Co-enhancement of energy efficiency and dynamic flexibility for indirect air-cooled systems under frequent peak-shaving: Trade-offs of cost, performance and robustness

Deep peak-shaving operation of coal-fired power units requires heat-exchange equipment to rapidly track varying loads. Indirect air-cooled units are central to power supply in water-scarce regions, and their cold-end system rejects heat to ambient air and possesses substantial thermal inertia, so its heat-rejection capacity responds sluggishly to changes in unit load and ambient temperature. Existing heat-exchanger designs address only rated-condition heat rejection, without jointly considering air-side, water-side, and heat-capacity effects, so steady-state efficiency and dynamic flexibility cannot be improved concurrently. To address this gap, an integrated optimization framework is proposed for heat-exchanger design, with the steady-state back pressure P s,end and the dynamic recovery time Δt as coordinated optimization objectives. Air-side correlations obtained from computational fluid dynamics are embedded in a cold-end dynamic model driven by turbine off-design boundaries. Neural-network surrogates with prediction errors below 0.33% accelerate the genetic-algorithm search, while Shapley-value analysis interprets the design parameters. For a 2 × 350 MW unit, the optimal structure reduces P s,end from 8.25 to 7.55 kPa and shortens Δt from 765.85 to 655.05 s, but requires higher circulating-pump power. Under the reference economic assumptions, the coal-cost saving (about 528 CNY/h) does not offset the additional pump-power cost (about 1867 CNY/h), giving net economic effect near −1339 CNY/h. This performance oriented structure is therefore not the lowest-cost design. The study provides a structural route to reconcile efficiency and flexibility in air-cooled units, supporting their peak-shaving role under high renewable penetration.