Its about “Real Time” to reduce Methane Emissions
What IIOT can do for Methane Emissions in Oil and Gas
by Keerthana Jayakumar and Gaurav Jayakumar
The time between when a methane leak is detected in an oil and gas site and when it is repaired could be as bad as 16 days¹.
This is not surprising considering that current regulations only mandate periodic leak checks ex — monthly inspections². If there are 60m³ ¹ of methane released per hour, this means ~300000 tonnes of CO2 equivalent gas will be released into the atmosphere before a fix is applied. To put this into a metric we can visualize, that’s 300000 cars on gasoline driving for half a year continuously.³
What if we continuously monitor emissions?
There has been growing interest in developing continuously monitoring solutions to detect methane emissions immediately. This will have the obvious benefit of minimizing these undetected emissions.
In order to implement a robust end to end continuous monitoring solution you need these pieces:
The EPA and ECCC prescribe two types of sensors to monitor methane emissions:
- Organic vapor analyzers like photoionization devices
- Optical gas imaging (OGI) cameras⁴
These are generally handheld devices but there is currently a lot of activity in developing continuous monitoring methane sensors in the IOT space. Gaurav and I felt the buzz at the IOT North Conference in Calgary this past November. For a glimpse into what’s going on, check out this article from the University of Calgary.
Oil and Gas companies are great at the last piece of this flow diagram as well. Actions here are defined as the repairs required to mitigate a leak.
What do we do with the data?
Where we see potential improvement is the three pieces in the middle: Data Collection, Data Analysis and Insights.
There have been recent exciting waves in the IIOT space that make these tasks exponentially easier.
It is difficult to generate Insights (ex- Machine learning model results and reports) using Industrial data. Id like to explain why using an analogy.
Imagine your boss needs you to organize 5 minute standup meetings everyday with key workers in your department. This seems like a simple enough task but you find yourself running into two challenges.
- The workers are in remote parts of the facility and do not have phones. You need to go out and bring them into your meeting. Sometimes this is easy and sometimes they offer resistance. Some of them are even in completely inaccessible locations.
- They speak multiple languages and do not understand each other. When you bring them into the meeting, you need a translator present to decipher what they are saying and relay the message to you. The translator doesn’t always understand the business so there is a potential for errors in the translations.
With these problems, a 5 minute meeting can turn into 5 hours.
Now imagine the workers are sensors and applications on the site, the languages they speak are device protocols, you are the machine learning model trying to derive insights and the translators are the data engineers cleaning the data for you.
Depending on the protocol the workers (sensors) use, some are easier to connect to than others. Its not uncommon for data engineers to take months to connect to and clean data from disparate sources.
Introducing IIOT (Industrial Internet of Things)
An IIOT architecture is the equivalent of having everyone in the same meeting room speaking the same language. All the devices and applications from the site are connected to each other and communicating in real time. This has the potential to shorten the time to insight from months to weeks. We can focus on the models and reports instead of wasting time and money cleaning data. In addition to this, results from machine learning models can be sent to existing legacy applications in the site to do things like trigger SCADA alarms or automatically issue work orders.
In the next few articles of our series, we are going to discuss how this can be achieved using a popular IIOT protocol and standard (MQTT/SparkPlugB), create solutions in the two biggest cloud platforms (AWS and Azure) and compare the results.
If you missed the last article in our methane monitoring series, here it is.
References:
1. bc-cas-mefs-chapters-1–3-err2–20200224.pdf (bcogc.ca)
2. Highwood Emissions Management — Emissions Reduction Experts
3. What exactly is 1 tonne of CO2? We make it tangible. — Climate Neutral Group
