Multiphase corrosion risk analysis

Paul Ellerton

October 3, 2014

Paul Ellerton business unit manager at Apollo discusses how to identify corrosion risk using multiphase flow assurance tools.

CFD on a tubing hanger.

Aging assets have a greater need for more routine maintenance inspections. Offshore inspections, particular those involving subsea pipelines, can be time, resource, and cash consuming.

Transporting unprocessed multiphase fluids from wells or satellite platforms via steel pipelines to the processing platform is common on many developments, but can lead to flow assurance issues.

In a recent maintenance inspection in the North Sea, there was suspicion of corrosion and wax build-up in an oil export pipeline. The 20km-long, 14in. diameter pipeline had been operational for 20 years, with little in the way of physical inspection. Due to its length, a full inspection of the pipeline was impractical, so the operator decided to target inspection at the perceived worst case locations.

Diver inspection was carried out at several low-points, where it was thought that the combination of high temperatures and water held up in the line would cause localized corrosion issues and loss of wall thickness. Ultrasonic wave testing showed almost no loss of material in the line at any locations. Due to the length of pipeline service, almost no corrosion seemed unlikely, which led the operator to question if the inspected locations were actually worst-case locations.

A study was commissioned to conclusively understand whether the pipeline was in a good state of repair or if the perceived likely high-corrosion locations were incorrect. Apollo carried out a staged multiphase simulation program to develop an accurate picture of what had happened over time, what was going to happen in the future and how the pipeline inspection program could be improved.


Fig. 1: Pipeline hold-up predictions using different input frequencies.Images from Apollo.

There are many pipeline corrosion prediction tools available. When used in the right application they can lead to accurate wall-loss predictions. The algorithm-based tools are often conservative and give an overall picture of wall-loss, as opposed to location specific information. Apollo’s technique to corrosion diagnosis is built on a staged flow assurance approach to identify a “region of risk”:

1.    Undertake dynamic hydraulic modeling

a)    Sensitivities to validate method (frequency of data used) and match to operational data
b)    Analyze temperature, liquid hold-up and amount of time spent in region of risk to determine multiple points of interest

2.    Carry out Computational Fluid Dynamic (CFD) modelling to confirm findings at points of interest

3.    Review water and solids build-up and conclude on several likely locations

4.    Recommendations and apply learnings. 

The first stage was carried out using OLGA transient flow assurance software. A steady state approach can be used to give a good indication, but in reality the flow conditions take weeks or months to attain equilibrium – an important factor when trying to understand which areas are at the highest risk. The operator had 15 years’ daily production data, including oil and water rates. A simulation was carried out replicating the conditions in the pipeline throughout its history, including line flushing at known intervals. Each stage of the production history was simulated with the end of a flow period used as the input for the next.

Amongst the many sensitivities required to validate the work was one on the frequency of flow data to be used. This highlighted the importance of using weekly (results used for every three days) as opposed to monthly flow data. Performing the simulations at less frequent rates (monthly) tended to over-predict the amount of water hold-up in the pipeline, in this case. Specific pipeline corrosion prediction tools will base the results on the average results over a period of time, whereas a detailed approach gives specific contact time for the water.

Figure 1 shows the effect on hold-up amounts using input data at different frequencies for a particular location. The analysis which used more frequent data points was more accurately predicting the movement and slugging of liquids.

Left: Fig. 2: Water hold-up against temperature graph at a specific location to highlight the time spent in the corrosion risk region. Right: Fig. 3: Plot highlighting risk regions on a section of the pipeline length.

Corrosion risk regions

Fig. 5: Plot showing the velocity match between OLGA and CFD for the same section of pipeline

To understand corrosion risk regions, the hold-up data needs to be married to temperature data. Corrosion will be worse when the surfaces are wetted during production and at elevated temperatures. During shut-ins, water will gather at low points, which would usually be associated with corrosion. However during a shut-in the pipeline cools to ambient sea conditions and the corrosion risk can be lower. The relationship between hold-up and temperature was derived for each position along the pipeline throughout its production history.

Figure 2 shows the relationship at one particular location, where every dot represents a data point throughout the location’s operational period. For the conditions in this system it was determined that corrosion will be highest at temperatures above 20°C and with water hold-up greater than 10%. At this location, it was shown that it had only been in the risk region for a fifth of its operational life.

An overview showing the amount of time each location spent in the risk region is plotted in Figure 3. From this work a number of points of interest along the pipeline were chosen based on their time in the corrosion region of interest. Further work highlighted that if free water was present (i.e. water does not have to be held up to cause corrosion) then all locations would have a similar corrosion risk.

A transient hydraulic analysis was undertaken using the 1D predictor tool OLGA. CFD analysis was used to understand if there were localized differences caused by momentum of the fluids or turbulence that would alter hold-up and highlight different corrosion risk areas. The CFD analysis conducted matched closely to that seen in the OLGA simulations for both hold-up and velocity predictions. Figure 5 shows the close match between CFD and OLGA in terms of velocity.

One of the reasons for the strong correlation between the CFD and OLGA results was because of the low fluid velocities contained within the pipeline. At higher flow rates you might expect more movement of the fluid phases and, as both approaches differ in the calculations of their interaction, you would start to see a small deviation in result predictions.

If the CFD results had matched poorly with those from OLGA, then the OLGA metrics would have had to be adjusted. The CFD analysis can give more accurate predictions of velocity at point locations, where increases in velocity would reduce the risk of stagnate water and corrosion. The close match in this case demonstrated that the highlighted risk regions were appropriate.

Finally, analysis was undertaken to determine the amount of wax in the pipeline for pigging purposes. Testing had indicated the total amount of reduced volume in the pipeline, but it did not discern whether this was standing water or solids. Although hydraulic software includes wax deposition models, the timing of events offshore during analysis led to a more hands-on approach. An iterative method was undertaken to estimate the amount of standing water by gradually altering the internal diameter of the pipeline. The method matched the results from the subsequent pigging operation with the difference in the amount removed believed to be due to solids left in the pipeline.

Avoiding pitfalls

Flow analysis using CFD.

The UT monitored locations were shown to have covered a good representation of where the highly corroded sections would be, although several other locations were highlighted as potentially having a more corrosive environment. The work highlighted the potential pitfalls in using engineering judgment for inspection purposes. But, although the locations were shown to be a good representation in this case, it could have been quite different. By inspecting the bathymetry of the pipeline, it may seem that there are obvious locations. However, analysis showed that corrosion can happen at other less intuitive places i.e. just offset from the lowest points or where there are multiple low points gathered. There is no recommended obvious place to carry out corrosion inspections.

To ensure relevant inspections are carried out, it is recommended that, prior to any offshore planning, a simulation program similar to the example above is carried out – as standard. Detailed upfront work can pin-point where the potential issues are in order to reduce mobilization time and costs.

It should always be understood that no simulation can fully represent a real inspection. However, staged encompassing analysis can take weeks to perform and give great insight into not only what has happened, but also why and so in turn can be used as a preventative method for reducing issues in the future. By viewing physical inspection results in light of the simulation program, the prediction methods can only become more accurate, which in turn will bear cost benefits throughout the industry.

Click Here to participate in our poll: What’s the biggest challenge related to pipelines and flowlines?  

Paul Ellerton
, business unit manager at Apollo, is a chartered engineer with over 10 years’ experience in the energy industry.  His specialization in flow assurance has been used in advanced analysis to develop practical engineering solutions globally for subsea and topsides applications. Ellerton has a master’s degree in CFD analysis.