Systematic data-driven controller tuning for offdesign operation of a recuperated sCO2 cycle over a wide range of power levels and ambient temperatures

ABSTRACT This study presents a systematic and data-driven methodology for proportional-integral (PI) controllers tuning in a gain scheduling approach in a 5 MW th air-cooled recuperated supercritical carbon dioxide (sCO 2 ) Brayton cycle. Off-design operations of sCO 2 systems pose substantial control challenges due to strong nonlinearities in fluid properties near the critical point and pronounced interactions among multiple control loops. Most existing studies rely on manually tuned controllers or simple case-specific gain adjustments, limiting their applicability across wide operating conditions. Although gain scheduling has shown clear benefits, a generalized and fully systematic approach has not yet been developed. To address this gap, this work proposes a fully automatable framework that systematically determines optimal controller parameters as a function of boundary conditions. The methodology comprises: (1) steady-state analysis across a discretized grid of ambient temperatures (5–45 °C) and power levels (25–100%); (2) transfer function identification at each operating point via step-response testing; (3) controller parameter derivation using both MATLAB built-in tools and Internal Model Control (IMC) approach; and (4) systematic testing via transient simulation across the operating conditions. Application to Compressor Inlet Temperature (CIT) control of the cooler demonstrates that the IMC approach with a closed-loop time constant equal to 1/10 of the open-loop one achieves stable and efficient controller performance across the majority of operating conditions. Performance degrades only at extreme conditions of load and air temperature, where air mass flow rates becomes very low (10-30 % of nominal). This regime that should be avoided operationally or mitigated through cooler modularization.