Working with the largest independent E&P company in the UKCS, we provided data science support to their asset teams to help improve the reliability of the gas compression systems on two of their assets. Here’s how we did it.

The Data

Existing operational data was taken from the gas compression systems on each of the assets. The range of coverage included 16 individual pieces of equipment, 681 tags and 1 million daily data points.

Using this data, our team built a range of predictive models, which allowed us to understand and predict the behaviour and performance of these systems on a continuous basis.

Our Approach

Using bespoke predictive models, our team focused on anticipating failures and identifying process upsets at the earliest possible opportunity, so our customer could:

  • Avoid production losses
  • Reduce man hours
  • Cut maintenance costs
Insights

During the first 12 months of application we helped this customer to achieve:

27

gas compression system trips anticipated and avoided

5%

production loss avoided

60

actionable insights provided by OPEX

Some examples of the types of issues we helped to identify on this system:

  • Early detection of machinery protection instrumentation failure
  • Identification of pressure control valves that were not responding in the expected manner
  • Erratic level control on the Glycol Contactor, due to manual valve loader
A practical example of an insight on this system

Dry gas seal integrity management

OPEX’s data analysts identified that the primary seal gas supply to the booster compressor was cycling erratically.

The onshore support team were informed, who then instructed offshore to clamp the valve in manual until such time that the valve could be repaired during an opportune shutdown.

Undetected, frequent large fluctuations of supply pressure may have caused damage to the dry gas seals.