ChampionX’s Event Detection AI Model | Defend Your Assets from Super Emitters and OLRE

The U.S. Environmental Protection Agency (EPA) recently approved the first third-party surveyor for the Super Emitter Program, signaling the official launch of this initiative.

With the Super Emitter program in motion, it is essential for the operators to understand the program, how it interacts with Subpart W regulations, and why it's important to deploy advanced technology. Read our blog to learn more.

Understanding the Super Emitter Program: What Operators Need to Know

What is the Super Emitter Program?

The Super Emitter Program, introduced under OOOOb, directly impacts how the operators manage methane emissions. Under this initiative, the EPA authorizes the third-party organizations to identify and report methane emission events exceeding 100 kg/hr, classified as "super-emitter events." Once a super-emitter event is detected, these third parties notify the EPA with the leak location and relevant details for verification. The EPA then informs the operator to investigate and determine the cause of the leak. If the leak is unpermitted or results from equipment out of compliance, the operator is placed on a list available to the public.

What are the changes to Subpart W?

The EPA has added a new reporting category to Subpart W called “Other Large Release Events (OLRE),” requiring the operators to report methane emissions exceeding 100 kg/hr of their Subpart W calculated emissions. For example, if a valve that was previously calculated to emit 2 kg/hr under Subpart W suddenly starts emitting 102 kg/hr or more, the leak must now be categorized under this section. Additionally, if the leak source does not have a category under Subpart W, any leak above 100 kg/hr will be categorized under OLRE.

Accurate reporting of these events depends on precise start and stop times – also known as bookending – to calculate the total volume of gas released.  Start times can be determined using monitored data or assumed to begin at your most recent survey, up to 91 days before the event was identified. End times are based on when emissions stop or when repairs are completed. These changes mean you need robust monitoring and reporting systems to ensure compliance and avoid unnecessary exposure.

Introducing Soofie® Event Detection AI Model

As the Super Emitter Program launches and Subpart W undergoes significant updates, it is critical to have precise event alerting and tracking. Emissions Technologies at recently released our new Event Detection AI model. This model is designed to identify methane leak events in real time using the minute-by-minute methane concentration data from the Soofie® sensors in the field. By grouping leak events and identifying their start and end times, the system sends alerts if the user-set emission mass threshold is breached during an event

Key Features:

• State-of-the-art anomaly detection AI model: Utilizes advanced algorithms to detect leaks.
• Integrated UI logic: Ensures accurate event detection and grouping.
• Customizable emission thresholds: Allows users to set specific thresholds for alerts.
• Real-time alerts: Provides immediate notifications for quick action.

The Event Detection AI model identifies normal behavior by finding patterns and relationships in the dataset, by learning directly from the data without needing pre-assigned labels. Anomalies or leak events are detected automatically, and the system generates smart alerts via the UI Event Bifurcation Logics. This enhances event-based quantification for mitigation strategies and reporting. Users can set thresholds for the leaked mass of methane, beyond which they will receive AI alerts. Data analytics on historical AI results can help identify seasonality in site emissions and potentially equipment-based emission seasonality so that user can Identify the recurring reasons for leaks and develop mitigation strategies.

How it works

Within our Enterprise platform, users can set up one or multiple alerts based on mass thresholds for detected events. For example, if the AI detects the start of an event and the user has set a threshold of 25 kg for methane emissions, the UI will calculate the methane emitted during the event based on emission rates and time. Once the emitted methane exceeds the user-set threshold, the event is posted on the Enterprise system, and email and/or mobile alerts are sent out. This method of threshold based smart AI alerting provides the users with the option to focus on only considerable emission events and avoid any alert fatigue. See Figure 1 for an example of how the event is recorded in our system.

Figure 1. Event Detection AI - event start (Red dotted line) and event end (Blue dotted line), based on the user set methane emission mass threshold.

In the figure above, there are three events detected by the Event Detection AI and posted by the UI logics after the emission mass was greater than the set thresholds on this site. The events marked indicate the start and end of the events, corresponding to the changes in the methane concentration data. As soon as the threshold is breached, the event is posted on the Enterprise platform, and the alerts are sent out. Once the event has ended, the UI logic is active for some time to ensure that the event has indeed ended. Using the information provided in the alerts, the event-based quantification can be used for mitigation strategies.

In conclusion, the new Event Detection AI model represents a significant advancement in methane leak detection and management. By leveraging state-of-the-art AI technology and real-time data from Soofie® sensors, this model provides users with precise, actionable insights to mitigate emissions effectively. The customizable thresholds and real-time alerts ensure that users can focus on critical events without being overwhelmed by unnecessary notifications. This innovative solution not only enhances operational efficiency but also supports environmental sustainability efforts. We invite our customers to explore the capabilities of this new model and experience the benefits of enhanced emissions monitoring and management.

Learn more about Soofie®: Continuous Emissions Monitoring | SOOFIE® | ChampionX