You are Home   »   News   »   View Article

Learning from others in analytics

Wednesday, August 2, 2017

Jess Kozman, Singapore regional representative with the Professional Petroleum Data Management (PPDM) Association, presented some ideas about where the oil and gas industry can learn from other industries.

There are many industries which are advanced in their use of analytics, including weather, defence, aeronautics, transport, healthcare, pharmaceuticals, genetics and sport, said Jess Kozman, Singapore regional representative with the Professional Petroleum Data Management (PPDM) Association.

Mr Kozman presented some ideas about where the oil and gas industry might learn from some of these industries, speaking at the Digital Energy Journal conference in KL in October, 'Connecting Subsurface, and Drilling Expertise with Digital Technology.'

Car sharing service Uber has developed an interesting business model, with a proposal for making money from contributing to the Singaporean government's objective to increase fertility rates through demographic analysis, he said.

Through its UberPool 'ride sharing', Uber tries to match people together, who can then share a ride to work. It looks for potential matches based on people's shopping habits and Facebook postings.

Uber has proposed to receive a bonus from the government based on any new Singaporean Citizens which are produced from an Uber-created match.

'Uber would get a bonus for performing that service for the country,' he said. 'So a completely unexpected outcome.'

The lesson for the oil and gas industry might also be that benefits from analytics can come from directions you do not expect, he said.

Another interesting example is power company GE, which provides power equipment for airlines and many oilfield applications, including compressors and pumps.

A few years ago its aerospace division decided to sell 'thrust', rather than the heavy equipment itself - in other words GE takes responsibility for supplying and maintaining the equipment, and it is not actually owned by the operator. This puts the company with the greatest expertise in the equipment in charge of keeping it running.

Perhaps oil and gas service companies could use analytics to package their services in a better way, he said.

Many rail networks are finding ways to use analytics to reduce downtime, something many areas of the oil and gas industry could perhaps learn from.

One German rail operator has built a special railcar for managing data, where all the data about the operations of the train is gathered and processed. This system could be seen as similar to a pigging system for a pipeline, analysing for defects along a long line, he said.

Another user of analytics is parcel delivery service UPS, which tracks fuel usage. The company worked out that in many cases, it would be quicker to make three right turns and drive around a block, than make one left turn which involves waiting for a gap in traffic (this is in the US where vehicles drive on the right).

'The biggest challenge for UPS drivers was getting them to follow something which seemed counter intuitive,' he said. UPS' solution was to offer drivers a bonus if they could follow their own ideas, rather than the computer, and find it led to less fuel consumption. 'The drivers found the computer was right,' he said.

A similar principle has been used by companies running service trucks in US shale operations, with decisions about which route to take.

Another user of analytics was a research company which was gathering very detailed data about fatigue damage on metals. The company often had a problem from losing data, since it was being kept on various hard drives and files in different places.

By moving data to the cloud, at a cost of $1100 per terabyte, it found it was massively reducing the amount of data loss, and that saved $500k a year from not having to redo work, he said.

Sometimes analytics can help show areas which are worth paying special attention to. An example is Continental Airlines which discovered that there was quantifiable benefits to giving its 'Gold' customers special attention.

The company worked out that by having an employee greet the 'Gold' customers on arrival and escort them through the terminal, the person was far more likely to book another flight on Continental and spend money on-board the plane, leading to an additional $500m revenue.

Oil company PETRONAS has been an extensive user of analytics, gathering 50-70 GB a day of real time data, monitoring temperature, pressure and vibration - for its Formula 1 racing cars.

PETRONAS has shared many best practises with real time data operations with its real time operations centre for drilling, Mr Kozman said.

These are all examples of companies finding value from working with real time data in places where it might not be expected, he said.

The oil and gas industry is unique in that it is working with data at many different scales and timescales. 'It makes oil and gas data unique and more challenging. But it also means we share a lot of techniques and best practises,' he said.

An oil company might be analysing data for the past 70 years, which is very different from what an e-commerce company might do. And sometimes the most valuable data is the oldest data not the newest, such as handwritten notes by mud loggers, he said.

However, the bulk of the volume of oil and gas data has been gathered over the past few years, he said.

Sometimes a decision might be taken on the basis of only the most easily available data, rather than the best possible data, leading to an inferior decision, he said. 'We have a propensity to look at data which reflects the way we're looking at operations now,' he said. This leads to what data scientists refer to as 'mode bias' and can influence the result of analytics.

A key factor in making projects successful is that there should be a subject matter expert involved, who can provide a 'sanity check' on any results. You don't just have statistics experts working on the project, he said.

The oil and gas industry is perhaps too susceptible to claims from some software companies, who say 'you just put the data in the system and the data will tell you what to do,' he said.

Associated Companies
comments powered by Disqus


To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.


Latest Edition Oct-Dec 2023
Nov 2023

Download latest and back issues


Learn more about supporting Digital Energy Journal