CoVar Technologies Solutions

Adaptive Alarms™

Automated adaptive alerts are improving driller situational awareness and early, effective response to unexpected occurrences.

Modern drilling rigs have a variety of sensors and displays. These are continuously presented to the driller with real-time data about the state of the rig and the state of the well. Additionally, many rigs are remotely monitored by off-site operations centers. With the overwhelming quantity of data available on the rig and off-site, the challenge becomes reacting to the rig data in ways that result in improved safety and efficiency of drilling operations. When automatic alerts are available to process the data they are too often ignored because of false alarms (“the boy who cried wolf” problem). When automated alerts are not available, important drilling data is often ignored because it is buried in a slew of other data (“information overload”).

CoVar is addressing this problem at the rig site with the development of its proprietary Adaptive Alarms™ engine. By applying patented machine learning techniques to drilling alarm data, we can track real-time instrumentation data and alert the driller when the instrumentation is deviating from its normal fingerprint. As the well is drilled, the algorithm continually updates its fingerprint of the data, adapting to changing well conditions in ways that minimizes false alarms and maximizes sensitivity.

This algorithms underpinning our Adaptive Alarms™ software forms the foundation of our Influx Detection at Pumps Stop (IDAPS) module. Deployed to RTOC consoles, this software automatically detects pumps stop events such as connections and flow checks and monitors mud flow and pit volume. Successive connections build a fingerprint of a normal connection. As soon as deviations from this fingerprint occur that could be consistent with influx, an alert is issued to the well monitor, estimating probability of influx and influx volume. Using pattern recognition algorithms, the software is robust to sensor failures, data validity issues, and data drop-outs.

This unique software package seamlessly integrates with WITS feeds over TCPIP and can be installed on any Windows-based virtual machine. It offers a simple but powerful tool to operation centers worldwide, providing unprecedented reliability in early influx indications, an intuitive interface, and easy-to-use analysis tools for both routine connection monitoring and exception handling.