Alarms are a fundamental part of any SCADA or analytics system. Ideally, they direct attention to potential problems that cost oil companies time and lost production. What has happened, however, is an alarm flood. Every sensor can be set to alarm and it is not uncommon for hundreds of SCADA alarm notifications to trigger in a day.
Because of this, alarms have not done anything other than add more confusion and inaction in many oilfields. It becomes impossible to know what the critical alarms are and which ones are not. Most oil and gas companies don’t know the proper alarm management philosophy and how to implement it, so they live with the barrage of alarms and the fatigue that ultimately costs them.
Let’s put some definition around what alarm fatigue is:
Becoming numb to alarm notifications because there are so many that it becomes impossible to make sense of whether they are meaningful or not. As a result, they are either turned off for good or ignored like the neighborhood car alarm.
With so many alarms chirping, time and money are likely being lost somewhere. But it becomes impossible for engineers and operators to know where the real problems and opportunities lie in the oilfield. The alarms lose their meaningfulness. So the default alarm handling in SCADA is to let things play out and see which wells begin to shut down before taking action.
Alarm fatigue means no one is getting ahead of problems or significantly changing how they work. It ultimately renders a company’s big investment in data useless.
We have seen two basic strategies or alarm philosophies when it comes to alarm setting. Most companies opt for the Catch Everything Approach, which quickly creates the condition of alarm fatigue. It feels the safest for engineers and operators, but it doesn’t lead to great results. The greatest danger for Catch Everything is that alarms get ignored so important problems are missed.
In reaction to the swarm of alarms, some companies turn many of them off or dial down the thresholds so there is more silence. We call this the Stop the Madness Approach. On the opposite end of the pendulum, the greatest danger for Stop the Madness is completely missing key events. Stop the Madness is a false alarm management approach that yields no better results than Catch Everything.
At the center of the pendulum is where alarms that matter most alert engineers and operators to take action. Most companies don’t get to the center because they don’t have a guiding alarm strategy and the time to iteratively adjust and learn from different settings. This means alarms are rarely adjusted or stay calibrated to detect the most meaningful events that matter and require action.
eLynx is an expert in helping companies create their alarming strategy and make full use of the settings and notification capabilities of eLynx’s SCADA alarm management system. We have built and proven effective SCADA alarm management standards and practices. Our product has features related to schedules of reporting so they don’t all come at once and different tiers of alerting that take into account the immediate need to tend to an alarm. Our subject matter experts and customer service team will outline your strategy and help you put it into practice.
The most significant innovation to date in analytics is our anomaly detection application. As you see in the figure below, it shrinks the number of alerts you receive because it precisely identifies problems and types of anomalies with near certainty. There is no guessing or wondering if this is a meaningful alert or what the underlying cause might be. You will have the root cause identified and be able to take specific action to resolve it before it costs time and money.
eLynx’s Anomaly Detection is the game changer for the industry because this means virtually no false alarms or missed events. It provides the ability to catch and accurately identify problems when they first occur or predict if they will occur in the future. This is real-time and real-world predictive analytics.
While our SCADA alarm strategy will put your company more in the center of the pendulum, this is still an area that contains imprecisions. Alarms are manually input set points for only one variable at a time. They don’t distinguish between a normal fluctuation or a real problem either.
eLynx’s Anomaly Detection looks across a wide range of variables simultaneously using time series data. Our advanced machine learning algorithms combine supervised learning and unsupervised learning techniques to compute how these behavioral patterns match up to specific incidents in the past. It uses real-time data to do complex pattern matching across a vast and ever-growing data set for a well.
Our anomaly detection tool forecasts with close-to-perfect accuracy what the root cause of the problem is. If eLynx detects an anomaly, engineers and operators will be able to determine exactly how much time and money will be lost in the future if it goes unattended. They can immediately take action or schedule plans to perform maintenance based on the priority and financial impact of a problem. To learn more about Anomaly Detection, watch our video case study.
We help companies get to the center of the pendulum with their SCADA alarm systems and pinpoint problems and opportunities faster and with more accuracy with Anomaly Detection. Only then will the investment in better data to make better decisions finally pay off.