In this article, Sean Unrau, Product Line Manager, reveals the way Pason is helping reduce alarm fatigue.
Rig crews work in a stressful, loud environment where they endure a barrage of sensory input. Moving machinery, connections being made; instrumentation and sensors of every kind displaying values and trace graphs; alarms sounding.
In this environment detecting dangerous and expensive events such as kicks, mud losses, and spills can be challenging for crews. So one issue designers of detection systems need to consider is alarm fatigue.
Traditionally, systems are designed to alert the crew when mud levels and flow rates move outside a defined range, which can be a sign of lost circulation or impending kicks. However, mud levels and flow rates fluctuate widely during normal drilling operations, so even a correctly calibrated and functioning system will generate many alarms that do not signify lost circulation or kicks.
This fosters an environment where drillers can become desensitized to alarms. They might cope with this alarm fatigue by decreasing the sensitivity of alarms, which decreases the effectiveness of the detection system. Or they may turn off the alarms, either temporarily or altogether. This is like taking off a seat-belt because it rubs the shoulders.
In developing the Enhanced Pit Volume Totalizer (ePVT) Event Detection system, Pason has grappled with this challenge of reducing alarm fatigue. The system’s approach begins with an interface designed from the ground up in cooperation with drillers in the field. The ePVT interface, presented on the large, touch screen Rig Display, is designed so rig crews can get the information they need within seconds, and determine if they need to take action.
Under the hood, the newest version of the system includes a powerful feature called Adaptive Alarms, which exploits machine learning to enable a far more robust approach to the detection of kicks and mud losses.
Machine learning is a field of computer science related to pattern recognition and statistics. It enables computers to learn from data in order to make predictions, and to act like rational agents. Machine learning techniques are used to identify credit card fraud by recognizing deviation from normal spending behaviours. Machine learning is also behind targeted online advertising, search engines, and email spam filtering. It was also used to teach computers to automatically identify handwritten digits, improving correct digit classification from 92% in 1998 to over 99.75% correct in 2012, surpassing human performance.
Modern machine learning techniques provide powerful tools for automated control and warning systems. Pason has partnered with CoVar Applied Technologies to develop a system that uses machine learning to analyze drilling data and generate much smarter predictions of normal mud behaviour. This will help ePVT users with early detection of kicks, losses, and spills.
“An analysis of 12 holes drilled by 12 different rigs showed the standard alarm method generated an average of 14 false alarms an hour, while the adaptive alarm generated an average of one false alarm every 20 hours, for a 99% reduction in false alarms.”
Mud volume and flow rates fluctuate when rig crews carry out standard tasks, such as working pipe up and down, transferring mud between tanks, or when a connection is being made. Traditional alarms trigger whenever they detect a pre-defined variance in mud volumes and flow rates. Since these vary so much, alarms are a frequent sound on the rig.
The company’s adaptive alarms tool operates differently. Using machine learning techniques, this new system is able to estimate what the mud volumes and flow rates should be given the rig’s current operations. With the Event Detection upgrade, ePVT will only generate alarms when it detects changes that are not expected. That means drillers will have far fewer alarms to cope with every shift. An analysis of 12 holes drilled by 12 different rigs showed the standard alarm method generated an average of 14 false alarms an hour, while the adaptive alarm generated an average of one false alarm every 20 hours, for a 99% reduction in false alarms.
How does ePVT come up with the volume and flow estimates? It uses a wealth of data provided by the company’s Electronic Drilling Recorder (EDR). The system uses sensors to track pump rate (SPM), time and rates of change, block position, mud levels in every tank, and the flow out of well.
Unlike other kick detection systems, ePVT uses an array of standard sensors already found on most rigs. This is less expensive than manned services. With the algorithms provided by CoVar, sensors such as a flow paddle provide all the data necessary to run effective event detection.
This standard data from the EDR system is used as input for machine learning algorithms that generate expected flow rates and tank volumes in real time. Alarms only activate when the measured values differ from these expected values. Drillers can see graphs showing exactly how the measured and expected values differ.
This is especially useful during periods such as connections, when mud volumes fluctuate and flow is expected to drop to zero. In standard systems, rig crews tend to ignore the alarms that routinely go off during connections. Whether circulating or making a connection, the ePVT Event Detection system analyzes mud volume and flow to give advanced warning of kicks and losses. The first few times the crew makes a connection, the Event Detection system records the flowback data during the connection and comes up with a profile for that hole. In subsequent connections, the alarms only activate if the flowback data differs substantially from the range of values captured in the profile. With this method, kicks and losses can be identified much earlier.
Having the data is one thing. It is just as important to provide that data to drillers in a way that is easy for them to understand and use at a glance. That is why the company has continued to enhance the ePVT interface to meet the needs of users. The interface can be pulled up at the touch of a screen and shared across the rig site and can also be used to create digital trip sheets.
Where drillers have traditionally needed to fill in trip sheets by hand every time they trip out of the hole, they can now record trips using ePVT, which automatically captures trip information in a simple interface. This makes the task quicker to complete, and provides the information immediately to anyone else following the well’s progress. It is all designed to keep the driller focused on what he needs to watch, and minimize any distracting information.
Questions? To learn more about Event Detection contact [email protected] or visit pason.com.