Artificial Intelligence: One answer to reducing greenhouse gases at industrial facilities
October 03, 2018
October 03, 2018
Data collection and analyzation can lead to cleaner, more efficient energy—a must in the coming decades
There is a paradoxical conversation currently taking place within the energy sector. On one hand, energy consumption globally is expected to grow by 28% between 2015 and 2040 according to The U.S. Energy Information Administration’s latest International Energy Outlook. There is a debate about which sources of energy will meet this demand, but technology development, cost, and availability will ultimately settle this dispute.
During the same period, many nations have committed to reducing 2005 greenhouse gases (GHG) globally by 30% by 2030 as part of the Paris Climate Accord. That means we are living in a time where we know we need to produce significantly more energy while simultaneously drastically reducing the GHGs of our existing infrastructure.
To meet these two potentially opposing demands, two fundamental things need to happen: new sources of energy need to have lower associated GHGs and existing energy infrastructure needs to reduce GHG emissions significantly.
It is the later of those two options that will have the biggest impact. I am not alone in this thinking, according to DNV-GL, an international accredited registrar, “efficiency gains play a far greater role in helping to cut emissions over the coming two decades than the combined contribution of the switch to wind, solar, and electric vehicles.” So how do we make these existing infrastructure projects more efficient?
The solution on the surface sounds simple: be more efficient. However, performing energy efficiency reviews of existing infrastructure is specialized work.
Energy efficiency reviews for older facilities involve gathering significant amounts of data, in multiple forms ranging from engineering drawings to datasheets, from excel spreadsheets to scanned equipment catalogs.
This data, when organized and understood is compared with live operational data, which enables us to understand how a facility is set up and how energy is being consumed.
The process of organizing and analyzing data is currently the most time-consuming element of performing energy efficiency reviews. In the past 10 years the amount of data available has grown by 50 times and is growing exponentially. This increasing amount of data is a bottleneck in our ability to perform analyses. It also limits our ability to look at larger and more complex energy facilities.
Recognizing the significant increases in the amount of data available from energy facilities and the increasing demand from customers to find energy efficiencies, it’s clear a new method is needed.
To find a new method we partnered with global IT experts, including IBM. When choosing partners, Stantec was looking for organizations that are members of the Canadian Oil Sands Innovation Alliance (COSIA) and are trying to find ways to bring big data and artificial intelligence (AI) solutions to the natural resources industry.
We are living in a time where we know we need to produce significantly more energy while simultaneously drastically reducing the GHGs of our existing infrastructure.
Together we began working with a large midstream gas processing facility in Alberta, Canada, to digitize live energy usage and GHG emissions to identify energy and GHG savings. Utilizing Stantec engineering expertise with IBM big data and AI capabilities, we greatly simplified data organizing tasks. As an example, one task that previously took 30 days to organize and analyze was now complete in 10 minutes!
Work is still ongoing to refine the workflow process so energy efficiency and GHG savings processes are synchronized. The long-term goal is to have a repeatable program for other customers and leverage the insights of AI further.
We are also looking into a large sized project with the specific aim to reduce energy and GHG emissions by 5%. Correlating the data using conventional techniques would have been a yearlong project. Utilizing AI should trim the process to six months.
Oil and gas facilities are not much different than your home. You know you want to reduce your carbon footprint, so you analyze your personal use. For example, you realize your old fridge is consuming a lot of energy. You might consider replacing it. Also, you turn down the furnace or air conditioning to save costs while you’re away at work. In both cases you reduce your carbon footprint and save on your energy bill. Similarly, the goal of all this data and analysis is to increase energy savings for our clients and reduce GHGs for the planet.
Digitizing and understanding energy usage on an hourly and daily basis is a critical step in first identifying major power users within a plant. Corresponding the energy usage to GHG also provides customers with a visible baseline to start meaningful efficiency.
The approach so far has been to correlate vast amounts of data from a design perspective and complete live data correlations to give us a complete energy usage picture.
All this data is enabling us to identify patterns in energy inefficiencies. Patterns often indicate areas for further investigation and optimization.
If the dual challenge to increase energy production while reducing GHG emissions is going to be met, a new level of innovative solutions is going to be needed—and fast!
Bringing AI into the mix is an essential first step. Early indications of how much energy and GHG can be eliminated are highly encouraging. We are on the right path, with major progress being completed, expect to hear a lot more from this as we look to deploy this globally.