PointCross

Smart Oilfield Part I: The Principles Behind ‘Smart’ Technologies

Nothing can demonstrate the possibilities and the value of modern day sensing, telemetry, and cascading close loop controls as well as the use of remotely controlled drones (aka Remotely Piloted Vehicles) for surveillance and air to ground attacks.

Flown from across the world by pilots operating from a cockpit simulator with visuals representing the images being captured by the drone, the pilots control the aircraft using the visual and sensor cues. The pilot’s experience is almost the same as being in the aircraft except that the pilot is not in any physical danger. Modern sensor and control technology combine digital sensors for position, attitude, altitude, pitch, actual visuals, infra-red and electronic surveillance to create a fusion of data back at the pilot station.

One might be tempted to think that all one has to do is to present the information to the pilot and provide the pilot with controls that drive the actual flight controls in the drone and all is well. Not quite. The tough problem is that it takes a certain amount of time for the sensor information to reach the pilot station on the other side of the world, and the controls applied by the pilot take as much time to get back to the drone.

Imagine driving a car at 60 miles an hour, when a deer suddenly crosses your path 50 feet ahead.  There’s a 1/4^th second delay as your mind registers the need to act; there’s another 1/4^th of a second for your brain to send a signal to your foot; and one more 1/4^th of a second for the car to apply the brake to the wheels after you’ve hit the brake pedals. That represents a total of 66 feet traveled before the car even begins to slow  down  – and by then, it’s already too late. The reaction time of such critical, real-time decisions must be faster than the action to which it’s responding. The same is true for drones; by the time the pilot tries to reverse the bank of the drone – caused by a gust of wind – the drone may have already banked a lot more than it appeared to the pilot. The pilot ends up initially under correcting, then over correcting and quite soon the drone is uncontrollable.

But we know that drones fly well and very reliably. That is because the short term, quick reacting responses needed to keep the drone flying is handled by a smart controller or computer in the drone itself. A gust banks the aircraft to the left – no problem – the computer instantly commands the left wing aileron down and the right wing aileron up to correct the bank. The pilot does not need to worry about these kinds of low-level technicalities, leaving her free to focus on decisions that require her human judgment, like which heading to fly and what altitude to maintain.

Meanwhile, the mission director is making an even higher level of decision; he’s interested in completing the mission – whether it is surveillance or finding and destroying a designated target. He uses the same current location data and other sensor data from the remote drone to decide the safest route to the target, the best way to manage the total flight to it and back to the base.

Notice that there are three control loops at work here – a control loop being loosely defined as a combination of “observe-decide-act” by a human, automatic control system, or hybrid of both. The fast-responding controls within the aircraft; the tactical controls with the pilot in the loop; and the longer strategic controls with the mission director in the loop.  All three loops work in a cascading manner one above the other, and all three use the same sensor data. But each one samples the data differently and they use different logic to meet their specific ends. Together they execute the mission at each level smoothly, seamlessly, each control or decision loop extracting the kind of information it needs from the very same single source of sensory data.

This is not a unique case of cascading control loops. Every modern refinery, most modern automobiles, fly-by-wire aircraft like the Airbus or F-16, our human nervous systems, and many social organizations such as well run companies are all examples of such cascading control loops. In each case there are, broadly, three levels of control loops each one nested within the other. The innermost loop takes care of the fast reacting, operational responses. The middle layer samples the same sensory data and applies a higher level of smart feedback controls usually by inserting a human being into the control loop so that tactically significant decisions can be made. Typically these are medium term decision in terms or time and response. Outermost are the strategic decisions made with long range understanding of the sensory data, and carefully contemplated long term decisions whose impact will be also be felt in the longer term.  For the past decade or more there has now been a concerted effort to apply these concepts to the oil field for a number of reasons. 

Stay tuned for our next post where we will explore cascading control loops in E&P companies, and how ‘Smart’ Oilfields should apply these concepts.