You are Home   »   News   »   View Article

When senior managers take an interest - EAGE panel discussion

Friday, November 30, 2018

A 'digital transformation' panel discussion at this year's EAGE in Copenhagen showed how everything can change when senior managers take an interest in digital technology.

A 'digital transformation' panel discussion at this year's EAGE (European Association of Geoscientists and Engineers check) in Copenhagen on June 11 showed how everything with digital technology changes, once senior managers take an interest.

The panel included the SVP E&P for North Sea and Russia with Total, the director of geoscience with Repsol, and the EVP technology with Schlumberger. Also a distinguished advisor for seismic imaging with BP and the head of the subsurface functional excellence group with Woodside (Australia).

The reason everything changes is because senior managers are primarily interested (or solely interested) in the organisations' overall performance today. Digital technology is a component of that, but so are people - whether good people can be recruited, motivated and continually developed. Probably, in their minds, people are much more important.

So they can see digital technology in a way which technology and IT people don't - as a tool to support people. They observe that the best business results often come from computers and people working together, maximising the strengths of both.

The aim of business is 'to be the most efficient provider of primary sources of energy, not necessarily be in the forefront of machine intelligence,' BP's John Etgen said. 'If you want to be the most efficient provider of primary energy - you need ways to combine machine and human performance,'

They recognise the importance of making data more widely available to people, rather than locked up in software applications only available to a handful.

They recognise that advanced technology can play a role in making the industry more attractive to the top students, but also that the industry does not have to be a world leader in technologies such as cloud and AI - its primary business is.

They ask if technology is making work more enjoyable, such as helping test out the hypotheses which people develop. They ask about maximise the capabilities of the human brain. They ask how technology can do more to reduce the less interesting work, and they care about how easy the technology is to use.

They ask which vendors of technology are most capable of delivering this, observing that both gigantic cloud system vendors and small oil and gas specialist IT companies have something to offer which the traditional oil and gas large software companies and consultancies don't have. These seem like important questions but are not asked very often in the pages of Digital Energy Journal.

We heard how Total is evaluating where specifically analytics can add value to the business, in a program called 'DAVE' (Data analytics Value Exercise, and also the name of one of Total's asset managers). The program found that the fastest payoff for analytics was in predicting production from wells.

We heard Repsol's director of geoscience and digitalisation saying he sees data adding value to subsurface analysts similar to how it adds value for oncologists interpreting cancer images, where in experiments typically computers get it right 30 per cent of the time, people get it right 70 per cent of the time, and people plus computer get it right 70 per cent of the time.

John Etgen, seismic imaging distinguished advisor to BP, recommended that oil companies should not necessarily just look for the large IT contractors to solve their problems. 'If you only look at big companies you will miss stuff,' he said. 'There is a whole ecosystem [of small IT companies] out there.'

John Etgen, BP

John Etgen, distinguished advisor for seismic imaging with BP, noted that seismic data was the original 'digital' part of the industry, being digital since 1955.

The 'digital' emphasis at BP today has three components - connecting people and data (and making data available to many different people across BP); connecting physical and digital assets (including reservoirs, wells and facilities); and connecting machine intelligence to business decisions.

There are about 50 machine learning experts and data scientists in BP's Houston office (out of a staff of about 5000), and probably their work 'touches' about 2500 people.

Mr Etgen stressed that you should not just go to big IT companies.
'There are things they are knowledgeable and capable in and have a track record,' he said. 'But if you only look at big companies you are going to miss stuff. There's a whole ecosystem out there [with lots of small IT companies]. The question is figuring out who to work with.'

Ashok Belani, Schlumberger

Ashok Belani, EVP Technology, Schlumberger said that until now, most work in oil and gas has been structured around software applications - a large software package which includes database, processing and the user interface.

In the future all of these layers will be separated, with data kept in various storage systems, and made available all the time to different applications. This brings about new capabilities.

The oil industry is still not able to work with all of the data that it has. For example, 'We should be able to work with all the North Sea data, but we don't do that,' he said.

Mr Belani thinks artificial intelligence could enable a computer to copy the way a person does a task. So a person could do 10 per cent of a salt interpretation, and that 10 per cent could be used to train the computer to do the salt interpretation for the rest of the data set.

It could be possible for computers to make decisions if you had a stack of technologies which together can work out all the parameters to make a decision. In this case, computing is an 'enabler' of the whole thing, performing tasks seamless to a user.

But today, an operator is directly pretty much every part of the task. 'In the future system will do the task and serve up to the user.'

Mr Belani said that for technology development, Schlumberger engages with many different companies. ''There is no way to keep innovation in house any more,' he said. 'You have to have an ecosystem.'

Darryl Harris, Woodside

Darryl Harris, chief geophysicist at Woodside (based in Australia check) and head of Woodside's subsurface functional excellence group, said that his company sees digitalisation as a drive to 'collective intelligence', where it can have access to all of its company experience all at once. It means the company can be more data driven, with a 'show me' [the data] rather than a 'tell me' approach to making decisions. 'We could make decisions a lot faster if we had this collective intelligence, make decisions based on data.'

Francisco Artigosa, Repsol

Francisco Ortigosa, Director of Geoscience and digitalization, Repsol, said that the company is keen to use digital technology to help reduce 'time to first oil', and one component of that is the time taken to process data and make decisions around subsurface.

The time taken to process seismic, interpret subsurface data and simulate reservoirs takes a 'significant impact on NPV of the project.'

Repsol also sees that digital technology can help make work more enjoyable, similar to the way that people enjoy a new smart phone.

At the moment, the fragmentation of digital technology means that key decisions, such as to apply for licenses, do exploration drilling, are made on seismic interpretation, reservoir characterisation and so on, are not made using the core data.

Repsol sees cloud technology as a way to improve this. For example, until recently, only 12 people could directly access Repsol's supercomputer for subsurface data. Now, all of its 500 geology / geophysics professionals can access it through the cloud.

Mr Artigosa also believes AI can make a big contribution to reducing 'time to first oil'. He showed an example of a seismic cube 'interpreting itself,' with a computer mapping out the faults and horizons in a few minutes.

Repsol plans to stop using the name IT and call it 'operational technology' instead, emphasising that the technology is support the operations. IT has 'a sort of history in every company,' he said.

It will also employ specialist 'data practitioners', handling everything related to data, including the infrastructure and cybersecurity.

Michael Dorrell, Total

Michael Borrell, SVP E&P North Sea & Russia, Total, said that company has a research budget of $1bn a year, and about 10 per cent of that is specifically spent on digital. That's about a third of its exploration and production research budget.

Mr Dorrell sees digital as all about the interface people have with the 'digital world' - which should ultimately 'make operator more effective, with better safety, having better performance and profit. That's what we are about as an organisation.'

Total puts its digital development in three areas - 'subsurface' (including drilling), 'industrial' including platforms, pipelines, refineries and terminals, and 'work practises' - how people work.

One of the most interesting applications for Mr Dorrell is when digital technology can help 'flatten our organisations', by making data available to all staff.

One example of changing the way people work is the 'smart rooms,' where people from different disciplines can work together collaboratively, he said.

Total recently embarked on a project called 'Data analytics Value Exercise', to work out where data analytics could add value. The acronym 'Dave' was also the name of one of Total's North Sea asset managers.

The most promising use was making predictions of production from 'cyclic wells', he said. The project started three months before the conference (April 2018) and was expected to make a payback by July 2018.

Maersk Drilling [part of Maersk Oil, acquired by Total in __ check] had a long standing relationship with IBM to develop predictive drilling.

Total also has a collaboration with Google to try to generate 'digital assistants' to help its geoscience and geophysics staff, automatically doing some of the number crunching and data sorting.

Total put in place a 2 year plan, but it is turning into a 10 year plan.

Mr Dorrell estimates that geoscientists typically spend about half their time doing repetitive tasks and about half their time doing value adding tasks, and it would prefer if the amount of time on value adding tasks could be increased. 'So it is not just about digital it is also about people,' he said.

The company has a special 'digital officer' in the role of keeping a connection between the operations teams and the digital teams, working out if it is possible to develop solutions in different areas.

Cloud or not?

There were many interesting comments about the decision making process of moving data and software to the cloud.

Schlumberger's Mr Belani sees it as a one way street. 'One way or another, most computer infrastructure will be in the cloud in future,' he said. 'You can take it slower or faster - the faster you move the better of you are. For each company to maintain its infrastructure is a thing of the past.'
In future, 'no-one will own a HPC in their own,'

Also bear in mind that the oil and gas industry's needs for cloud computing today are miniscule compared to the needs of industries such as social media. He estimates that the entire oil and gas industry only needs the equivalent of 0.1 per cent of the computing power which Google and Microsoft have. 'For them, to handle this part for speed and performance is a piece of cake. We can let these industries invest in the cloud infrastructure and feed on it. '

In terms of security, Mr Belani believes that cloud companies will always have much more competence in managing a data centre safely than an oil company. 'It is just like that, the rest of the industry doesn't have a shot at this,' he said. 'Gmail is one of the safest e-mail platforms in the world.'

For standard computing requirements, it will always be cheaper to do it in the cloud than on your own computers. But it may be better to have your own computers if you have a special need, such as full seismic wave conversion on a computer with a ratio of 8 GPUs to 1 CPU.

Woodsides' Mr __ agreed. ''sometimes in oil and gas we get a bit arrogant. The IT companies have been working on something a lot harder than we have. We think we have better security than Google.'

Mr Etgen portrays the cloud vs in-house discussion like the decisions people make about their own toasters or printers. Nobody cares about sharing printers, but when it comes to toasters, it is different. When we want toast, 'you want it right now, exactly as you want it.

Taking the same argument to cloud, if you want to do very specialised research and development, you may find the ability to finely control your computer power as still valuable.

Total's Mr + noted 'our system don't allow WhatsApp - which is one of the most secure messaging systems.'

Mr Repsol noted the cloud 'is significant cheaper.'

Recruitment

The panel was asked what kind of skills and young people they need. 'That's easy, we want people who run towards problems, as simple as that,' BP's Mr Etgen said. 'Jobs are always changing. We want people curious, eager.

'You hire them based on skills which seem commensurate with the task- but that will have to change.'

Also, 'domain expertise is not going away. We'll want people trained classically and all that stuff. We need people who understand how the earth works - that's not going away. People who are domain experts but know how to play in the data science world.'

There could a trend towards even more specialisation, for example instead of being a geophysicist we will have subdomains of geophysics like 'signal processing geophysicist. I don't know how you function with that many specialists. Perhaps people will need to multiple skill sets,' he said.

Repsol's Mr Ortigosa said there is something of a bidding war for data scientists, with stories about some making billions of dollars. 'The people you recruit, they know this,' he said.

Total's Mr Borrell said that the sort of people the industry needs are 'obviously well qualified, but also the right mindset, open, inquisitive, looking for solutions, able to work in a collaborative environment, people who can concentrate on a new problem. We want people passionate about data, because they can find value for us.'

Mr Borrell said that with the world moving so quickly, a key characteristic in recruitment is the ability to embrace change. As a starting point, you want 'People with an education which enables them to fit into our organisation.' But beyond that, you want people who can be 'open, inquisitive, demonstrate they are open to change.'

But oil companies also need people who are 'absolutely passionate about rocks,' like the ones he saw at a recent university talk he gave to geology students. 'Reserves will be found by people not machines,' he said.

Mr Borrell believes it is important that companies like Total are perceived to be in the forefront of technology - in particular for recruitment reasons. The industry was in the forefront in the past, and has 'certainly been caught up' by other industries now.

This view was echoed by Woodside's Mr Harris. 'I think it is important we are leading in terms of attracting the right capability. 'If you don't attract the right p3ople they are going to go to places where they can see clear leadership.'

Future of geoscientists and machines

The panel were also asked what they see as the future role of geoscientists, and whether they expect people to be replaced by machine.

Schlumberger's Mr Belani noted that a great deal of work is already done by machine. For example, machines are used to do the initial processing of seismic data, getting it to a point where a person can start to understand it, perhaps reducing the file sizes by 10,000 times.

It is very difficult for someone to make any understanding of the so-called 'first break' raw seismic data. 'People are going to find reserves that will continue for a long time.'

The question is more whether machines can help make people more 'performant', and perhaps also make their work more enjoyable. For example, a computer can enable a person to generate and test out more hypotheses, and provide better visualisation.

Another future concept is that computers will do most of the work, and people will be in a quality control role. But this is a continuation of something which has also been happening for many years, with people doing quality control of seismic, said BP's Mr Etgen. And also, 'I can't say it's so satisfying as an occupation.'

Computers can be more useful if they present information in a way which is easier for people to work with, or which 'maximises our brain bandwidth,' Mr Etgen said.

'The human's value is understanding what's really going on down there.
Not messing with mice and data. Our carbon based computer is actually really well suited to some of these problems'

Computers should be put to use going through large data volumes. 'No-one ever looks at all the data we make, it's not possible,' he said.

Total's Mr Borrell agreed that for now, and in the medium term, the company sees computers as a tool to help geoscientists to achieve more. 'We are trying to take some of the grunt work out. 'We're about people having time to create value from data.'

Woodside's Mr __ said that a computer could test out a hypothesis, but 'Coming up [with a hypothesis is still in the realm of the geoscientist.'
'Maybe we get to point hypothesis can come from a computer. I can't see that, it may be out there.'

All decisions and work could be described as algorithmic, but the more complex algorithms can only be followed by people, not programed into computers.

One example of people and machines working together could be for people to create different subsurface models (hypotheses) of what the subsurface might look like or how it was formed, and then computers can see which of the models is most likely to match reality. 'We're not at the stage of the computer being able to conceive of a model,' he said.

Repsol's Mr Ortigosa said he had seen a number of studies showing the strengths of computer plus machine, for example showing that a computer can diagnose cancer 30 per cent, the best oncologist, 70 per cent, together 90 per cent. So we should see something similar in oil and gas, empowering people with machines.

Slow pace

The panel was asked why the pace of proceeding to digital is so slow.

BP's John Etgen says that in the high oil price era, there was room for inefficiency and obstruction to change. But now, that has changed.

'The pace of transformation could be a lot faster,' said Schlumberger's Mr Belani. 'There are people trying to protect jobs - pace of change is affected. There's a lot of concerns where they don't need to be'.



Associated Companies
» Digital Energy Journal
comments powered by Disqus

CREATE A FREE MEMBERSHIP

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

FEATURED VIDEO

A modern data platform to improve predictions and company performance
Jane McConnell
from Teradata

DIGITAL ENERGY JOURNAL

Latest Edition Sept-Oct 18
Sep 2018

Download latest and back issues

COMPANIES SUPPORTING ONE OR MORE DIGITAL ENERGY JOURNAL EVENTS INCLUDE

Learn more about supporting Digital Energy Journal