Industry buzz continues to grow around predictive maintenance (PdM), which is seen as a significant contributor to improving efficiency in the oil and gas sector.
The importance of mitigating downtime and reducing maintenance burden is hard to overstate.
While the UK Oil & Gas Authority (OGA) estimates that UKCS reserves of 5.7 billion barrels of oil equivalent could sustain production for another 20 years or more, the continued profitable development of this resource depends on performance optimisation.
The prize is perhaps best exemplified by the Oil & Gas Technology Centre (OGTC) estimate that $7.2 billion of North Sea revenues are lost each year because of low productivity.
It is little surprise then that the OGTC announced a partnership this week with upstream operator Repsol Sinopec Resources UK (RSRUK), software firm Spartan Solutions and the Strathclyde University to develop machine learning software capable of predicting asset failure.
PdM is being implemented to great effect in the renewables, nuclear, aviation and other industries, and it now features heavily in the oil and gas sector’s push for more data analytics as part of the Digital Transformation or Industry 4.0.
While operators already capture and store large volumes of sensor data from rigs, refineries, etc., their track record in making use of this information is patchy.
In part, this is because of a lack of a standardised taxonomy, meaning that data which is collected and stored is inconsistent and very difficult to collate and analyse.
This problem is recognised, and the industry is gradually making moves to comply with ISO14224:2016, which seeks to standardise international taxonomy for the collection of equipment reliability and maintenance data.
The OGTC/RSRUK/Spartan/Strathclyde effort seeks to achieve this from the outset by integrating with Spartan’s PHALANX digital operations platform and maintaining a common language for capturing equipment reliability and maintenance data.
According to John Glen, Spartan’s chief financial officer, PHALANX will be used to integrate with back-office systems to rapidly plan, execute and record the necessary interventions to maximise uptime.
Mr Glen said the combination of PHALANX with the new machine learning software offers a “full-circle solution for maintenance experts: predict, action, report”.
He noted that the captured ISO14224 compliant data will retrain the machine learning predictor to increase accuracy.
Meanwhile, harnessing this fully digital field service data allows for cost savings beyond the headline-grabbing production numbers.
Alluding to the potential cost savings, OGTC’s digital transformation solution centre manager Stephen Ashley said that of the few North Sea and international operators that are already benefitting from data analytics technologies, examples included a 65% reduction in system outages and maintenance savings of more than £1 million per year.
Mr Glen added: “Our approach is to provide a sensor agnostic software-as-a-service PdM application, focused on the critical equipment in the upstream, that is quick and cost-effective to deploy and does not require our customers to hire dedicated data scientists or external equipment experts.”
With data analytics becoming evermore important throughout the full spectrum of industrial sectors, solutions like this could be utilised to predict and avoid faults and failures on a wide range of oil and gas equipment.