Cost benefits of designing carbon capture processes for part-load operation of power plants
math.OC
Submitted: 2026-10-26
Updated: 2026-10-26
Comments: 32 pages, 11 figures
License: http://creativecommons.org/licenses/by/4.0/
The gist: As fossil power plants increasingly cycle to balance variable renewable generation, carbon capture plants must be designed for a wide range of part-load operating conditions…
Abstract
As fossil power plants increasingly cycle to balance variable renewable generation, carbon capture plants must be designed for a wide range of part-load operating conditions, to avoid excessive capital costs. However, accounting for the variability in design optimization yields higher computational complexity, and is therefore not a current design methodology. Here, we apply two-stage stochastic optimization to identify optimal carbon capture plant designs accounting for part-load flue gas conditions. We reduce computational complexity through a data-driven approach-to-equilibrium model of the absorption and desorption processes. Operating scenarios are derived from real-world part-load operation data of coal and natural gas combined cycle power plants. Our results demonstrate that designs optimized over multiple operating scenarios substantially reduce equipment sizes, lowering capital costs by 5-9 % at the expense of higher operating costs, resulting in a moderate reduction in total annualized cost by 0.5-1.6 %. The lowest costs are achieved when targeting an average capture target across operating scenarios, which allows CO2 capture to be shifted from full-load to part-load periods to reduce peak equipment sizes without compromising overall CO2 emission reductions. These cost benefits are robust across sensitivity analyses and increase with higher shares of part-load operation. Thus, part-load variability should be considered at the design stage, particularly for plants subject to variable load.
Related papers
- An Accelerated Proximal Gradient Method for Bures-Wasserstein Space
- FOMTools: Development and AI-Assisted Code Generation for First-Order Methods
- Lions and Muons: Optimization via Stochastic Frank-Wolfe under Heavy-Tailed Noise
- Dynamic Service Recommendation with Congestion-Dependent Joining: Near-Optimal and Constant-Factor Approximation Algorithms
- Integrating Learning and Optimization for Open Transport Capacity Pooling under Disruption
- Zeroth-Order Gradient Estimators under Decision-Dependent Noise