| In response to China’s push towards carbon neutrality and carbon peak,the development of nearly zero-energy buildings has become a priority.However,the relevant technology systems,norms,and standards for nearly zero-energy public buildings are not yet mature in China,resulting in many problems in the construction process.This study focuses on nearly zero-energy office buildings in Shandong Province,and aims to provide new design and optimization approaches for the region’s building industry.To achieve this goal,this study comprehensively expounds the relevant definitions of nearly zero-energy buildings and summarizes the technical means of performance research at home and abroad.Based on this,a research method and overall technical path is established.A comprehensive survey and analysis of the performance characteristics,spatial characteristics,and enclosure structure of nearly zero-energy office buildings in Shandong Province is conducted,and a benchmark model is abstracted.Using this model,a performance model is constructed using Grasshopper.The study further investigates the influence mechanism of the enclosure structure parameter design on energy consumption through single-factor quantitative simulation comparison between office buildings at the general level and nearly zero-energy level.Energy consumption-oriented design parameter recommendation values for the enclosure structure of nearly zero-energy office buildings in Shandong Province are obtained through comprehensive optimization of the energy consumption target.The accuracy and feasibility of the optimization results are verified through the actual project of the Comprehensive Experimental Building of Shandong Jianzhu University.Furthermore,a Gaussian process regression model is used to construct a proxy model of the benchmark performance model,thereby improving the prediction efficiency of the energy consumption target.Based on the shortcomings of the single-objective optimization research method,this study proposes and constructs a GANN prediction and multi-objective controllable optimization system.The system establishes energy consumption,thermal comfort,and cost as the optimization objectives,and outputs 4000 sets of structured data containing27 variables and 3 objectives.The artificial neural network multi-layer perceptron model is trained and verified in Python using this dataset,and a proxy model for the multi-objective parametric performance model of nearly zero-energy office buildings in Shandong is obtained.The proxy model is used as a function and input into the NSGA2 algorithm programmed with controllability for design variable parameters and objective parameters in Python for multi-objective optimization.The study concludes by providing a theoretical basis for the design and research of other types of buildings in other regions,in addition to addressing many problems in the design of nearly zero-energy office buildings in Shandong under local climate conditions.To facilitate the use of the prediction and optimization system by designers,a "Shandong Province nearly zero-energy office building prediction and controllable optimization software" and "model generation visualization plugin" are encapsulated using Python and Grasshopper,respectively.Designers can quickly complete the prediction,optimization,and model generation of building performance in the early stages of design according to their own needs through the system software.Overall,this study offers valuable insights for the design and optimization of nearly zero-energy office buildings in Shandong Province and beyond. |
