Avoiding methane emissions charges with the help of predictive maintenance

Methane is a component of natural gas. It is also labeled as a greenhouse gas (GHG) which, according to the Oxford’s English dictionary, is “a gas that contributes to the greenhouse effect, global warming effect, by absorbing infrared radiation similar to carbon dioxide and chlorofluorocarbons.” However, according to the U.S. Environmental Protection Agency (EPA), methane is 25 times more potent in trapping atmospheric heat than carbon dioxide,¹ and concentrations of methane have doubled in the last 200 yr. The practice of gas flaring has existed since oil production began due to its presence restricting the production of the more energy intensive crude oil from oil reservoirs.

Methane discharges are seen as one of the main causes of carbon emissions. Consequently, the world is enforcing stricter regulations that confine flaring practices. Still, methane leaks occur from oil production during periods of distressed operations.

One of the latest initiatives to reduce carbon emissions was delivered in the August 2022 update of the “Inflation Reduction Act Methane Emissions Charge: In Brief.”² Through the act, the oil and gas industry is facing a first-of-its-kind requirement in 2024 to report their GHG emissions to the U.S. EPA’s Greenhouse Gas Emissions Reporting Program (GHGRP) with charges due for methane discharges. This is the first time the federal government has directly imposed a charge, fee or tax on GHG emissions. However, the act also provides financial considerations for operators to change equipment and execute technological improvements. Preventing methane discharge is an urgent priority for the energy sector in North America and requires urgent action.

How do methane discharges occur? Historical discharges include deliberate gas flaring to encourage greater release of the crude oil product. Such flaring is now ostensibly forbidden or restricted and, if it does occur, it will be subject to hefty charges. This may restrict or seriously change operational activities, but certain emissions can be unintended consequences of operations activities. Such releases can be sudden, often resulting from a process or mechanical failure that causes an unplanned or emergency shutdown. Nominally, these occur without warning, affecting profitable operations. However, they can also subject the operation to distressing side effects, such as personnel safety issues and/or cause the over-pressurization to lift automatic flare valves or vessel safety valves, causing the release of hydrocarbons.

During normal operations, a hydrocarbon process is usually within safe operating conditions. However, as much as 10% of operations can occur during transient conditions such as shutdowns, startups or transitions.³ During this time, an astonishing 50% of all safety incidents occur. Clearly, avoiding such transients is the key to preventing safety and gas release incidents. The question becomes: How does a company do that?

An approach to prevent methane discharges. Some of the answers can be found in advanced analytical solutions immersed in machine learning- (ML-) based artificial intelligence (AI). Such analytical prowess is the basis of novel methods of predictive and prescriptive maintenance which can proactively prevent methane discharges. Such systems learn the patterns of normal equipment behavior so that abnormalities in mechanical or process operations that are harbingers of failure conditions stand out. Equally, such software also detects explicit patterns of degradation that will lead to failure if unattended, giving weeks and months’ notice of when equipment will fail.

This is the heads-up that the operators need to adjust operations to completely avoid a shutdown or for maintenance to plan and prepare a safe, orderly shutdown well in advance of any equipment distress. Such early warnings ensure action to permit timely repairs and avoid carbon releases. Additionally, the leading solutions also provide self-learning and knowledge-sharing to save cognitive advice that results in prescriptive guidance on what to do the moment an event is detected. A self-learning system is one way in which software producers can enhance the skills and productivity of new employees that have not witnessed the same experiences as the seasoned staff. This will also contribute to discharge avoidance.

Extreme forecasting of degradation events also provides more optionality on when to take the downtime. Time to plan downtime coupled with a comprehensive view of the entire operation permits plant personnel to see how an operational maintenance decision affects the entire organization. It is evident how the downtime event impacts plant scheduling, determines feedstock or delivery, and oversees risk and safety issues. Considerations can be made across cross-functional activities to explore the optimum repair window while minimizing production losses, considering the needs of the whole business from production to maintenance and from supply chain to engineering.

Protecting the bottom-line. Knowing that one emissions event can release more carbon than the whole rest of the year, predictive maintenance emboldened by ML/AI can show which pieces of equipment will fail and when, and deliver sufficient lead times to prevent carbon discharges. If a leak does occur, the software enables action to stop it in short order. Preventing methane discharges has a direct impact on environmental, social and governance (ESG) and sustainability priorities, as well as the bottom-line performance. GP

 

LITERATURE CITED

1. U.S. EPA, “Global Methane Initiative: Importance of Methane,” June 2022, online: https://www.epa.gov/gmi/importance-methane
2. Congressional Research Service, “Inflation Reduction of Methane Emissions Charge: In Brief,” August 2022, online: https://crsreports.congress.gov/product/pdf/R/R47206
3. Miklovic, D., “ Why 10% of your operations cause 50% of your safety incidents,” LNG Research, January 26, 2016, online: https://blog.lnsresearch.com/why-10-of-your-operations-cause-50-of-your-safety-incidents

 

Mike Brooks is the Global Director, APM Solutions, at AspenTech. Previously, he was Chief Operating Officer of Mtell, which pioneered ML for managing the health of industrial equipment. Brooks has also served as a venture executive with Chevron Technology Ventures and has held senior roles at five startups. He began his career as an engineer at Esso and Chevron.