| In response to the double challenge of building energy conservation and environmental quality improvement in the urbanization in China in the new era,vigorous development of green buildings has reached consensus among all parties.Studies have shown that more than 40% of the building's energy-saving potential comes from the early design stage.In this context,this study hopes to propose a new approach on energy conservation optimization and daylighting optimization for office buildings at the early stage,and provide a new idea for government managers and owners in green building design.Firstly,through the review of the latest literatures at home and abroad,this paper summarizes the main obstacles that the performance-driven design process is difficult to be applied at the early design stage,including: the performance optimization tools and methods are difficult to achieve full integration with the architectural design,the building parameters have very large uncertainties,and architectural design software and performance simulation tools do not match well.In Chapter 2,this paper proposes the ideas and implementation steps of the two-way optimization design process for building performance,introduces the self-developed multi-objective performance optimization design software platform(MOOSAS).In Chapter 3,the building energy consumption prediction model was improved from 3 aspects: expansion of building plans,surface radiation amount correction,refinement of energy consumption algorithm,resulted in multi-level building energy consumption prediction algorithm.The accuracy of the algorithm was verified by comparison with a variety of international software simulation results and ARHRAE 140 standard test method.In Chapter 4,an energy consumption prediction method combined with artificial neural network was proposed.For the complex building body shape,the building blocks are decomposed first and the energy consumption of the decomposed blocks is predicted;finally,the energy consumption of these blocks is summed up.The training method of energy consumption prediction model was given and the accuracy of the proposed energy consumption prediction method was verified.In Chapter 5,a prediction method of daylight performance and lighting energy consumption combined with artificial neural network was proposed.The artificial neural network was used to realize the rapid prediction method of the dynamic daylighting performance indicators,and the lighting energy consumption of the external area and atrium.The artificial neural networks were used to solve the problem of high computational costs.Finally,based on the proposed optimization framework and performance prediction methods,4 case applications were carried out to demonstrate the proposed method: the inverse generation of the building scheme with the lowest energy consumption,real-time simulation optimization,the sensitivity analysis of the design parameters,and the two-way optimization of the energy consumption and daylighting performance.The innovations of this study include 3 points:(1)A multi-level prediction model of building energy consumption at the early design stage was proposed.The accuracy of the model has been verified through ASHRAE 140.The model can be used to forecast the energy trends at the early design stage.(2)A prediction method of building performance(energy consumption,daylight performance metrics)combined with artificial neural network was proposed.The method can be better applied to the simulation and inverse optimization of complex building forms at the beginning of the project.(3)The above innovations form the core of the multi-objective performance optimization design software platform(MOOSAS)for early design stage,and the validity of the method is tested through some cases' application and analysis.The research of this paper can provide ideas for the innovation of performancedriven green building design method and engineering application. |
PDF Full Text Request