The path to net-zero emissions requires the rapid deployment of carbon capture and utilisation (CCU) technologies. However, the commercial rollout of these technologies is hindered by significant challenges: from fluctuating energy prices and diverse policy landscapes to technological risks and capital intensity.
A study published in the Elsevier Journal of CO₂ Utilisation presents a novel, structured approach to address these challenges by integrating real options valuation (ROV) with reinforcement learning (RL). This combined framework offers a more dynamic and flexible strategy for assessing CCU deployment under uncertainty.
The proposed methodology consists of three core steps: Surrogate modelling to simplify complex process simulations; Uncertainty modelling to represent key market and technical variables; Assessment modelling using RL to determine optimal investment pathways over time.
Unlike conventional methods, the ROV-RL framework allows for strategic flexibility, such as delaying, expanding, or abandoning a project in response to evolving market conditions.
The framework of this study was demonstrated using a case study of CO₂ hydrogenation to methanol in Europe, a technology with a high technology readiness level (TRL). By modelling uncertainties such as methanol demand and hydrogen carbon intensity, the study showed that flexible investment strategies led to higher project value and greater cumulative emissions reductions.
This research highlights that investment flexibility is not just economically advantageous but also environmentally impactful, supporting broader climate goals. The framework also offers a foundation for incorporating additional ESG metrics, such as energy efficiency and product safety, into future decision-making models.
KEY FINDINGS
🔵 CCU technologies face major uncertainty and upfront risk.
🔵 Traditional economic models cannot handle multi-stage, flexible decisions well.
🔵 ROV offers flexibility (e.g., delay, expand, abandon).
🔵 RL helps solve complex, multi-period investment problems.
🔵 A new 3-step framework was tested on a CO₂-to-methanol case in Europe.
🔵 RL-guided decisions improved project value and reduced emissions.
🔵 The method can be extended to include broader ESG and risk metrics.
Find the study here.
