Case Study: Pigging Operations
Three Little Pigs - Field Applications to Improve Efficiency
This Little Piggy...
The ‘Three Little Pigs’ is an English fable, published in 1890, about three pigs who were sent off into the world to build their fortune, with each building their house out of different materials. Each took a different - shall we say “risked-based approach” - to managing their investment. In oil/gas operations, we rely on different risk mitigation tools to manage inventory in our pipelines, often in areas that we cannot accurately access, monitor, or remediate without some help. Sometimes, we need to look far down the rabbit hole (pipeline) to see the truth.
Enter a tool known as “Pipeline Inspection Gauge”…or PIG (a convenient acronym, although not entirely sure that’s the actual name origin).
Pigging is common across the oil/gas industry to manage issues such as:
Commissioning activities (i.e. dewatering)
Inspection / surveillance (i.e. integrity management
Liquid inventory control
Paraffin (or solid / debris) removal
The tools must be made to fit, either tightly or with some flexibility around pipe bends and any small diameter changes within the network. In some cases, the better choice is a highly-viscosity ‘gel pig’ that could even be made sacrificial with enough applied shear, to prevent pigs from getting permanently stuck. Pigging technology has become advanced enough to enable pigging of multi-diameter pipelines, with the appropriate design considerations.
And in some cases, we need to condition the line before pigging. If there is risk of debris / solids that may place a large burden on the mechanical removal capability of the pig, a high-rate volume flush to remove bulk solids may be an effective way to limit the risk of blockage when the pig comes through. A pigging appetizer, if you will (i.e. Outback’s “Rusted Onion”)
The purpose of this discussion is to evaluate three different pigging case studies, all with a focus on operability impacts and how those were modeled at the engineering stage. By having a system-wide view of the operational risks before pigging, it can reduce some of the uncertainty and provide good KPIs to monitor during the actual operation. Knowing what to look for and how the system is expected behave can highlight any obvious red flags. Better to see those early.
Case Study #1: Integrated Pipeline / Process
In this first example, we consider a long distance (offshore) pipeline designed primarily for gas service. However, an opportunity to convert to multiphase service by transporting condensate may enable favorable project economics to monetize the liquids. As such, a procedure to ensure that there are no adverse operational impacts must be adopted.
The primary concerns for conversation to multiphase operation are:
Increased back-pressure due to hydrostatic head of liquid
Onshore liquids handling constraints for a system not designed for liquids service
Due to terrain challenges, liquids will accumulate in low spots, be pushed out by the pig, and quickly re-establish their equilibrium holdup level.
In some cases, the frequency of the pigging can be adopted to control the accumulated liquids. In the image below (LEFT), we see liquids increasing (post pigging run) steadily for a period of 7+ days. In this case, the pigging frequency may be optimized to reduce the net-recovered liquid volumes by pigging more frequently. However, in the image on the RIGHT, equilibrium for a different operating mode is reached within 1.5 days. As such, pigging >2 days will result in the same answer. Operations has the ability to stretch their pigging frequency as a result, unless they are committed to launching every 1 day…
Specific to this case, the liquid inventory in the pipeline was in far excess of the onshore liquid handling constraints. We did consider bypass (see next section) for reducing volumes, but this did not bring received liquids down to an acceptable level.
Constraints:
Operationally, we believed that we could “throttle back” at the plant inlet to control the liquids, but there were several constraints we were needed to optimize:
Liquid slug size:
With an existing vessel, there is a finite liquid volume which can be ‘stored’ in the inlet separator / slug catcher
Liquid drain rate:
Downstream condensate stabilization, as well as existing valve/orifice sizes, limited how fast we could dump and process liquids from the separator
Gas starvation:
If we receive too large of a liquid slug, there is a risk of gas starvation into the compressor while processing liquids, resulting in the plant operating in recycle
Back-pressure (pipeline):
If we throttle back too much to control the inlet liquid arrival rates, this will increase back-pressure on the pipeline, potentially putting the operating pressure above the offshore compressor discharge pressure limits
Temperature / Hydrates:
By holding too much back-pressure, we can induce a large dP at the facility, resulting in a dT (expansion cooling via Joule-Thomson effect) that could lead to hydrates and/or freezing. While chemical injection may be available, the potential for contamination (MeOH) of the condensate must also be considered
Within the onshore plant, there were several options for liquid level (LCV) and pressure (PCV) control.
This could be done manually with a series of trial-and-error, by manipulating the inlet control valves. Initial screening work indicated that the pig-induced liquid slug would flood the slug catcher without any inlet control, as well as with only minimal (“5%”) control. However, a more aggressive scheme (“1%”) was able to control liquid level to acceptable values, but delay the full process by ~12 hours as the liquids were backed-up in the pipeline and slowly allowed in.
As the incoming liquids were constrained, the pig slowed down (stalled) and the back-pressure increased.
We could have cycled through the pipeline / process models until we iterated on a solution. However, we opted to build an integrated / dynamic simulator, coupling the pipeline (OLGA) and plant (Hysys), facilitated by Billington Process Technologies (BPT) to integrate the models. By passing the incoming multiphase flow behavior from the pipeline directly to the dynamic process model, the control logic can be implemented to manage liquids, whilst passing back the impact on the pipeline for any process control.
NOTE: This is similar to a previous case study, where we also combined the pipeline and facility process control (see below). Same general idea, different optimization routine.
Outcome:
By coupling the pipeline to the facility, we were able to build a robust operating scenario that met all individual constraints. Not only was a suitable pigging philosophy developed, the integrated model reduced the total pigging time by optimizing the constraints to find a ‘shortest duration’ solution. The development of the integrated model serves as a ‘lite digital twin / operator training simulator’ that enables operations to continue to update the pigging program with changes in incoming liquid volumes, pigging performance, or any other changes at the plant. An ‘evergreen’ pigging program that can be kept updated, without a manually labor-intensive process - or worse - a static, paper/hard copy technical note with a few stale graphs.
Case Study #2: To Bypass or not To Bypass
In this example, we evaluate the need to pig for paraffin (wax) removal of an offshore (deepwater) asset. For pipelines that operate without direct physical access (i.e. hot taps, hot oiling, and/or coiled tubing access), pigging is one many tools that are considered for wax removal. Aside from downtime impact to production, the system needs to be made “piggable”, either with remote launchers or via looped pipelines. And, what if we launch the pig and it gets stuck? After all, we are trying to remove an (often unknown) quantity of hard, sticky material that could contain a number of “nasties” that make removal challenging. Sometimes, the side effects (pig stuck) may be worse than the disease (increased pipeline dP).
Enter bypass pigging. The goal of designing a pig for a solids removal program is to ensure that we do not concentrate the removed solids (paraffin) ahead of the pig, which risks plugging or creating a large volume of highly viscous solids that will wreck havoc on the receiving facilities. As such, a pigging bypass can be designed to moderate the fluid behavior ahead of the pig.
To avoid a wax accumulation ahead of the pig, the rate of bypass flow from the pig must be greater than the rate of wax approaching the pig.
Bypass sizing
The pig design can therefore be based on the following two criteria (A O’ Donoghue, Pipeline Research Ltd):
Bypass flow through the pig:
Rate at which wax approaches the pig (area of max deposit)
By setting the bypass rate to be equal to (or greater) than the wax accumulation rate, both an appropriately-sized bypass (A, %) and a pigging frequency (N, days) can be determined.
In this case, we modeled the potential wax build-up of our pipeline using the OLGA Wax Deposition Module, as well as the Pigging Module. It should be noted that this is a multiphase production system, operating in highly turbulent flow, which complicates the accuracy of the wax modeling results. However, the paraffin behavior was fully characterized with available high-temperature gas chromatography of the liquids (HTGC), as well as wax appearance temperature and wax precipitation curve estimates via Differential Scanning Calorimetry (DSC), both provided by SPL.
Challenges:
Several challenges existed for accurate model predictions, which included:
Accuracy of wax deposition predictions
Fundamental to some of the pigging estimates (and pig sizing) is an expectation that we will accurately know the wax deposition behavior.
Wax model / pigging model combination
Numerically, there is “a lot going on” and an already-complicated phenomenon (wax deposition) is further complicated by adding the pigging numerics to the simulation
Accuracy of slurry rheology
One of the key variables here was ensuring adequate pressure to drive the pig through the round-trip flowline system. Understanding the fluid behavior during pigging, particularly ahead of the pig when the removed wax slurry is concentrated (but mobilized via bypass) represents a non-trivial dP component that must be accurately assessed.
Stalling vs. increased back-pressure vs. facility handling
Too large of a bypass and we risk stalling the pig by allowing too much flow through. Too small of a bypass and we risk sticking the pig behind a wall of viscous fluids. And all of this must fit within facility handling constraints (pig receiver/trap, inlet separator, etc.). Very quickly, our “Three Little Pigs” fairy tale turns into a “Three Little Bears (Goldilocks)” fairy tale of needing to get the pig design just right.
In this case, pigging was evaluated after various increments of time (up to ~1 year) to quantify the impact on delaying pigging operations. It is no surprise that longer intervals resulted in higher pressure requirements due to the incremental paraffin removal required. However, it is interesting that the results are not linear in this case - the impact of delaying from 30-60 days is similar to the impact from 60 - 365 days. As such, within this bounds, operations can work within the available pressure constrains (and facility wax handling requirements) to design the pigging program.
Interesting to note that the reported liquid phase viscosity ahead of the pig (paraffin deposit removal + liquids) reached a maximum value of >10,000 cP in the model, which is no doubt contributing to the back-pressure requirements. So again, accurate understanding of the fluid rheology coupled with pig design are directly tied to the pressure required.
Also interesting to note that the bypass rate selected here was 8-10%. A lower bypass rate was - as expected - leading to the pig becoming ‘stuck’ due to extremely high incremental back-pressures. A higher by-pass was leading to the pig also becoming ‘stuck’ due to insufficient motive force behind the pig, not to mention a very poor wax removal efficiency.
The optimal pigging program is likely to rely on returns from the initial pigging runs to revamp and adjust throughout operations. Combination of back-pressure monitoring, facility wax returns, pig data logger diagnostics, fluid sampling, and wax deposition modeling are all important factors in adjusting subsequent pigging runs. The goal(s) should be to minimize the adverse effects to operational efficiency (uptime) and production deferrals (back-pressure)…whilst not risking a stuck pig. A delicate balance. As Mr. Miyagi said in Karate Kid, “Better learn balance….Balance is key…Balance good, karate good. Everything good. Balance bad, better pack up, go home”
Case Study #3: Onshore Perspective
In our final example, we shift to the world of onshore pigging. Here, the reasons for pigging are generally the same - liquid or solids removal - but the stakes can be slightly lower (it’s easier to deal with a stuck pig when we can have better access). However, in return for this reduced risk, we sometimes aim to squeeze the last bit of optimization from each pigging run, as these can be a constraint on available manpower. This presents its own challenges, given the size of the pipeline networks generally involved. Plus, relative to offshore where every care is taken to have a viable pigging path, we may encounter more ‘unpiggable’ areas in the network .
For this project, pigging was performed to reduce liquids in the network, helping to minimize back pressure on wells. Interestingly, separation (gas / liquids) occurs at the respective wellheads prior to entering the pipeline network. In most cases, the gas is sent into the network, while the liquids are trucked away. However, for a few select wells, the liquids are also produced into the system in a “dual-phase” manner. This results in substantial costs savings vs. trucking the liquids, as well as offering a host of environmental benefits (i.e. reduced greenhouse gas emissions). Additionally, fewer truck-miles has lower potential HSE impacts due to the inherent danger of road driving.
The challenges of pigging this network are summarized as:
Handling liquids received:
Like example 1, this system is designed to receive gas with only small amounts of liquids. The introduction of higher liquids in the system, through dual-phase operation, must be balanced with what can be handled at the receiving point when pigging. Increasing receiving facility costs may outweigh the other potential benefits.
Timing:
Related to the challenge above, how do you pig effectively so that you are ideally pushing liquids from further upstream areas into downstream sections so that you don’t end up replacing pigged lines with liquids from pipelines immediately upstream? This must be balanced with the liquids handling issues mentioned directly above. After all, timing is everything…
Fluids uncertainty:
Unlike an offshore system with fewer wells and a very good grasp on the fluids in the pipeline, for large network systems with a high number of wells, approximations must be made in the representative fluids in the system to achieve reasonable model run times. These uncertainties in the fluids can lead to more or less liquids condensing than there are in reality. Plus, we are more exposed to weather variations which can have a significant impact on liquid drop-out and, in turn, the required pigging operations may very well be ‘seasonal’.
Automation:
The level of automation has a direct impact on pigging procedures. If we are willing to sit at the board and bring every pig in ‘manually’, we can get to a finely-tuned level of control (LEFT). However, this requires more attention/manpower. Alternatively, we can opt for a more passive control (RIGHT), which is less labor-intensive…but comes with operational risks (longer durations and higher pressures). Again, finding the right balance is critical since both alternatives have their challenges.
Outcome:
By simulating different scenarios of starting pigs in various upstream locations, we can track liquids as they flow through the system into downstream pipelines. Running these simulations with upper and lower sensitivities on potential liquid volumes provides insight into the implications of varying liquid amounts on handling abilities.
Ultimately, we screened scenarios to identify successful pigging operations that align with existing equipment capabilities, while eliminating problematic options. In the end, we found multiple pigging scenarios that work and eliminated many others. This offers flexibility to the operations team that is directional, but not overly prescriptive. Because, as we all know, not everything always goes to plan. And we may miss a run here or there…so how do we recover?
Future Study: Machine Learning
The pigging case studies here show a variety of challenges from a modeling and operational perspective. And, how to determine when those “squiggly lines” are behaving the way they should or truly indicating a problem. Pigging will induce stress on the system and we get transient responses that are not typical, but may not be abnormal. We need to understand what is likely to occur and the associated physics / dynamics that go along with it. As the “Three Little Pigs” fable tells us, “only a person who builds a solid base can face such hazards”.
And while the last case study starts to hint at additional complexity in pigging approaches for multi-pipeline networks, these were all still largely single pig scenarios (in = out) for the most part. Pretty straight-forward and easy to predict (when we take the time to evaluate). But what if we have multiple pigs in the pipeline network at the same time? More specifically, what about up to 20? And, what is those pigs have unknown bypass rates (liquid slippage) and - potentially - unknown launch/receive times? Its often hard to get it right when we know everything…but even harder when things become unpredictable / unreliable.
Stay tuned for an upcoming case study where we stretch the limits of traditional pipeline models, leveraged with AI and machine learning, to attempt optimizing a pigging program across a complex multiphase network in West Texas. Can machine learning help make sense of sporadic data, potentially helping solve the age-old “garbage in / garbage out” risk that plagues most traditional modeling approaches? Make sure to Subscribe to stay tuned for when that case study is released.
How come no reference to the Pig man !? :)