Design of Experiments for Dynamic Test Runs in Solvent-Based CO₂ Capture Pilot Plants.

Area of Science:

• Chemical Engineering
• Process Systems Engineering

Background:

• Pilot plant test runs are resource-intensive and often rely on lengthy steady-state data collection.
• Dynamic test runs offer a more efficient method for gathering substantial data within shorter timeframes.

Purpose of the Study:

• To present a systematic design of dynamic experiments (DoDEs) for enhanced model parameter identifiability.
• To compare the effectiveness of different dynamic input signals for parameter estimation.

Main Methods:

• Designed dynamic experiments utilizing pseudorandom binary sequence (PRBS) and Schroeder-phase multisine input signals.
• Conducted tests in the Pilot Solvent Test Unit (PSTU) to collect transient data.
• Applied dynamic data reconciliation and parameter estimation techniques to the collected data.

Main Results:

• Dynamic experiments yielded superior parameter estimates compared to hundreds of steady-state runs in a nonreactive system.
• Accurate parameter estimation was achieved using only 24 hours of dynamic data for the MEA-H2O-CO2 system.

Conclusions:

• Dynamic experiments are a highly efficient approach for pilot plant studies, significantly reducing data collection time.
• The designed DoDEs enable robust parameter estimation, improving model accuracy and resource utilization in pilot plants.