Research On Morphological Optimization Design Of Office Building Complex In Cold Regions Based On Machine Learnin

The current construction of office building complexes in cold regions of China is large and shows a trend of intensification.The contradiction between the performance needs of buildings such as energy saving,comfort and capacity brought about by the rapid development of the construction process and the environmental crisis such as global warming and energy scarcity is becoming more and more prominent.Architects need to create architectural solutions with a new perspective and improve building performance by drawing on the results of multidisciplinary developments.Firstly,the shape optimization design of office building complexes in cold regions is analyzed from the aspects of shape characteristics and performance-driven design modes of office building complexes,and the optimization design process is built based on technical methods such as parametric modeling,performance simulation,machine learning,multiobjective optimization and multi-attribute decision-making.Secondly,the shape optimization model is constructed according to the construction process,which includes three sub-models:the sampling survey method is used to obtain the shape quantitative indicators of typical buildings,its value range and step size are constrained,and the parametric modeling of typical office building complexes in cold regions is completed by relying on Rhino and Grasshopper;Based on the Lunch Box plug-in,the shape performance mapping model is established,and the training data of the model is obtained by the simulation of Ladybug & Honeybee performance analysis software,and the trained model can realize rapid prediction of building performance;the built-in optimization algorithm of Octopus,multi-attribute decision and parallel coordinate diagram are used to build an optimization decision model,which can support designers to optimize the building form and make design decisions on the optimal solution of Pareto with performance as the driving force.Thirdly,the sensitivity analysis of building form and performance is carried out,the key form design parameters are extracted,and based on this,the optimization model is used to optimize the form according to different optimization design scenarios,and based on the analysis results,suitable design strategies are proposed at three levels: parameter selection,form design,and solution decision.Finally,taking a practical project as an example,the process,model and strategy are verified,and the results show that the energy saving,thermal comfort and application potential of PV systems can be improved through optimization.The proposed optimization model can improve the efficiency and decision-making accuracy of shape design of office building complexes in cold regions,and enhance the tradeoff between the requirements of composite performance of buildings;the research analyzes and refines the shape characteristics of office building complexes in cold regions,which can provide quantitative reference for the shape design of office building complexes;the sub-model of shape performance mapping of office building complexes in cold regions can quickly predict energy intensity,indoor thermal comfort,and solar radiation in different forms.The established optimization decision-making sub-model can comprehensively weigh the form and performance information,reduce the difficulty of decision-making,and improve the scientific decisionmaking.In addition,the proposed design strategy can provide suggestions and guidance for the shape design of office building complexes from the aspects of key parameter selection,shape design tendency,and optimal scheme decision.