Tom Moroney, the Deepwater Technology Deployment Manager at Shell Oil, gave a presentation on Shell Oil’s approach to control towers, Big Data, and analytics at ARC’s Forum in February. They don’t call it a “control tower” – for Shell it is an “Integrated Operational Center.” And it is not a supply chain control tower – it is focused on asset and production management. Nonetheless, their approach offers real insights into what is possible in terms of generating big data, and then analyzing, visualizing and contextualizing that data, and ultimately driving value.
Their control tower in New Orleans monitors and helps to optimize their deepwater drilling operations in the Gulf of Mexico and off the coast of Brazil. They are monitoring 200 wells connected to 9 floating and 4 fixed platforms. These wells produce several hundred thousand barrels of oil and several hundred million feet of natural gas per day.
Shell Oil has a large number of sensors out in the Gulf of Mexico. Over 17,000 instruments over all equipment types are being evaluated, generating data on vibration, pressure, temperature, flow rates, and other things as well. This generates massive amounts of data. Each instrumented piece of equipment is generating anywhere between 10 and 10,000 pieces of data. This, in aggregate, means Shell is consuming 430 million data points per day from which their analytics’ engines do an additional 310,000 calculations.
To optimize production, promote safety, and avoid environmental issues the framework for the predictive analytics needs to be approached intelligently. When new wells are drilled or platforms built, how should the equipment be instrumented? What data is needed and how does that data need to be analyzed? What actions will be taken based on the analysis? These things all need to be thought through.
At Shell, answering these questions meant a disciplined approach to how things like alarms, alerts, and events are defined:
- An Alarm is an operational notification that requires immediate and urgent attention. These are focused on the safe control of equipment.
- An Alert is an engineering notification that indicates a parameter is trending outside of desired range. This indicates that threatening situation may be developing.
- An Event is a single or combination of ALERTS that indicate a defined production anomaly.
- A Service is the defined set of steps (standard operating proceedures) to ‘treat’ an event.
Together, alarms-alerts-events-services create a framework of the 20 deepwater issues that most threaten equipment performance and thus production. These include things like compressor reliability, the rate and phase behavior of the well, or gas lift performance. For each of these critical performance areas, there are defined data elements – simple calculations or even analysis models.
If an alarm needs to be validated or a service performed, the system has not only role-based collaborative workflows, but the information the engineer needs to help resolve the situation in easy and intuitive drill down paths; the necessary information could range from immediate access to a well’s test performance, and equipment manual, to a well’s flow history.
To close the loop, the root cause of the problem and resolution actions need to be recorded. In some cases the analysis will suggest a new service best practice that needs to be documented and built into new workflows.
This advanced set of capabilities has been in operation since 2010. Shell Oil received a payback in 6 months and since then have had a fourfold payback their original investment.