Abstract
ABSTRACT
The increasing penetration of variable renewable generation requires combined‐cycle gas‐turbine plants to follow load more frequently, while repeated cycling reduces part‐load efficiency and increases thermal stress. The unresolved problem is whether a utility‐scale plant retrofitted with two‐tank molten‐salt storage can follow a daily dispatch while respecting inventory, thermal, control, degradation, and economic constraints. This study develops a component‐based MATLAB/Simulink model of a 550 MW plant and a 488.6 MWhth storage system, including the gas‐turbine island, heat‐recovery steam generator, steam cycle, molten‐salt heater and steam generator, tanks, pumps, heat tracing, supervisory control, direct emissions, a repairable‐component availability framework, a normalized thermal‐cycle severity indicator, and a screening‐level economic model. The design point is calibrated with deviations below 2%. Dynamic credibility is assessed through conservation checks, state‐bound and limiting‐case tests, and comparison with an internal transient benchmark trajectory; independent plant transient validation remains unavailable. Under a 24 h renewable‐rich dispatch, net output ranges from 313.542 to 557.990 MW and daily generation is 10 296.98 MWh, 58.34 MWh above the storage‐free trajectory. The storage charges 230.50 MWhth and delivers 399.20 MWhth using both initial inventory and newly stored heat; a closed tank‐side balance gives a final state of charge of 9.06%. Direct carbon intensity varies from 0.302 to 0.364 tCO
2
/MWh. The value 1.70 × 10
−4
is retained only as a normalized cycle‐severity indicator, not a life fraction. With 31 USD/kWhth installed cost, an 8% discount rate, a 25‐year life, and 300 equivalent cycles per year, the levelized cost of stored heat is 15.39 USD/MWhth and ranges from 11.54 to 23.08 USD/MWhth for 400 to 200 cycles. The results demonstrate that storage can widen the dispatch range and shift recovered exhaust heat to high‐value periods; however, design and investment decisions require plant transient data, disclosed component maps and controller settings, material‐specific degradation laws, vendor reliability data, and annual market‐based optimization.