Urban building energy modeling (UBEM) plays a crucial role in analyzing building energy use and has shown that large-scale UBEM can drive energy efficiency and sustainable development through urban retrofitting. However, large-scale UBEM presents challenges, including data acquisition workload, frequent parameter adjustments, and long simulation times. Moreover, the workflow connection between UBEM and urban retrofitting pathways remains unclear. Thus, this study proposes a framework that combines a fast, large-scale UBEM method in a Python environment with renewable energy integration to create energy demand-supply networks. The proposed UBEM method utilizes R-tree for geometric repairs, while EPPY efficiently batch-sets simulation parameters based on building function and performs batch simulations with EnergyPlus to quantify energy demand. Energy demand-supply networks are constructed through an improved gravitational model that considers location and functional mix, along with social network analysis. The framework was applied to Nanjing's historic city center in Jiangsu, China, covering 23,279 buildings across 551 blocks with six functional categories, totaling 54,232,464 m² of building area. The energy use map reveals that high energy use intensity blocks (over 175 kWh/(m²·year)) are distributed in the southern, particularly in commercial and old residential areas, while educational blocks have the highest photovoltaic (PV) potential. The simulation time using the multi-threaded EPPY method was only 14.1% of that with the conventional Ladybug tool for 75 buildings, and about 46.2% for ten urban blocks. Even with PV potential considered, 84.2% of blocks have energy demand exceeding supply, necessitating additional retrofitting. Combined retrofits are more effective than individual retrofits, achieving up to a 16.7% energy savings. This study provides new insights into large-scale UBEM and offers valuable decision-making support for energy-efficient urban retrofitting.

Typical energy-efficient retrofit studies based on urban building energy models face challenges in quickly obtaining appropriate retrofit solutions and often ignore the unexpected outcomes caused by inherent model uncertainty. To solve it, this study proposes a decision support framework that integrates a hybrid urban building energy model (UBEM) method, NSGA-Ⅱ, and TOPSIS to obtain rapidly the optimal energy-efficient retrofit solutions that take into account model uncertainty. The study took the building groups in Sipailou campus as a case study and identified 76 “stable solutions” and 149 “active solutions” that minimize energy consumption, carbon emission, and life-cycle cost (LCC) over 30 years from 40,353,607 retrofit schemes. Key findings include that when considering model uncertainty, the quantities, types, and ranks of optimal retrofit solutions have changed. When the error of baseline UBEM validation is within ±5% and considering uncertainty transmission from energy simulation to ANN model, the energy-saving potential of optimal retrofit schemes has expanded from [63.78, 65.05]% to [60, 68.75]%, carbon-saving potential has shifted from [63.69, 64.09]% to [59.92, 67.79]%, and the LCC has changed from [–40.68, 14.59] × 10⁶ to [–38.25, 16.97] × 10⁶ Yuan. This study provides decision makers with a scientific approach to consider the potential uncertainties and risks associated with optimal retrofit solutions.

The students receiving higher education boosted a total increase of 416.45% in China in last 20 years, resulting in newly built campuses reaching over 4.4 billion m². Therefore, implementing low-carbon development on university campuses is an important part of achieving carbon neutrality in China. In this study, the old and new campuses of Southeast University in China were selected and the Rhino Grasshopper tool was used to create and calibrate their energy model with real electricity data to ensure the 20% error range. The calibrated energy model was used to set up four base scenarios under different development paths in year 2030 and 2050, including natural development, campus construction, policy-oriented, and sustainable development. The simulation indicates that campus construction leads to the greatest increase in carbon emissions, with the old campus and new campus experiencing a 16.7% and 162.9% rise, respectively, compared to the current situation. In contrast, policy-oriented scenarios result in the most significant reduction in emissions, decreasing by 121.4% and 114.5% for each scenario, respectively. Only policy-driven approaches will enable both campuses to achieve carbon neutrality by 2050. The driving factor decomposition analysis indicates that in no-policy-intervention scenarios, the primary contributors to carbon emissions are short-term climate fluctuations and aging equipment. Conversely, in scenarios with government intervention, the pivotal elements are the implementation of renewable energy and the development of low-carbon technologies. The results of the static scenario combination show that the old campus has a significant lower average carbon emission of 7,080 t than 279,090 t of the new campus in 2050. However, the new campus shows higher potential, with a proportion of 38.3% achieving carbon neutrality in the combination results, compared to 17.2% for the old campus. The study results offer insights into the pathway for universities to achieve carbon neutrality, emphasizing the significance of policy direction and the adoption of renewable energy.

This research investigates retrofitting strategies for multifunctional spaces within educational buildings, employing agent-based and performance-based modeling to support decision-making. An experimental matrix was developed, reflecting three usage scenarios (reading, exhibition, lecture) across four retrofitting schemes. An agent-based model was developed to delineate intricate human behaviors in space and examined the self-organizing behaviors of 30 agents for each scheme in every scenario, evaluating six metrics on spatial efficiency and visual experience. Calibrated models, derived from real data and processed through DesignBuilder software, evaluated three metrics: energy use, thermal comfort, and visual comfort. The research then incorporated metrics from the agent-based model and performance simulation to develop a method for discussing the decision-making process in retrofit strategies. The findings indicate that the optimal retrofitting solution for multifunctional spaces is heavily influenced by the distribution of usage scenarios. Given the substantial influence of space metrics on selecting the optimal retrofit scheme, the proposed framework effectively facilitates decision-making for building retrofits by providing a holistic evaluation of both spatial and energy criteria.

Retrofitting a historic building under different national goals involves multiple objectives, constraints, and numerous potential measures and packages, therefore it is time-consuming and challenging during the early design stage. This study introduces a systematic retrofitting approach that incorporates standard measures for the building envelope (walls, windows, roof), as well as the heating, cooling, and lighting systems. Three retrofit objectives are delineated based on prevailing Chinese standards. The retrofit measures function as genes to optimize energy-savings, carbon emissions, and net present value (NPV) by employing a log-additive decomposition approach through energy simulation techniques and NSGA-II, yielding 185, 163, and 8 solutions. Subsequently, a weighted sum method is proposed to derive optimal solutions across multiple scenarios. The framework is applied to a courtyard building in Nanjing, China, and the outcomes of the implementation are scrutinized to ascertain the optimal retrofit package under various scenarios. Through this retrofit, energy consumption can be diminished by up to 63.62%, resulting in an NPV growth of 151.84%, and maximum rate of 60.48% carbon reduction. These three result values not only indicate that the optimal values are achieved in these three aspects of energy saving, carbon reduction and economy, but also show the possibility of possible equilibrium in this multi-objective optimization problem. The framework proposed in this study effectively addresses the multi-objective optimization challenge in building renovation by employing a reliable optimization algorithm with a computationally efficient reduced-order model. It provides valuable insights and recommendations for optimizing energy retrofit strategies and meeting various performance objectives.