Urban microclimates can substantially alter building loads and the performance of distributed energy systems. However, most existing studies on integrated building–energy system optimization rely on standard typical meteorological year (TMY) data, whereas microclimate research has largely remained at the diagnostic stage of load impact assessment. Few studies have further propagated microclimate-induced load variations into the capacity sizing and operational optimization of active energy systems. To address this gap, this study proposes a novel and computationally efficient four-stage framework that couples local climate zone (LCZ) classification, mixed-integer linear programming (MILP), and automated machine learning (AutoML)-based surrogate modeling. The framework employs the NSGA-II algorithm to perform global optimization within a 13-dimensional continuous design space, thereby evaluating the Pareto trade-off between life cycle cost (LCC) and life cycle greenhouse gas emissions (LCGHG). A typical high-rise office building located in the hot-summer and cold-winter climate zone of Wuhan, China, is selected as the case study. The TMY baseline scenario is compared with three urban microclimate scenarios: LCZ 2, LCZ 5, and LCZ 6. The results indicate that microclimate impacts are strongly morphology-dependent. The urban heat island effect reduces winter heating loads across all LCZ scenarios; however, in the compact mid-rise zone represented by LCZ 2, mutual shading between buildings outweighs the adverse effect of higher ambient temperatures, resulting in the lowest cooling load. The surrogate models achieve high predictive accuracy, with mean absolute percentage errors (MAPEs) below 3.26% for LCC and 2.11% for LCGHG. High-fidelity verification of representative Pareto solutions further confirms the robustness of the surrogate models, with average errors of 2.36% and 1.95%, respectively. Under the TMY scenario, the balanced solution increases LCC by only 22.10% relative to the cost-optimal solution while reducing LCGHG by 38.90%. By contrast, the environment-optimal solution achieves a 49.90% emissions reduction but increases LCC by 102.90%. Under the balanced preference, the LCZ scenarios increase LCC by 0.12%–5.61% and LCGHG by 0.20%–0.62% relative to the TMY scenario. Moreover, directly applying the TMY-optimized design to LCZ 6 increases LCGHG by 12.20%, revealing a substantial design–reality gap caused by neglecting microclimate heterogeneity. This study demonstrates that incorporating microclimate effects into integrated building–energy system optimization is not merely a correction of boundary conditions, but a necessary step toward robust low-carbon design. Academically, the proposed framework advances microclimate research from load impact assessment to integrated passive–active system optimization. Practically, it provides architects, energy engineers, and urban planners with a rapid LCZ-oriented decision-support tool for early-stage design.
Chen et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: