The group I’m working with is responsible for defining strategy, principles, practices and re-usable components for upstream data integration projects globally. Currently I spend half my time in the UK and half in California.
It’s not as glamorous as it sounds. Nice as it would be to have an oilfield on the outskirts of San Francisco, unfortunately its in the middle of the Californian desert, close to Death Valley and just far enough away from Vegas, San Fran or LA to make them weekend-only destinations. And on top of that, it seems I’m flying myself to an early grave!
In the UK I spend my time on delivering stuff for the global program that I work on. When I’m in California its working on the latest project to adopt the SoBI approach and principles.
If you’ve read the first two blog posts and managed to wade through the SoBI white paper, you may be wondering what we have been doing in the 12 months since it was conceived, written and published. The short answer is implementing it.
A longer answer is that this is a big project – the project in Aberdeen is approaching two years since conception with around 25 people staffing it at peak times. The architecture is new, and fundamentally delivering an architecture is in itself quite a challenge in Upstream O&G. Upstream is the part of the Oil business that delivers the dollars. Basically, the more hydrocarbons that can be extracted, the more revenue the company makes. Historically this has led to quite a tactical approach to integration – the relative wealth of Oil & Gas companies ensures that best of breed applications are bought, but they are implemented to solve specific business problems, often without a holistic view on data integration between applications (“Give me the shiny things NOW!!”). It also means that in comparison to companies in other verticals I've encountered, upstream O&G has lots of applications out there, many of them storing similar data and accessing data from other applications in a very tactical, point-to-point manner.
Upstream Oil and Gas is also a very complex environment in which to work. In almost 10 years of working in the area of data warehousing and business intelligence, I’ve never come across a combination of:
an environment so littered with candidate data sources
systems implemented so myopically
with such a reliance on Excel as a data store
supporting such a complex business
Add to that a new architecture and throw in agile as a delivery mechanism and you have a significant challenge.
"Upstream" refers to the process of exploration and production of oil up to the point where it is transferred on for sale.
I’m surprised by the number of people who think that producing oil involves drilling a straight hole down to a conveniently located pool of oil and catching it as pressure forces it to the surface, before selling it on at huge profit.
Firstly, there is no magical pool of oil. What you have, are layers of sedimentary rock into which the oil flows and sits, along with water and gas and lots of other chemicals in pores between the molecules of rock. Think of it as the process of sucking water out of a sponge. The amount of stuff that the rock can hold is its porosity. The ease at which the stuff can flow through the rock is its permeability.
Skipping over how a potential oil reservoir is identified, the next thing to do is to drill. Depending on where you are in the world, this can be an expensive exercise. In the Californian desert where the oil reservoir isn’t too far below the surface, things aren’t too bad, but if you are working in deepwater exploration, things are significantly more complex and expensive. Take a look at the Jack #2 well announced by Chevron in deepwater Gulf of Mexico where the well was drilled to a depth of 28,000 feet. To put that in perspective, Mount Everest is 29,028 feet high. This is space age technology, and it doesn’t come cheap. Cost estimates for field development in the vicinity of Jack #2 are about $80-120 million per well drilled, with an additional $1.3 to $1.5 billion for subsea facilities. Upstream Oil and Gas is very much a risk/reward business and we tend only to hear about the rewards (which admittedly can be substantial).
It’s also surprising to find out that wells don’t have to be sunk vertically. The drills are directional so it’s possible for a well to take a circuitous route to the oil bearing formation. In some circumstances wells are drilled horizontally – in this case the well starts off vertical but curves as it goes down eventually flattening out to a horizontal line. It’s also possible that each surface well could have a number of wellbores within it – i.e. multiple paths leading off from the main well to different parts of a reservoir. Each wellbore also needs to be completed – completing a well consists of a number of steps;
The process of completion usually involves blowing holes in the casing at the bottom of the wellbore to expose it to the oil bearing rock. Each wellbore can be completed multiple times and you may wish to complete the well such that hydrocarbons from two or more formations may be produced simultaneously, without mixing with each other.
So what’s left? Stand at the top of the well with a barrel and collect the oil as it flows to the surface? If you’re lucky, the pressure in the reservoir will force fluid to the surface. Usually however this isn't the case, and even when oil does flow under its own pressure, this reduces over time as the fluid is extracted, reducing the pressure in the reservoir. This is your bottomhole pressure and it’s something you have to keep an eye on. It’s checked regularly by performing a well test which gives an indication of the pressures and temperatures at the bottom of the wellbore and gives an indication of the theoretical production rate of the well.
A good reference for O&G buzzwords is the Schlumberger Oilfield Glossary.
I think that’s enough for one post. Next time I’ll cover all the stuff that happens once the fluid is actually out of the ground – and what on earth any of this has to do with data integration.
Some news to finish. The head of the Technology Team, Iyas AlQasem, has just informed us that Conchango will end 2006 larger than it’s ever been. I presume this relates to number of employees now in the company rather than the average weight of its consultants, though it should be noted that with so many developers working in Egham during the day, the man who runs the donut shop is now driving round in a Bentley Continental.