Market Insights

Advancing Carbon Capture: AI Model Could Boost Efficiency of CCS Technologies

February 01, 2024 by Claire Turvill

Researchers believe an artificial intelligence model can increase the reliability and efficiency of carbon capture and storage technologies.

The capacity and reliability of renewable energy have seen substantial growth. However, the significant emissions from the current power sector highlight the need for nations to address power-related emissions and align with global climate goals. Carbon capture and storage (CCS)  is important in the shift towards net zero emissions. It entails capturing and storing carbon emissions from industrial and power generation processes. 

 

The Boundary Dam Power Station in Saskatchewan uses carbon capture and storage.

The Boundary Dam Power Station in Saskatchewan uses carbon capture and storage. Image used courtesy of Wikimedia Commons

 

Despite several operational industrial-scale CCS projects capable of capturing and storing over a million tons of carbon dioxide (CO2) annually, its widespread adoption is hindered by the high energy costs associated with solvent-based carbon capture methods employed in these systems.

Researchers from the University of Surrey have employed an artificial intelligence model to manage variabilities in energy supply and carbon dioxide concentration to improve the overall efficiency of carbon capture systems.

 

Inefficient Carbon Capture Technology

Fuel combustion in power plants releases CO2 as a byproduct. If left alone, the emission contributes to the buildup of greenhouse gases in the atmosphere. Directing the flue gas through water containing limestone can capture the CO2 and form bicarbonate, a benign byproduct. This CCS technique is called enhanced weathering.

The continuous operation of CCS systems relies on energy supply from renewable sources and faces challenges due to the volatility in energy availability. Fluctuations in the quantity of CO2 released further contribute to these issues, leading to extreme efficiency losses and energy inefficiencies in many continuously operating CCS systems.

 

Conventional carbon capture and storage process

Conventional carbon capture and storage process. Image used courtesy of National Energy Technology Laboratory 

 

To reduce some of the challenges associated with industrial carbon capture, scientists from the University of Surrey have proposed a Model Predictive Control framework focused on optimizing both the rate of carbon dioxide capture and energy consumption in a reactor specifically designed for capturing CO2 emissions from a coal power plant’s flue gas.

Initially developed and demonstrated using an Electrolysis-Wet scrubbing reactor, the framework can be applied to other CCS technologies and relates to broader areas of reaction chemistry and engineering applications.

 

AI Application for Carbon Capture

The Surrey researchers’ framework focused primarily on carbon capture technology's efficiency and energy utilization. They instructed the AI system to monitor variations in energy supply and CO2 concentration and make adjustments to the CCS to result in significant energy savings.

Addressing these challenges, researchers created a multi-objective predictive optimization framework to proactively optimize reactor conditions, particularly the superficial gas and liquid flow rates. This optimization aims to maximize the CO2 capture rate while simultaneously minimizing non-renewable energy consumption.

 

The AI-enhanced model predictive control framework.

The AI-enhanced model predictive control framework. Image used courtesy of the authors

 

The scientists trained the model system to anticipate the event fluctuations, enabling it to reduce water pumping when CO2 capture needs are lower or when renewable energy supplies are diminished. 

Although this model was tested on an enhanced weather CCS system, researchers hope the framework’s principles can apply more broadly. The model can potentially assist anyone seeking to capture and store more CO2 with reduced energy requirements, regardless of the specific process.