R&D Examples

R&D EXAMPLE

As an example, corrosion activity can be monitored using noise spectra as shown in the illustration below.

The output allows for diagnostic analysis and preventative maintenance to be undertaken before critical component failure.

DC and low frequency corrosion monitoring

• Real time corrosion monitoring: ability to monitor whole system for corrosion activity including mixed metal systems.

• Traditional DC monitoring with additional low frequency to enhance detection of galvanic corrosion between dissimilar metals.

• Stainless and mild steel have clear spectral differences even when they have similar DC potentials and could be considered passivated.

• More testing / field recording is required but indications are positive.

• Implementation will be straightforward.

R&D EXAMPLE

Sand deposits in separator vessels can displace significant liquid volumes. This consequently increases the total mass of the system.

By monitoring natural (and potentially imparting low energy forced vibrations) it is possible to estimate the change in mass of a large vessel and therefore give early warning of sand build-up - external equipment only, not required to drain or modify the vessel.

Timelapse shows scale model with 0 to 10kg of sand added

2nd trace: resonance in vertical direction reduces frequency as sand mass increases

Retrofittable vibration monitoring

• Low power vibration sensors with AI capability available

• Packaging for Ex zone 1 environment required

• Next step to deploy test hardware to flow loop or real location to gather data

• Could be applied to flow lines for wax / scale deposition and possibly gas breakthrough

R&D EXAMPLE

An existing dial pressure gauge needs to be closely monitored for changes prior to a well intervention

A WellCam can be temporarily installed to monitor and upload data from the remote location

Automatic image recognition

• Camera can be configured / trained to read the gauge and upload the result – lower and upper pressures and confidence level – lowest bandwidth/no human intervention

• Image taken and processed every minute, update to server configurable or if a significant change is detected

• Images stored locally and can be sent if bandwidth allows for human verification