Unlocking the collective intelligence of Woodside Energy with IBM Watson
- Summary:
- Woodside Energy is transforming itself into a 'data driven' business. Its plans are ambitious but they are not without significant technical and cultural challenges.
Kalms, the Data Science Manager of the company's Perth based FutureLabs, described how the innovation group has three main objectives within the company; to build and deliver projects based on Woodside's data assets along with developing an internal data driven culture.
The vision in data science is to unlock the collective intelligence of the entire organisation past and present.
Over the first 18 months FutureLabs has grown from four staff to the equivalent of twenty full time employees with a mix of contractors, graduates and staffers. With such a small team, outside partners become essential.
We also partner with IBM Watson with our cognitive computing, and then with our analytics we work with Accenture, Optika and some niche players. Our infrastructure is on AWS. The collaboration is really important, it helps with our innovation.
As with other innovation and digital transformation advocates, Woodside's FutureLabs team looked for an early win to establish their bonafides within the organisation.
One of the things we wanted to do was introduce plant wide analytics, we started with a small part of the plant. If we can test a hypothesis then we can scale.
FutureLabs had the opportunity to test this at one of the companies Northern Australian gas processing terminals where part of the plant was creating a production bottleneck.
The operators knew if they had additional sensors along the length of the distillation column they could relieve the bottleneck.
Traditionally the engineer would do the scoping, they would work out how much it would cost to do the hard wiring and the labour and penetration into the vessel which would require a shutdown and shutdowns in our business are big, big dollars. Often it becomes too hard or expensive so we just leave things the way things are.
So we put magnetic temperature sensors along the column, that gave our operators enough information and confidence to alleviate the capacity bottleneck. It was small percentages but added tangible value – small percentages in our business are millions of dollars.
In Kalm's view the analytics side is relatively easy, a bigger task is applying cognitive computing to the company's deep sea exploration and drilling operations where geological information is analysed to anticipate potential problems for the platforms.
It increases our ability to avoid drilling 'events', it reduces cost and reduces risk. Reducing risk in our industry is important.
Watson allows us to process massive amounts of information and our people are better employed in interpreting it.
Instead of spending eighty percent of our time reviewing information we're now spending eighty percent reviewing information. We're in that smart quadrant and that's really powerful.
The FutureLabs project has also been working on its Willow project that creates an intuitively searchable knowledge base of the company's documentation and ties together a number of the organisation's other systems for easy access.
A constant challenge in all of these projects is making sure the information going into the system is suitable.
A lot of people think you can pay for Watson, flick a switch and save a lot of money.
We've had instances in what I'd call the data curation in getting the analytics right – it takes significantly more time than what people think. When we're doing our development now we have to manage our expectations of the data.
We're only just realising that data is an asset. The quality of data has been a challenge, it's really important to look after it.
Despite the successes on a small scale, Kalms still sees a long way to go in changing the company's mindset.
You don't change culture overnight.
One of the key challenges in an oil and gas company is building the confidence and trust of operations teams. Data scientists and Engineers approach analytics with very different mindsets and the latter demanding high levels of proof – something understandable given the potentially disastrous consequences of a mishap on a drilling rig or processing plant.
Despite the scepticism from engineers, management's support has been one of a key driver of the company's efforts to change its culture says Kalms
One of the big factors of success we've had has been the CEO's backing.
So in some respects, oil and gas companies aren't too different from other organisations in needing senior management to drive change.
My Take
That the oil and gas industry is extending its adoption of data analytics shouldn't be surprising. The sector requires huge amounts of capital and is high risk from a both a financial and environmental perspective. Understanding as much about an exploration field or production processes as possible is an obvious management and engineering priority.
Woodside's FutureLabs shows us the direction of the oil, gas and mining industries as engineers and business analysts get access to more data and better tools. As these businesses become more automated, increasingly work is going to be done in a comfortable Perth office rather than in the often hostile locations where the drilling rigs and processing plants are found.
The key challenge for these 'data driven' organisations is going to be ensuring the quality of information coming in provides the right answers. IBM Watson is part of that equation. That is something Shelley Kalms highlighted and which has been an issue across every industry that pursues this goal. Ensuring data lakes aren't poisoned is as much a challenge for companies that have previously been more focused on not polluting the physical world.
The main lesson though from Kalms' story is the importance of management support. In almost every story of successful or failed transformation projects, the key factor has been executive and board commitment to change. That's a constant across all sectors.