Research On Refined Optimization Of Passive Building Performance And Time-Domain Influencing Of Climate

Passive building design can reduce energy demand by adapting the building to climate characteristics and site conditions,and has become a broad consensus in building energy-saving design.With the improvement of building energy efficiency standards and the development of near-zero energy buildings,refined and collaborative passive building design will become the new design norm in the future.At the same time,the refined design also put forward higher requirements on the completeness and accuracy of the information of the uncertain influencing variables in the passive design process.Therefore,carrying out the analysis of uncertain influencing factors and improving the reliability of the increasingly refined and complicated building passive design system will become a problem that designers need to solve.This study re-examined the relationship between climate time-domain characteristics,passive building design methods,and passive performance improvement.Taking the refined design as the mainline,through establishing a random generation model of dynamic passive performance,proposing a multi-objective collaborative optimization method based on accuracy improvement,considering the impact of dynamic uncertain variables on passive performance,and combining with the passive performance adaptability analysis of the climate time-domain,systematically carried out the basic theory and method research of building passive performance improvement.The specific research work is as follows:(1)To quantify the transmission law of uncertain information of influencing variables,this thesis established a random generation model of building dynamic passive performance.Firstly,the key uncertain variables that affect the passive performance were identified,and a quantitative model of the uncertain influencing variables was constructed based on probability theory.Based on multiple design orientations including building energy demand,thermal comfort level,light comfort,economy,and environmental benefits,a comprehensive evaluation index system for building passive performance was established.Finally,the Monte Carlo simulation framework for generating random passive performance was constructed,which quantifies the transmission law of uncertain characteristics of influencing variables.This research laid the foundation of the model for obtaining building sample data for subsequent research.(2)The high-dimensional design space defined by multiple design parameters interactively will make the passive performance refined optimization model complicated and difficult to solve efficiently.By combining redundancy analysis,gradient boosting decision tree,and genetic algorithm,this research proposed a multi-level multi-objective optimization method for passive performance.To cover the set of design variables with complete and significant influence information in the optimization model,while reducing the model complexity,the redundancy analysis method was used to identify the significant design parameter variables from the perspective of multi-optimization objective integration.A meta-model based on Gradient Boosted Decision Trees was proposed,and the mapping relationship between significant design parameter variables and optimization objectives was established.Finally,the passive performance meta-model is used as the fitness function of the genetic algorithm to participate in the multi-objective optimization process,which improved the optimization accuracy while efficiently solving the “high-dimensional uncertain design parameter variable ? multiple passive performances” refined optimization model.Through case verification,compared with two conventional optimization methods,the proposed multi-level optimization method has advantages in improving optimization accuracy with error reduction by more than 8%.(3)Aiming at the passive design dominated by steady-state uncertain influencing variables such as design parameters,ignoring the influence of dynamic uncertain variables,this thesis studied the adverse impact of passive strategies on the passive design process,and constructs a passive performance optimization model under uncertain design scenarios.Firstly,the driving factors of two typical passive strategies for window ventilation and shading adjustment was discussed.Furthermore,guided by light comfort,thermal comfort,and energy-saving performance,an integrated control simulation model of passive strategy was established with the building operation mode as the constraint,and then an uncertain design scenario is set.Finally,the sensitivity analysis method,the uncertainty analysis method,and the multi-objective optimization method were used to explore the influence of the uncertain design scenario on the passive design from the three aspects of weight distribution of design parameter variables,building performance distribution,and optimization scheme.This research improves the information completeness of the uncertain influencing variables in the passive performance optimization model and responds to the orientation of the refined design.(4)Building climate zoning is an important tool for formulating energy-saving goals and guiding the building design,but the building climate zoning method based on the single meteorological parameter cannot meet the needs of refined design.Taking China’s cold climate zone as the research object,this research established a refinement model of building climate zoning that couples integrated climate variables and full design space.Firstly,the uncertainty analysis and sensitivity analysis methods were used to study the consistency of building energy demand characteristics and design characteristics within existing climate zones.Besides,the characterization capabilities of traditional climate zoning indicators(degree days)for building energy characteristics was also explored.Finally,to fully cover the correlation between different meteorological variables and the coupling between climate and passive full design space,based on statistical theory,this thesis proposed a zoning index that can reflect the characteristics of random cooling and heating energy demand.Based on this index,the combination of K-means cluster analysis and Bayesian discriminant analysis were used to divide the cold climate zone into four sub-climate zones.The smaller and homogeneous building climate zoning method helps to improve the rationality of energy-saving target formulation and the accuracy of passive building design guidance.(5)The existing building performance predictions are mainly based on the typical meteorological year and specific design schemes,which are difficult to respond to re-fined design requirements and climate change on time.Combining Monte Carlo random generation and statistical downscaling methods,this study established a passive performance prediction model for future buildings considering multi-dimensional variables including climate change,building design,and operating characteristics.First,based on the statistical downscaling method,the hourly weather data of five typical climate regions in China in the future mid-and long-term climate scenarios were obtained.Secondly,taking the design boundary of the near-zero energy building as the constraint,the passive random design space under the interaction of multiple thermal parameters was constructed using the random generation method.The evaluation index systems of two operation modes of free-running and mechanical cooling/heating were constructed respectively.Finally,based on hourly weather data and random design space,the Monte Carlo simulation method was used to construct the passive performance simulation models of typical residential buildings and office buildings in two operating modes.This prediction model reveals the influence rule of mid-to-long-term climate change on the passive performance of future buildings.