.

In this case the transient simulation is fed

with pressures, rates and PVT that are
pertinent to the time in question (this may not
always be the case when modelling efforts are
isolated), and automatically triggered based
upon events in the steady state model.
It is only after evaluating slug volumes and
velocities that decisions can be made to
impose control logic, or indeed modify the
system (e.g. Wellhead chokes) to mitigate the
above response.

A similar approach could be readily applied to

a whole host of flow assurance issues
(Waxing, Asphaltene dropout, hydrate
formation, slugging, etc.) that have historically
had the modelling engineer hopping from one
programme to the next, in an isolated way.

© Text and images contained herein are subject to copyright laws, and should not be reproduced without permission of Petroleum Experts. Petroleum Experts is a registered trademark.